WO2003003307A1 - Image processing apparatus and method, and image pickup apparatus - Google Patents
Image processing apparatus and method, and image pickup apparatus Download PDFInfo
- Publication number
- WO2003003307A1 WO2003003307A1 PCT/JP2002/006385 JP0206385W WO03003307A1 WO 2003003307 A1 WO2003003307 A1 WO 2003003307A1 JP 0206385 W JP0206385 W JP 0206385W WO 03003307 A1 WO03003307 A1 WO 03003307A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- pixel
- background
- pixel data
- frame
- foreground
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N5/00—Details of television systems
- H04N5/222—Studio circuitry; Studio devices; Studio equipment
- H04N5/262—Studio circuits, e.g. for mixing, switching-over, change of character of image, other special effects ; Cameras specially adapted for the electronic generation of special effects
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Definitions
- Image processing apparatus and method and imaging apparatus
- the present invention relates to an image processing apparatus and method, and an image capturing apparatus, and more particularly to an image processing apparatus and method that takes into account a difference between a signal detected by a sensor and the real world, and an image capturing apparatus.
- motion blur occurs when the moving speed of the object is relatively high.
- the present invention has been made in view of such a situation, and has an object to specify an area of a background image, an area of an image of a moving object, and an area of an image in which mixing occurs.
- the purpose is to:
- the image processing apparatus is configured to convert pixel data of a peripheral frame around a frame of interest corresponding to a pixel of interest of the frame of interest of image data into a background object serving as a background among a plurality of objects of the image data. And extract the target pixel data of the target pixel existing in the target frame.
- a relational expression generating means for generating a plurality of relational expressions indicating the relation between the target pixel data and the background pixel data for the eye pixel; and a plurality of relational expressions in the real world corresponding to the target pixel based on the relational expression.
- a mixing ratio detecting means for detecting a mixing ratio indicating the mixing state of the object; a prediction error calculating means for calculating a prediction error by substituting the mixing ratio detected by the mixing ratio detecting means into a relational expression;
- the area to which the pixel of interest belongs is a mixed area in which a plurality of objects are mixed, and is formed at the tip of the foreground object, which is the foreground of the plurality of objects, in the movement direction.
- An uncovered background area that is a covered background area or a mixed area that is formed at the rear end of the foreground object in the direction of motion.
- a covered background area for specifying whether there is an object, an uncovered background area specifying means, and an area to which the pixel of interest belongs is a foreground area consisting only of foreground object components constituting a foreground object, or a background object.
- a foreground area background area specifying means for specifying whether or not the background area consists only of the background object component.
- the mixture ratio detecting means detects a foreground object component included in the pixel of interest based on the relational expression, corresponding to the pixel of interest, and detects a mixture ratio.
- the prediction error can be calculated by substituting the detected mixture ratio and the foreground object component included in the pixel of interest into the relational expression.
- the relational expression generation means extracts the pixel data of the peripheral frame corresponding to the target pixel as background pixel data corresponding to the background object, and extracts the pixel data of the target pixel and the vicinity of the target pixel in the target frame. Extracts neighboring pixel data of the neighboring pixels located, and, for the pixel of interest, a plurality of relationships indicating the relationship between the pixel data of interest, the neighboring pixel data, and the background pixel data corresponding to the pixel data of interest or the neighboring pixel data. An expression can be generated.
- the relational expression generating means performs the first approximation that the foreground object components included in the target pixel data and the neighboring pixel data are equal, and the change of the mixture ratio in the mixed region.
- a plurality of relational expressions can be generated based on a second approximation that the transformation is linear with respect to the position of the pixel in the mixed area.
- the relational expression generating means performs the first approximation that the foreground object components included in the target pixel data and the neighboring pixel data are equal, and that the change in the mixture ratio in the mixed region is based on the position of the pixel in the mixed region.
- a plurality of relational expressions can be generated based on a second approximation that is assumed to be planar.
- the mixture ratio detection means can detect the mixture ratio by solving a plurality of relational expressions by the least square method.
- the relational expression generation means extracts the pixel data of the frame before the target frame as background pixel data, and determines that the target pixel belongs to the uncovered background area.
- the pixel data of the frame following the frame of interest can be extracted as background pixel data, and a plurality of relational expressions can be generated.
- the relational expression generating means extracts mixed pixel data in which a plurality of objects are in a mixed state from the attention frame and the surrounding frame based on the movement of the foreground object in accordance with the target pixel, and Based on the motion of the background object, the background pixel data corresponding to the background object is extracted from a frame different from the frame from which the mixed pixel data was extracted, and multiple relational expressions are calculated. Can be generated.
- the relational expression generating means is based on the first approximation that the foreground object components corresponding to the mixed pixel data are equal, and the second approximation that the mixed pixel data extracted from the frame of interest and the surrounding frame is constant. Thus, a plurality of relational expressions can be generated.
- the relational expression generating means extracts mixed pixel data in which a plurality of objects are in a mixed state from the attention frame and the surrounding frame based on the movement of the foreground object in accordance with the target pixel, and Corresponding to each of the following, based on the motion of the background object, from the frame before the frame from which the mixed pixel data was extracted, The background pixel data corresponding to the background object is extracted, a plurality of relational expressions are generated, and the power bird background area and the uncovered background area specifying means uncover the area where the prediction error is equal to or greater than a predetermined threshold. It can be specified as an area.
- the relational expression generating means extracts mixed pixel data in which a plurality of objects are in a mixed state from the attention frame and the surrounding frame based on the movement of the foreground object in accordance with the target pixel, and Based on the motion of the background object, background pixel data corresponding to the background object is extracted from the frame following the frame from which the mixed pixel data is extracted, and a plurality of relational expressions are generated.
- the bird background area / covered background area specifying means can specify an area where the prediction error is equal to or larger than a predetermined threshold value as a covered background area.
- pixel data of a peripheral frame around a frame of interest corresponding to a pixel of interest of the frame of interest of image data is converted into a background object as a background among a plurality of objects of the image data.
- target pixel data of the target pixel existing in the target frame is extracted, and a plurality of relational expressions indicating the relationship between the target pixel data and the background pixel data are obtained for the target pixel.
- a relational expression generation step to generate, and a mixture ratio detection step for detecting a mixture ratio indicating a mixture state of a plurality of objects in the real world based on the relational expression.
- the area to which the pixel of interest belongs is a mixed area formed by mixing a plurality of objects, and a covered back formed at the tip of the foreground object, which is the foreground of the plurality of objects, in the direction of movement.
- the mixture ratio detection step detects a foreground object component included in the pixel of interest corresponding to the pixel of interest based on the relational expression, detects a mixture ratio, and calculates a prediction error.
- the prediction error can be calculated by substituting the mixture ratio detected by the above processing and the foreground object component included in the pixel of interest into the relational expression.
- the pixel data of the peripheral frame corresponding to the target pixel is extracted as background pixel data corresponding to the background object, and the target pixel data of the target pixel and the vicinity of the target pixel in the target frame are extracted. Extracts neighboring pixel data of the neighboring pixels located and generates a plurality of relational expressions for the pixel of interest, showing the relationship between the pixel data of interest, the neighboring pixel data, and the background pixel data corresponding to the pixel data of interest or neighboring pixel data. You can make it.
- the relational expression generation step includes a first approximation that the foreground object components included in the target pixel data and the neighboring pixel data are equal, and a change in the mixture ratio in the mixed region is linear with respect to the position of the pixel in the mixed region. It is possible to generate a plurality of relational expressions based on the second approximation that is appropriate.
- the relational expression generation step includes a first approximation that the foreground object components included in the target pixel data and the neighboring pixel data are equal, and a change in the mixing ratio in the mixed region is a plane relative to the position of the pixel in the mixed region. It is possible to generate a plurality of relational expressions based on the second approximation that is appropriate.
- the mixture ratio can be detected by solving a plurality of relational expressions by the least square method.
- the relational expression generation step is as follows. If the target pixel belongs to the covered background area, the pixel data of the frame before the target frame is extracted as background pixel data, and the target pixel is set in the uncovered background area. If they belong, The pixel data of the frame next to the frame of interest can be extracted as background pixel data, and a plurality of relational expressions can be generated.
- the relational expression generation step extracts mixed pixel data in which a plurality of objects are in a mixed state from the frame of interest and peripheral frames based on the movement of the foreground object in accordance with the pixel of interest. Based on the motion of the background object, the background pixel data corresponding to the background object is extracted from a frame different from the frame from which the mixed pixel data was extracted, and multiple relational expressions are extracted. Can be generated.
- the relational expression generation step includes a first approximation that the foreground object components corresponding to the mixed pixel data are equal, and a second approximation that the mixed pixel data extracted from the frame of interest and the surrounding frame is constant. Based on this, a plurality of relational expressions can be generated.
- the relational expression generation step extracts mixed pixel data in which a plurality of objects are in a mixed state from the frame of interest and peripheral frames based on the movement of the foreground object in accordance with the pixel of interest. Based on the movement of the background object, the background pixel data corresponding to the background object is extracted from the frame before the frame from which the mixed pixel data is extracted, and a plurality of relational expressions are generated.
- the power bird background area uncovered background area specifying step may specify an area where the prediction error is equal to or greater than a predetermined threshold to be an covered background area.
- the relational expression generating step extracts mixed pixel data in which a plurality of objects are in a mixed state from the frame of interest and peripheral frames, based on the movement of the foreground object, in accordance with the pixel of interest. Based on the movement of the background object, the background pixel data corresponding to the background object is extracted from the frame following the frame from which the mixed pixel data was extracted, and multiple relational expressions are generated.
- the bird's-background area uncovered background area specifying step is a step in which the area where the prediction error is equal to or more than a predetermined threshold is regarded as the covered background area. Can be specified.
- the program of the recording medium is a program for converting a pixel data of a peripheral frame around a target frame corresponding to a target pixel of the target frame of image data into a background object as a background among a plurality of objects of the image data. And extracting the target pixel data of the target pixel existing in the target frame and generating a plurality of relational expressions indicating the relation between the target pixel data and the background pixel data for the target pixel.
- a mixture ratio detection step for detecting a mixture state indicating a mixture state of a plurality of objects in the real world based on the relational expression, based on the relational expression, and a mixture ratio detection step,
- a prediction error calculation step of calculating a prediction error by substituting the calculated mixture ratio into a relational expression.
- the region to which the pixel of interest belongs is a mixed region formed by mixing a plurality of objects, and is formed at the leading end in the movement direction of the foreground object that is the foreground of the plurality of objects.
- a force background area that determines whether the area is a covered background area or a mixed area, which is an uncovered background area formed at the rear end of the foreground object in the direction of motion.
- the step of identifying the covered background area and the area to which the pixel of interest belongs is either a foreground area consisting only of the foreground object components that make up the foreground object, or a spine consisting only of the background object component that makes up the background object.
- a foreground area / background area specifying step for specifying whether the area is a scene area.
- the mixture ratio detection step detects a foreground object component included in the pixel of interest corresponding to the pixel of interest based on the relational expression, detects the mixture ratio, and calculates the prediction error.
- the prediction error can be calculated by substituting the mixture ratio detected by the above processing and the foreground object component included in the pixel of interest into the relational expression.
- the pixel data of the peripheral frame corresponding to the target pixel is extracted as background pixel data corresponding to the background object, and the pixel data of the target pixel is extracted. Extract the pixel data of interest and the neighboring pixel data of the neighboring pixel located in the vicinity of the pixel of interest in the frame of interest, and for the pixel of interest, correspond to the pixel data of interest and neighboring pixel data, and the pixel data of interest or neighboring pixel data A plurality of relational expressions indicating the relation between the background pixel data can be generated.
- the relational expression generation step includes a first approximation that the foreground object components included in the target pixel data and the neighboring pixel data are equal, and a change in the mixture ratio in the mixed region is linear with respect to the position of the pixel in the mixed region. It is possible to generate a plurality of relational expressions based on the second approximation that is appropriate.
- the relational expression generation step includes a first approximation that the foreground object components included in the target pixel data and the neighboring pixel data are equal, and a change in the mixture ratio in the mixed region indicates a position of a pixel in the mixed region.
- a plurality of relational expressions can be generated based on a second approximation that is assumed to be planar.
- the mixture ratio can be detected by solving a plurality of relational expressions by the least square method.
- the relational expression generation step if the target pixel belongs to the covered background area, the pixel data of the frame before the target frame is extracted as background pixel data, and the target pixel belongs to the uncovered background area.
- the pixel data of the frame next to the frame of interest can be extracted as background pixel data, and a plurality of relational expressions can be generated.
- the relational expression generating step extracts mixed pixel data in which a plurality of objects are in a mixed state from the frame of interest and peripheral frames, based on the movement of the foreground object, in accordance with the pixel of interest. Based on the motion of the background object, the background pixel data corresponding to the background object is extracted from a frame different from the frame from which the mixed pixel data was extracted, and multiple relational expressions are extracted. Can be generated.
- the relational expression generation step the first approximation that the foreground object components corresponding to the mixed pixel data are equal, and extracted from the frame of interest and the surrounding frame A plurality of relational expressions can be generated based on the second approximation that the mixed pixel data is constant.
- the relational expression generation step extracts mixed pixel data in which a plurality of objects are in a mixed state from the frame of interest and peripheral frames based on the movement of the foreground object in accordance with the pixel of interest. Based on the motion of the background object, the background pixel data corresponding to the background object is extracted from the frame before the frame from which the mixed pixel data was extracted, and multiple relational expressions are generated.
- the power bird background area uncovered background area specifying step can specify an area in which the prediction error is equal to or greater than a predetermined threshold as the uncovered background area.
- the relational expression generation step extracts mixed pixel data in which a plurality of objects are in a mixed state from the frame of interest and peripheral frames based on the movement of the foreground object, corresponding to the pixel of interest, and extracts the mixed pixel data. Based on the motion of the background object, the background pixel data corresponding to the background object is extracted from the frame following the frame from which the mixed pixel data is extracted, and a plurality of relational expressions are generated. In the uncovered background area specifying step, the area where the prediction error is equal to or larger than a predetermined threshold can be specified as the covered background area.
- the program according to the present invention stores, in a computer, pixel data of a peripheral frame around the frame of interest corresponding to the pixel of interest of the frame of interest in the image data, as a background object as a background among a plurality of objects of the image data. Is extracted as background pixel data corresponding to, the target pixel data of the target pixel existing in the target frame is extracted, and a plurality of relational expressions are generated for the target pixel, indicating the relationship between the target pixel data and the background pixel data.
- a relational expression generation step a mixture ratio detection step of detecting a mixture ratio indicating a mixture state of a plurality of objects in the real world based on the relational expression, corresponding to a target pixel, and a mixture ratio detection step Prediction error calculation step for calculating the prediction error by substituting the mixture ratio detected by
- the region to which the pixel of interest belongs is a mixed region in which a plurality of objects are mixed, and the motion direction of the foreground object, which is the foreground of the plurality of objects, based on the Is a covered background area formed at the front end of the foreground object, or an uncovered background area formed at the rear end of the foreground object in the movement direction of the foreground object.
- the area to which the pixel of interest belongs is a foreground area consisting only of the foreground object components that make up the foreground object, or a background object that makes up the background object
- a foreground area background area identification step for identifying whether the area is a background area consisting only of components is executed. It is characterized by having
- the mixture ratio detection step detects a foreground object component included in the pixel of interest corresponding to the pixel of interest based on the relational expression, detects the mixture ratio, and calculates the prediction error.
- the prediction error can be calculated by substituting the mixture ratio detected by the above processing and the foreground object component included in the pixel of interest into the relational expression.
- the pixel data of the peripheral frame corresponding to the target pixel is extracted as background pixel data corresponding to the background object, and the target pixel data of the target pixel and the vicinity of the target pixel in the target frame are extracted. Extracts neighboring pixel data of the neighboring pixels located and generates a plurality of relational expressions for the pixel of interest, showing the relationship between the pixel data of interest, the neighboring pixel data, and the background pixel data corresponding to the pixel data of interest or neighboring pixel data. You can make it.
- the relational expression generation step includes a first approximation that the foreground object components included in the target pixel data and the neighboring pixel data are equal, and a change in the mixture ratio in the mixed region is linear with respect to the position of the pixel in the mixed region. It is possible to generate a plurality of relational expressions based on the second approximation that is appropriate.
- the relational expression generation step includes a first approximation that the foreground object components included in the target pixel data and the neighboring pixel data are equal, and a mixing ratio in the mixed region.
- a plurality of relational expressions can be generated based on a second approximation that the change of the pixel is planar with respect to the position of the pixel in the mixed area.
- the mixture ratio can be detected by solving a plurality of relational expressions by the least square method.
- the relational expression generation step if the target pixel belongs to the covered background area, the pixel data of the frame before the target frame is extracted as background pixel data, and the target pixel belongs to the uncovered background area.
- the pixel data of the frame next to the frame of interest can be extracted as background pixel data, and a plurality of relational expressions can be generated.
- the relational expression generation step extracts mixed pixel data in which a plurality of objects are in a mixed state from the frame of interest and peripheral frames based on the movement of the foreground object in accordance with the pixel of interest. Based on the motion of the background object, the background pixel data corresponding to the background object is extracted from a frame different from the frame from which the mixed pixel data was extracted, and multiple relational expressions are extracted. Can be generated.
- the relational expression generation step includes a first approximation that the foreground object components corresponding to the mixed pixel data are equal, and a second approximation that the mixed pixel data extracted from the frame of interest and the surrounding frame is constant. Based on this, a plurality of relational expressions can be generated.
- the relational expression generation step extracts mixed pixel data in which a plurality of objects are in a mixed state from the frame of interest and peripheral frames based on the movement of the foreground object in accordance with the pixel of interest. Based on the movement of the background object, the background pixel data corresponding to the background object is extracted from the frame before the frame from which the mixed pixel data is extracted, and a plurality of relational expressions are generated.
- the bird's-eye background area uncovered background area specifying step may specify an area where the prediction error is equal to or greater than a predetermined threshold as an uncovered background area.
- the relational expression generation step includes extracting mixed pixel data in which a plurality of objects are in a mixed state from the frame of interest and peripheral frames, based on the movement of the foreground object, corresponding to the target pixel. Based on the motion of the background object, the background pixel data corresponding to the background object is extracted from the frame following the frame from which the mixed pixel data was extracted, and multiple relational expressions are generated.
- the power bird background area uncovered background area specifying step an area in which the prediction error is equal to or larger than a predetermined threshold can be specified as a covered background area.
- An imaging device includes: an imaging unit that outputs a subject image captured by an imaging element having a predetermined number of pixels having a time integration effect as image data including a predetermined number of pixel data;
- the pixel data of the peripheral frame around the frame of interest corresponding to the pixel of interest is extracted as background pixel data corresponding to the background object that is the background of the plurality of objects of the image data, and
- a relational expression generating means for extracting a plurality of relational expressions indicating the relation between the target pixel data and the background pixel data for the target pixel by extracting target pixel data of an existing target pixel.
- a mixing ratio detecting means for detecting a mixing ratio indicating a mixed state of a plurality of objects in the real world.
- a prediction error calculating means for calculating a prediction error by substituting the mixing ratio detected by the mixing ratio detecting means into a relational expression; and, based on the prediction error, a region to which a target pixel belongs and a plurality of objects mixed.
- a force or a mixed area which is a cover background area formed on the leading end side in the direction of movement of the foreground object which is a foreground of the plurality of objects,
- a covered background area for identifying whether or not an uncovered background area is formed at the rear end of the foreground object in the direction of movement of the foreground object. Is the foreground area consisting of only the foreground object components that compose the background object, or the spine that composes the background object.
- Foreground area background specifying whether a background area consisting of only Obujuku preparative component Region specifying means.
- the mixture ratio detecting means detects a foreground object component included in the pixel of interest based on the relational expression, corresponding to the pixel of interest, and detects a mixture ratio.
- the prediction error can be calculated by substituting the detected mixture ratio and the foreground object component contained in the target pixel into the relational expression.
- the relational expression generation means extracts the pixel data of the peripheral frame corresponding to the target pixel as background pixel data corresponding to the background object, and extracts the pixel data of the target pixel and the vicinity of the target pixel in the target frame. Extracts neighboring pixel data of the neighboring pixels located, and, for the pixel of interest, a plurality of relationships indicating the relationship between the pixel data of interest, the neighboring pixel data, and the background pixel data corresponding to the pixel data of interest or the neighboring pixel data. An expression can be generated.
- the relational expression generating means performs the first approximation that the foreground object components included in the target pixel data and the neighboring pixel data are equal, and that the change in the mixture ratio in the mixed region is based on the position of the pixel in the mixed region.
- a plurality of relational expressions can be generated based on a second approximation that is linear.
- the relational expression generating means performs the first approximation that the foreground object components included in the target pixel data and the neighboring pixel data are equal, and that the change in the mixture ratio in the mixed region is based on the position of the pixel in the mixed region.
- a plurality of relational expressions can be generated based on a second approximation that is assumed to be planar.
- the mixture ratio detection means can detect the mixture ratio by solving a plurality of relational expressions by the least square method.
- the relational expression generation means extracts the pixel data of the frame before the target frame as background pixel data, and determines that the target pixel belongs to the uncovered background area.
- the pixel data of the frame following the frame of interest can be extracted as background pixel data, and a plurality of relational expressions can be generated.
- the relational expression generating means extracts mixed pixel data in which a plurality of objects are in a mixed state from the attention frame and the surrounding frame based on the movement of the foreground object in accordance with the target pixel, and Based on the motion of the background object, the background pixel data corresponding to the background object is extracted from a frame different from the frame from which the mixed pixel data was extracted, and multiple relational expressions are calculated. Can be generated.
- the relational expression generating means is based on the first approximation that the foreground object components corresponding to the mixed pixel data are equal, and the second approximation that the mixed pixel data extracted from the frame of interest and the surrounding frame is constant. Thus, a plurality of relational expressions can be generated.
- the relational expression generating means extracts mixed pixel data in which a plurality of objects are in a mixed state from the attention frame and the surrounding frame based on the movement of the foreground object in accordance with the target pixel, and Based on the motion of the background object, the background pixel data corresponding to the background object is extracted from the frame before the frame from which the mixed pixel data is extracted, and a plurality of relational expressions are generated.
- the bird's-eye-background-area-covered-background-area specifying means can specify an area in which the prediction error is equal to or greater than a predetermined threshold as an uncovered background area.
- the relational expression generating means extracts mixed pixel data in which a plurality of objects are in a mixed state from the attention frame and the surrounding frame based on the movement of the foreground object in accordance with the target pixel, and Based on the motion of the background object, background pixel data corresponding to the background object is extracted from the frame following the frame from which the mixed pixel data is extracted, and a plurality of relational expressions are generated.
- the bird background area / covered background area specifying means can specify an area where the prediction error is equal to or larger than a predetermined threshold value as a covered background area.
- the pixel data of the frame is extracted as background pixel data corresponding to a background object serving as a background of the plurality of objects of the image data, and the pixel data of the pixel of interest existing in the frame of interest is extracted.
- a plurality of relational expressions indicating the relationship between the target pixel data and the background pixel data are generated, and based on the relational expression, indicate a mixed state of a plurality of objects in the real world corresponding to the target pixel.
- a mixture ratio is detected, and a prediction error is calculated by substituting the detected mixture ratio into a relational expression.Based on the prediction error, a region to which a pixel of interest belongs is mixed with a plurality of objects.
- a covered background formed at the front end side in the movement direction of the foreground object which is the foreground of the plurality of objects in the mixed area It is specified whether the area is a mixed area or an uncovered background area formed at the rear end of the foreground object in the movement direction of the foreground object. It is specified whether it is a foreground region consisting only of the foreground object components that make up the background object or a background region consisting only of the background object components that make up the background object.
- FIG. 1 is a diagram showing an embodiment of an image processing apparatus according to the present invention.
- FIG. 2 is a block diagram illustrating the image processing apparatus.
- FIG. 3 is a diagram illustrating imaging by a sensor.
- FIG. 4 is a diagram illustrating the arrangement of pixels.
- FIG. 5 is a diagram illustrating the operation of the detection element.
- FIG. 6A is a diagram illustrating an image obtained by capturing an object corresponding to a moving foreground and an object corresponding to a stationary background.
- FIG. 6B is a diagram illustrating a model of an image obtained by capturing an object corresponding to a moving foreground and an object corresponding to a stationary background.
- FIG. 7 is a diagram illustrating a background area, a foreground area, a mixed area, a covered background area, and an uncovered background area.
- Figure 8 shows the time-wise expansion of the pixel values of the pixels that are adjacent to each other in a row in the image of the object corresponding to the stationary foreground and the object corresponding to the stationary background. It is a model figure.
- FIG. 9 is a model diagram in which pixel values are developed in the time direction and a period corresponding to the shutter time is divided.
- FIG. 10 is a model diagram in which pixel values are developed in the time direction and a period corresponding to the shirt time is divided.
- FIG. 11 is a model diagram in which pixel values are developed in the time direction and a period corresponding to the shirt time is divided.
- FIG. 12 is a diagram illustrating an example in which pixels in a foreground area, a background area, and a mixed area are extracted.
- Fig. 13 is a diagram showing the correspondence between pixels and a model in which pixel values are expanded in the time direction.
- Fig. 14 is a model in which pixel values are expanded in the time direction and a period corresponding to the shirt time is divided.
- FIG. 15 is a model diagram in which pixel values are developed in the time direction and a period corresponding to the shutter time is divided.
- FIG. 16 is a model diagram in which pixel values are developed in the time direction and a period corresponding to the shutter time is divided.
- FIG. 17 is a model diagram in which pixel values are developed in the time direction and a period corresponding to the shutter time is divided.
- FIG. 18 is a model diagram in which pixel values are developed in the time direction and a period corresponding to the shirt time is divided.
- FIG. 19 is a flowchart illustrating the process of adjusting the amount of motion blur.
- FIG. 20 is a block diagram showing the configuration of the mixture ratio calculation unit 103. As shown in FIG.
- FIG. 21 is a diagram showing an example of an ideal mixing ratio.
- FIG. 22 is a model diagram in which pixel values are developed in the time direction and a period corresponding to the shirt time is divided.
- FIG. 23 is a model diagram in which pixel values are developed in the time direction and a period corresponding to the shirt time is divided.
- FIG. 24 is a diagram illustrating selection of a pixel.
- FIG. 25 is a diagram illustrating selection of a pixel.
- FIG. 26 is a block diagram illustrating a configuration of the estimated mixture ratio processing unit 201.
- FIG. 27 is a flowchart illustrating a process of calculating the estimated mixture ratio.
- FIG. 28 is a flowchart illustrating a process of calculating a mixture ratio estimation using a model corresponding to the covered background region.
- FIG. 29 is a block diagram showing another configuration of the mixture ratio calculating section 103. As shown in FIG.
- FIG. 30 is a diagram illustrating a straight line that approximates the mixture ratio H.
- FIG. 31 is a diagram illustrating a plane approximating the mixture ratio a.
- FIG. 32 is a view for explaining the correspondence of pixels in a plurality of frames when calculating the mixture ratio ⁇ .
- FIG. 33 is a flowchart illustrating the process of estimating the mixture ratio using a model corresponding to the covered background area.
- FIG. 34 is a block diagram showing still another configuration of the mixture ratio calculating section 103.
- FIG. 35 is a block diagram showing the configuration of the area specifying unit 104. As shown in FIG.
- FIG. 36 is a diagram for explaining the determination process of the uncovered background area determination unit 303.
- FIG. 37 is a diagram for explaining the determination process of the uncovered background area determination unit 303.
- FIG. 38 is a diagram illustrating an example of an input image.
- FIG. 39 is a diagram illustrating an example of the result of the area determination.
- FIG. 40 is a diagram illustrating an example of the result of the area determination.
- FIG. 41 is a diagram illustrating an example of an input image.
- FIG. 42 is a diagram illustrating an example of the result of the area determination.
- FIG. 43 is a diagram illustrating an example of the result of the area determination.
- FIG. 44 is a block diagram illustrating another configuration of the area specifying unit 104.
- FIG. 45 is a block diagram showing still another configuration of the area specifying unit 104.
- FIG. 46 is a block diagram showing still another configuration of the area specifying unit 104.
- FIG. 47 is a flowchart for explaining the area specifying process.
- FIG. 48 is a flowchart illustrating a specific process of the covered background area.
- FIG. 49 is a flowchart illustrating a specific process of the uncovered background area.
- FIG. 50 is a block diagram showing an example of the configuration of the foreground / background separation unit 105.
- FIG. 51A is a diagram illustrating an input image, a foreground component image, and a background component image.
- FIG. 51B is a diagram illustrating an input image, a foreground component image, and a background component image.
- FIG. 52 is a model diagram in which pixel values are developed in the time direction and a period corresponding to the shirt time is divided.
- FIG. 53 is a model diagram in which pixel values are developed in the time direction and a period corresponding to the shirt time is divided.
- FIG. 54 is a model diagram in which pixel values are developed in the time direction and a period corresponding to the shutter time is divided.
- FIG. 55 is a block diagram illustrating an example of the configuration of the separation unit 600.
- FIG. 56A is a diagram illustrating an example of a separated foreground component image.
- FIG. 56B is a diagram illustrating an example of the separated background component image.
- FIG. 57 is a flowchart for explaining the process of separating the foreground and the background.
- FIG. 58 is a block diagram illustrating an example of the configuration of the motion-blur adjusting unit 106.
- FIG. 59 is a diagram illustrating a processing unit.
- FIG. 60 is a model diagram in which the pixel values of the foreground component image are developed in the time direction, and the period corresponding to the shutter time is divided.
- FIG. 61 is a model diagram in which the pixel values of the foreground component image are developed in the time direction and the period corresponding to the shutter time is divided.
- FIG. 62 is a model diagram in which the pixel values of the foreground component image are developed in the time direction, and the period corresponding to the time interval is divided.
- FIG. 63 is a model diagram in which the pixel values of the foreground component image are developed in the time direction and the period corresponding to the shirt time is divided.
- FIG. 64 is a diagram illustrating another configuration of the motion-blur adjusting unit 106.
- FIG. 65 is a flowchart illustrating a process of adjusting the amount of motion blur included in the foreground component image by the motion blur adjustment unit 106.
- FIG. 66 is a block diagram illustrating another example of the configuration of the motion-blur adjusting unit 106.
- FIG. 67 is a diagram illustrating an example of a model that specifies a correspondence between a pixel value and a foreground component.
- FIG. 68 is a diagram for explaining calculation of a foreground component.
- FIG. 69 is a diagram for explaining calculation of a foreground component.
- FIG. 70 is a flowchart for describing the process of removing motion blur in the foreground.
- FIG. 1 is a diagram showing an embodiment of an image processing apparatus according to the present invention.
- the CPU (Central Processing Unit) 21 executes various processes according to a program stored in a ROM (Read Only Memory) 22 or a storage unit 28.
- ROM Read Only Memory
- RAM Random Access Memory 23 programs executed by the CPU 21 and data are stored as appropriate.
- the CPU 21, the ROM 22, and the RAM 23 are interconnected by a node 24.
- An input / output interface 25 is also connected to the CPU 21 via a bus 24.
- the input / output interface 25 is connected to an input unit 26 including a keyboard, a mouse, and a microphone, and an output unit 27 including a display, a speaker, and the like.
- the CPU 21 performs various processes in response to commands input from the input unit 26. Execute. Then, the CPU 21 outputs an image, a sound, or the like obtained as a result of the processing to the output unit 27.
- the storage unit 28 connected to the input / output interface 25 is composed of, for example, a hard disk, and stores programs executed by the CPU 21 and various data.
- the communication unit 29 includes the Internet and other Communicate with external devices via the network. In the case of this example, the communication unit 29 acts as an acquisition unit that captures the output of the sensor. Alternatively, the program may be acquired via the communication unit 29 and stored in the storage unit 28.
- the drive 30 connected to 5 drives the magnetic disk 51, optical disk 52, magneto-optical disk 53, or semiconductor memory 54 when they are mounted, and drives the programs recorded there. And data. The acquired programs and data are transferred to and stored in the storage unit 28 as necessary.
- FIG. 2 is a block diagram illustrating the image processing apparatus.
- each function of the image processing apparatus is implemented by hardware or software. That is, each block diagram in this specification may be considered as a block diagram of hardware or a functional block diagram by software.
- an image to be imaged that corresponds to an object in the real world is referred to as an image object.
- the input image supplied to the image processing device is supplied to an object extracting unit 101, a mixture ratio calculating unit 103, an area specifying unit 104, and a foreground / background separating unit 105.
- the object extraction unit 101 roughly extracts an image object corresponding to the foreground object included in the input image, and extracts the extracted image object from the motion detection unit.
- the object extracting unit 101 roughly extracts an image object corresponding to a foreground object by detecting a contour of the image object corresponding to the foreground object included in the input image.
- the object extraction unit 101 roughly extracts the image object corresponding to the background object included in the input image, and extracts the extracted image object from the motion detection unit. Feed to 102.
- the object extracting unit 101 roughly extracts, for example, an image object corresponding to the background object from a difference between the input image and the extracted image object corresponding to the extracted foreground object. .
- the object extracting unit 101 determines the image object corresponding to the foreground object from the difference between the background image stored in the background memory provided therein and the input image.
- the image object corresponding to the background object may be roughly extracted.
- the motion detection unit 102 calculates the motion vector of the image object corresponding to the coarsely extracted foreground object by using, for example, a block matching method, a gradient method, a phase correlation method, and a perrecursive method. Then, the calculated motion vector and the position information of the motion vector (information for specifying the position of the pixel corresponding to the motion vector) are supplied to the mixture ratio calculation unit 103 and the motion blur adjustment unit 106. I do.
- the motion vector output by the motion detection unit 102 includes information corresponding to the motion amount V.
- the motion detection unit 102 may output the motion vector for each image object to the motion blur adjustment unit 106 together with the pixel position information for specifying the pixel in the image object. .
- the motion amount V is a value representing a change in the position of the image corresponding to the moving object in units of pixel intervals. For example, when the image of the object corresponding to the foreground is moving so that it is displayed at a position separated by 4 pixels in the next frame with respect to a certain frame, the motion amount of the image of the object corresponding to the foreground V is
- the object extraction unit 101 and the motion detection unit 102 are necessary when adjusting the amount of motion blur corresponding to a moving object.
- the mixture ratio calculation unit 103 Based on the input image, the motion vector supplied from the motion detection unit 102, and the positional information, the mixture ratio calculation unit 103 sets the pixels in the covered background area, which is one of the mixed areas. And the estimated mixture ratio (hereinafter, the mixture ratio) ), And the corresponding mixture ratio relationship information is generated, and the pixel is assumed to belong to the uncovered background region, which is another one of the mixture regions. Then, the estimated mixture ratio and the mixture ratio relation information corresponding thereto are generated, and the generated two estimated mixture ratios and the mixture ratio relation information corresponding thereto are supplied to the region identification unit 104.
- the region identification unit 104 the region identification unit 104.
- the mixing ratio or is a value indicating a ratio of a component of an image corresponding to a background object (hereinafter, also referred to as a background component) in a pixel value, as shown in Expression (3) described below.
- the region specifying unit 104 determines each of the pixels of the input image based on the input image and the two estimated mixture ratios supplied from the mixture ratio calculation unit 103 and the mixture ratio relationship information corresponding thereto. , Foreground area, background area, or mixed area, and indicates whether each pixel belongs to the foreground area, background area, or mixed area (hereinafter referred to as area information). This is supplied to the section 105 and the motion blur adjusting section 106.
- the region identification unit 104 generates a mixture ratio based on the generated region information, the two estimated mixture ratios supplied from the mixture ratio calculation unit 103, and the mixture ratio relationship information corresponding to them.
- the obtained mixture ratio H is supplied to the foreground / background separation unit 105.
- the foreground / background separation section 105 only includes components of the image corresponding to the foreground object (hereinafter also referred to as foreground components).
- the input image is separated into a foreground component image composed of the following components and a background component image composed of only the background component, and the foreground component image is supplied to the motion blur adjustment unit 106 and the selection unit 107. It is also conceivable to use the separated foreground component image as the final output. It is possible to specify only the foreground and the background without considering the conventional mixed area, and obtain a more accurate foreground and background compared to the separated method.
- the motion-blur adjusting unit 106 includes a motion amount V and area information that can be determined from the motion vector. Based on, a processing unit indicating one or more pixels included in the foreground component image is determined. The processing unit is data that specifies a group of pixels to be processed for adjusting the amount of motion blur.
- the motion blur adjustment unit 106 includes a motion blur adjustment amount input to the image processing apparatus, a foreground component image supplied from the foreground / background separation unit 105, and a motion vector supplied from the motion detection unit 102.
- the motion blur included in the foreground component image, reducing the amount of motion blur, or increasing the amount of motion blur in the foreground component image based on the The amount of blur is adjusted, and the foreground component image in which the amount of motion blur is adjusted is output to the selector 107.
- the motion vector and its position information may not be used.
- the motion blur refers to a distortion included in an image corresponding to a moving object, which is caused by the movement of the object in the real world to be imaged and the characteristics of the sensor imaging. .
- the selection unit 107 includes, for example, a foreground component image supplied from the foreground / background separation unit 105 and a motion supplied from the motion blur adjustment unit 106 based on a selection signal corresponding to the user's selection.
- a foreground component image supplied from the foreground / background separation unit 105 and a motion supplied from the motion blur adjustment unit 106 based on a selection signal corresponding to the user's selection.
- One of the foreground component images whose blur amount has been adjusted is selected, and the selected foreground component image is output.
- FIG. 3 is a diagram illustrating imaging by a sensor.
- the sensor includes, for example, a CCD video camera equipped with a CCD (Charge-Coupled Device) area sensor that is a solid-state imaging device.
- the object 111 corresponding to the foreground in the real world moves horizontally between the sensor 112 corresponding to the background in the real world and the sensor, for example, from left to right in the figure.
- CCD Charge-Coupled Device
- the sensor captures the object 111 corresponding to the foreground together with the object 111 corresponding to the background.
- Sensor dew The light time can be 130 seconds.
- the exposure time is the period from when the sensor starts converting the input light into electric charge until the sensor finishes converting the input light into electric charge.
- the exposure time is also referred to as shirt time.
- FIG. 4 is a diagram illustrating the arrangement of pixels.
- a to I indicate individual pixels.
- the pixels are arranged on a plane corresponding to the image.
- One detection element corresponding to one pixel is arranged on the sensor.
- one detection element When the sensor captures an image, one detection element outputs a pixel value corresponding to one pixel forming the image.
- the position of the detection element in the X direction corresponds to the position in the horizontal direction on the image
- the position of the detection element in the Y direction corresponds to the position in the vertical direction on the image.
- a detection element such as a CCD converts input light into electric charges and accumulates the converted electric charges for a period corresponding to the shutter time.
- the amount of charge is almost proportional to the intensity of the input light and the time the light is input.
- the detection element adds the electric charge converted from the input light to the electric charge already stored during a period corresponding to the shutter time. That is, the detection element integrates the input light for a period corresponding to the shutter time, and accumulates an amount of charge corresponding to the integrated light. It can be said that the detection element has an integration effect with respect to time.
- the electric charge accumulated in the detection element is converted into a voltage value by a circuit (not shown), and the voltage value is further converted into a pixel value such as digital data and output. Therefore, the individual pixel values output from the sensor are the values projected into a one-dimensional space, which is the result of integrating the spatially extended part of the object corresponding to the foreground or background over the shirt time. Having.
- the image processing apparatus extracts significant information buried in the output signal, for example, the mixture ratio H by the accumulation operation of the sensor.
- the image processing device adjusts the amount of distortion caused by mixing of the foreground image objects themselves, for example, the amount of motion blur.
- the image processing device adjusts the amount of distortion caused by mixing the foreground image object and the background image object.
- FIG. 6A shows an object corresponding to a moving foreground and a stationary background.
- FIG. 9 is a diagram illustrating an image obtained by imaging a corresponding object.
- FIG. 6B is a diagram illustrating a model of an image obtained by capturing an object corresponding to a moving foreground and an object corresponding to a stationary background.
- FIG. 6A shows an image obtained by capturing an object corresponding to a moving foreground and an object corresponding to a stationary background.
- the object corresponding to the foreground moves horizontally from left to right with respect to the screen.
- FIG. 6B is a model diagram in which pixel values corresponding to one line of the image shown in FIG. 6A are expanded in the time direction.
- the horizontal direction in Fig. 6B corresponds to the spatial direction X in Fig. 6A.
- the pixels in the background area are the background components, that is, the pixel values from only the components of the image corresponding to the background object. Is configured.
- the pixel values of the pixels in the foreground area are composed of only the components of the foreground, that is, the components of the image corresponding to the foreground object.
- the pixel value of a pixel in the mixed area is composed of a background component and a foreground component. Since the pixel value is composed of the background component and the foreground component, the mixed region can be said to be a distorted region.
- the mixed area is further classified into a covered background area and an covered background area.
- the covered background area is a mixed area at a position corresponding to the front end of the foreground object in the traveling direction with respect to the foreground area, and is an area where the background component is covered by the foreground with the passage of time. .
- the uncovered background area is a mixed area at the position corresponding to the rear end of the foreground object in the traveling direction with respect to the foreground area. The area that appears.
- the image including the foreground area, the background area, or the covered background area or the uncovered background area is divided into the area specifying unit 104, the mixture ratio calculating unit 103, and the foreground / background separating unit 1.
- 05 is input as an input image.
- Figure 7 shows the background area, foreground area, mixed area, covered backdall
- FIG. 3 is a diagram illustrating a window area and an uncovered background area.
- the background area is a stationary part
- the foreground area is a moving part
- the covered background area of the mixed area is a part that changes from the background to the foreground
- the uncovered background area of the mixed area is the part that changes from the foreground to the background.
- Fig. 8 shows the timewise expansion of the pixel values of the pixels that are arranged in a row and adjacent to each other in the images of the object corresponding to the stationary foreground and the object corresponding to the stationary background. It is a model figure. For example, it is possible to select pixels that are arranged on one line of the screen as pixels that are adjacently arranged in one column.
- the pixel values of F01 to F04 shown in FIG. 8 are the pixel values of the pixels corresponding to the stationary foreground object.
- the pixel values B01 to B04 shown in FIG. 8 are the pixel values of the pixels corresponding to the stationary background object.
- the vertical direction in FIG. 8 corresponds to time, and time elapses from top to bottom in the figure.
- the position on the upper side of the rectangle in FIG. 8 corresponds to the time when the sensor starts converting the input light into electric charge, and the position on the lower side of the rectangle in FIG. 8 indicates the electric power of the light input to the sensor.
- the horizontal direction in FIG. 8 corresponds to the spatial direction X described in FIG. 6A. More specifically, in the example shown in FIG. 8, the distance from the left side of the rectangle indicated by “F01” in FIG. 8 to the right side of the rectangle indicated by “B04” is eight times the pixel pitch. That is, it corresponds to the interval between eight consecutive pixels.
- the light input to the sensor does not change during the period corresponding to the shutter time.
- the period corresponding to the shirt time is divided into two or more periods of the same length,
- the model diagram shown in FIG. 8 can be represented as a model shown in FIG.
- the number of virtual divisions is set according to the amount of movement V of the object corresponding to the foreground within the shutter time.
- the number of virtual divisions is set to 4 corresponding to the motion amount V of 4, and the period corresponding to the shutter time is divided into four.
- the top row in the figure corresponds to the first, divided period since the shirt opened.
- the second row from the top in the figure corresponds to the second, divided period when the shirt is open.
- the third row from the top in the figure corresponds to the third, divided period since the shirt opened.
- the fourth row from the top in the figure corresponds to the fourth, divided period since the shirt opened.
- the shirt time divided according to the movement amount V is also referred to as shirt time / V.
- the foreground component FO l / v is equal to the value obtained by dividing the pixel value F01 by the number of virtual divisions.
- the foreground component FO l / v is equal to the value obtained by dividing the pixel value F01 by the number of virtual divisions.
- F02 / v is equal to the pixel value F02 divided by the number of virtual divisions
- the foreground component F03 / v is equal to the pixel value F03 divided by the number of virtual divisions
- the foreground component F04 / v is the pixel It is equal to the value F04 divided by the number of virtual partitions.
- the background component BO l / v is equal to the value obtained by dividing the pixel value B01 by the number of virtual divisions.
- the background component B02 / v is equal to a value obtained by dividing the pixel value B02 by the virtual division number
- B03 / v is a pixel value B03 obtained by dividing the pixel value B03 by the virtual division number.
- B04 / v is equal to the pixel value B04 divided by the number of virtual divisions.
- the light corresponding to the background object input to the sensor does not change during the period corresponding to the shirt time, so the first The background component BO l / v corresponding to V, and the background component BO l / v corresponding to the second opening of the shirt, and the third background component BO l / v corresponding to the shirt time / V
- the corresponding background component BO l / v has the same value as the background component BO l / v corresponding to the fourth shutter opening time / V when the shutter is opened.
- B02 / v to B04 / v also have the same relationship.
- Figure 10 is a model diagram in which the pixel values of the pixels on one line, including the covered background area, are expanded in the time direction when the object corresponding to the foreground moves toward the right in the figure. is there.
- the foreground motion amount V is 4. Since one frame is short, it can be assumed that the object corresponding to the foreground is rigid and moves at a constant speed.
- the image of the object corresponding to the foreground moves so as to be displayed four pixels to the right in the next frame with respect to a certain frame.
- the leftmost pixel to the fourth pixel from the left belong to the foreground area.
- the fifth through seventh pixels from the left belong to the mixed area that is the covered background area.
- the rightmost pixel belongs to the background area.
- the image belonging to the covered background area At some point during the period corresponding to the shirt time, the component contained in the elementary pixel value changes from the background component to the foreground component.
- a pixel value M indicated by a thick line frame in FIG. 10 is represented by Expression (1).
- the fifth pixel from the left contains the background component corresponding to one shirt time / V and the foreground component corresponding to three shirt times / V, so the fifth pixel from the left is mixed The ratio is 1/4.
- the sixth pixel from the left contains the background component corresponding to the two shirt times / v and the foreground component corresponding to the two shirt times / v, so the mixture ratio of the sixth pixel from the left Is 1/2.
- the seventh pixel from the left contains the background component corresponding to the three shutter times / V and the foreground component corresponding to one shutter time / V, so the mixture ratio of the seventh pixel from the left is , 3/4.
- the object corresponding to the foreground is a rigid body and the foreground image moves at a constant speed so that it is displayed four pixels to the right in the next frame, for example, from the left in Fig. 10
- the foreground component F07 / v of the first pixel the shirt is open at the first time
- the fifth pixel from the left in Figure 10 is the second at the time of the shirt at the open time.
- the foreground component F07 / v is the foreground component of the sixth pixel from the left in Fig. 10 corresponding to the third shirt time / v when the shirt is open, and the left in Fig. 10 From the seventh pixel, the shirt is open and the fourth is equal to the foreground component corresponding to the shirt time / V, 0
- the object corresponding to the foreground is a rigid body and the foreground image moves at a constant speed so that it is displayed four pixels to the right in the next frame, for example, from the left in Fig. 10
- the foreground component F06 / v of the 1st pixel, the shutter opening / first shutter time / V corresponds to the 4th pixel from the left in Fig. 10, the second shutter time / v when the shutter is open Equal to the foreground component.
- the foreground component F06 / v is the fifth shirt from the left in Figure 10 Of the foreground component corresponding to the shutter time / v and the foreground component of the sixth pixel from the left in Fig. 10 corresponding to the fourth shutter time / v when the shirt is opened. ,.
- the object corresponding to the foreground is a rigid body and the foreground image moves at a constant speed so that it is displayed four pixels to the right in the next frame, for example, from the left in Fig. 10
- the foreground component F05 / v is the foreground component of the fourth pixel from the left in FIG. 10 corresponding to the third shirt time / V when the shirt is opened, and the left in FIG.
- the foreground component F04 / v of the pixel in which the shirt is open and the first shirtta time / V, is the foreground corresponding to the second pixel from the left in Fig. 10, in which the shirt is open and the second shirtta time / V Is equal to the component of Similarly, the foreground component F04 / v is the foreground component of the third pixel from the left in Fig. 10, which corresponds to the third shutter time / v when the shutter is opened, and the left pixel in Fig. 10.
- the fourth pixel is equal to the foreground component corresponding to the fourth shirt time / V when the shirt is open. Since the foreground area corresponding to the moving object includes the motion blur as described above, it can be said that the area is a distortion area.
- FIG. 11 is a model diagram in which pixel values of pixels on one line, including the uncovered background area, are developed in the time direction when the foreground moves toward the right side in the figure.
- the motion amount V of the foreground is 4. Since one frame is short, it can be assumed that the object corresponding to the foreground is rigid and moves at a constant speed.
- the image of the object corresponding to the foreground The image moves to the right by four pixels in the next frame with reference to a certain frame.
- the leftmost pixel to the fourth pixel from the left belong to the background area.
- the fifth to seventh pixels from the left belong to the mixed area that is the uncovered background.
- the rightmost pixel belongs to the foreground area.
- the pixels belonging to the uncovered background area At some point during the period that corresponds to the shirt time, the components included in the values change from the foreground components to the background components.
- a pixel value M ′ with a thick frame in FIG. 11 is represented by Expression (2).
- the fifth pixel from the left contains the background component corresponding to three shirt times / V and the foreground component corresponding to one shirt time / V, so the fifth pixel from the left is mixed The ratio is 3/4.
- the sixth pixel from the left contains the background component corresponding to the two shirt times / V and the foreground component corresponding to the two shirt times / v, so the mixture ratio of the sixth pixel from the left Is 1/2.
- the seventh pixel from the left contains the background component corresponding to one shutter time / V and the foreground component corresponding to three shutter times / V, so the mixture ratio of the seventh pixel from the left a Is 1/4.
- h is the mixing ratio.
- B is the pixel value of the background, and Fi / v is the foreground component.
- the object corresponding to the foreground is a rigid body and moves at a constant speed, and the movement amount V is 4, for example, the shirt pixel of the fifth pixel from the left in FIG. Is the first
- the foreground component FO l / v of shirt time / V is the foreground component of the sixth pixel from the left in Fig. 11, corresponding to the second shutter time / V when the shutter is opened. Equal to minutes.
- FO l / v is the foreground component of the seventh pixel from the left in Fig. 11 corresponding to the third shutter time / V when the shutter is open
- the eighth pixel from the left in Fig. 11 The pixel of the foreground component corresponding to the fourth shirt time / V when the shirt is open is equal to the respective pixels.
- the first foreground component F02 / v of the shirt time / V is equal to the foreground component of the seventh pixel from the left in Fig. 11 corresponding to the second shirt time / V when the shirt opens.
- the foreground component F02 / v is equal to the foreground component of the eighth pixel from the left in FIG. 11 corresponding to the third shutter time / V after the shutter is opened.
- the first foreground component F03 / v of the shirt time / V is equal to the foreground component of the eighth pixel from the left in FIG. 11 corresponding to the second shutter time / V when the shutter is opened.
- the number of virtual divisions has been described as four, but the number of virtual divisions corresponds to the amount of motion V.
- the motion amount V generally corresponds to the moving speed of the object corresponding to the foreground. For example, when the object corresponding to the foreground is moving so as to be displayed four pixels to the right in the next frame with respect to a certain frame, the motion amount V is set to 4.
- the number of virtual divisions is four.
- the motion amount V is set to 6, and the virtual division is performed. The number is six.
- Figures 12 and 13 show the mixed area consisting of the foreground area, background area, covered background area or anchored background area described above. And the relationship between the foreground component and the background component corresponding to the divided shutter time.
- Figure 12 shows an example of extracting pixels in the foreground, background, and mixed regions from an image containing the foreground corresponding to an object moving in front of a stationary background.
- the object indicated by A and corresponding to the foreground moves horizontally with respect to the screen.
- Frame # n + l is the next frame after frame #n
- frame # n + 2 is the next frame after frame # n + l.
- the pixel value in the foreground area is composed of four different foreground components corresponding to the time / V period, since the object corresponding to the foreground moves.
- the leftmost pixel of the pixels in the foreground area shown in FIG. 13 is
- Consists of FO l / v, F02 / v, F03 / v, and F04 / V that is, pixels in the foreground area include motion blur.
- the pixel value in the background area does not include motion blur.
- the pixel value of a pixel belonging to a mixed area composed of a covered background area or an covered background area is composed of a foreground component and a background component.
- FIG. 4 is a model diagram developed in the time direction.
- Frame #n is the next frame after frame ⁇ -l
- frame ⁇ + l is the next frame after frame # ⁇ .
- Other frames are similarly referred to.
- the pixel values B01 to B12 shown in FIG. 14 are the pixel values of the pixels corresponding to the stationary background object. Since the object corresponding to the background is stationary, the pixel value of the corresponding pixel in frame # ⁇ -1 to frame ⁇ + 1 does not change. For example, the pixel in frame # ⁇ and the pixel in frame # ⁇ + 1 corresponding to the position of the pixel having the pixel value of # 05 in frame # ⁇ -l each have the pixel line of # 05.
- Figure 15 shows the pixels in a row adjacent to one another in the three frames of the image of the object corresponding to the foreground moving to the right in the figure together with the object corresponding to the stationary background.
- FIG. 9 is a model diagram in which pixel values of pixels at the same position on a frame are developed in the time direction. The model shown in FIG. 15 includes a covered background region.
- Fig. 15 it can be assumed that the object corresponding to the foreground is rigid and moves at a constant speed, and the foreground image moves so that it is displayed four pixels to the right in the next frame.
- the motion amount V is 4, and the number of virtual divisions is 4.
- the foreground component of the leftmost pixel of frame ⁇ -1 in Fig. 15 at the time when the shutter opens and the first shutter time / V is F12 / v
- the second pixel from the left in Fig. 15 The foreground component of the pixel's second shirt time / V when the shirt is open is also F12 / v.
- the fourth pixel from the left in Figure 15 is the fourth pixel from the left pixel
- the foreground component of the shirt time / V is F12 / v.
- the foreground component of the second shutter time / V of the leftmost pixel of frame-1 in Fig. 15-1 is Fl l / v
- the second pixel from the left in Fig. 15 is also Fl l / v
- the foreground component of the third pixel from the left in Fig. 15 at the fourth shutter time / V when the shutter is open is Fl l / v.
- the foreground component of the leftmost pixel of frame ⁇ -1 in Fig. 15 at the third shutter time / V after the shutter is opened is FlO / v
- the second pixel from the left in Fig. 15 is also FlO / v
- the foreground component of the fourth shirt time / V when the shirt is opened is also FlO / v.
- the foreground component of the leftmost pixel of frame ttn-1 in FIG. 15 at the fourth shutter time / V from when the shutter has opened is F09 / v.
- the second pixel from the left of frame #nl in Fig. 15 It becomes v.
- the background components of the third pixel from the left of frame # n-l in Fig. 15 at the first and second shutter / V when shutter is opened are B02 / v.
- the background component of the fourth pixel from the left of the frame # n-l in FIG. 15 from the first to third shirt time / V from when the shirt is opened is B03 / V.
- the leftmost pixel belongs to the foreground area, and the second to fourth pixels from the left side belong to the mixed area which is the covered background area.
- the fifth through 12th pixels from the left of frame # n-1 in FIG. 15 belong to the background area, and their pixel values are B04 through B11, respectively.
- the first to fifth pixels from the left of frame ⁇ in FIG. 15 belong to the foreground area.
- the foreground component of the shirt time / v in the foreground area of frame # ⁇ is one of F05 / v to F12 / v.
- the object corresponding to the foreground is a rigid body and moves at a constant speed, and the image of the foreground moves so that it is displayed four pixels to the right in the next frame.
- the foreground component of the first shutter time / V of the fifth pixel from the left of the pixel is F12 / v
- the first pixel of the sixth pixel from the left in Figure 15 The foreground component of the second shirt time / V is also F12 / v.
- the time / V foreground component is F12 / v.
- the foreground component of the fifth pixel from the left of frame #n in Fig. 15 in which the shirt is open and the second shutter is opened is Fl l / v
- the sixth pixel from the left in Fig. 15 is The foreground component of the pixel, which is the third shirt time / V when the shirt is opened, is also Fl l / v.
- the foreground component of the seventh pixel from the left in Fig. 15 at the fourth shutter time / v when the shutter is open is Fl l / v.
- the foreground component of the 5th pixel from the left of frame #n in Fig. 15 at the third shutter opening time / V in the foreground is FlO / v
- the sixth pixel from the left in Fig. 15 The foreground component of the fourth shirt time / V, which is the same as when the shirt opens, is also FlO / v.
- the foreground component of the fifth pixel from the left of frame #n in Fig. 15 at the fourth shutter time / V after the shutter is opened is F09 / v.
- the background component of the sixth pixel from the left in the frame in Fig. 15 at the first shutter opening time / V when the shutter is open is B05 / v.
- the background components of the seventh and the seventh pixels from the left of the frame ttn in Fig. 15 from the left, and the first and second shirt time / V from when the shirt is opened are B06 / v.
- the background components of the eighth to the left pixels of the frame ⁇ in FIG. 15 from the left, the first to the third, and the background of the shirt time / V from the opening of the shirt are ⁇ 07 / ⁇ .
- the ninth to 12th pixels from the left of frame # ⁇ in FIG. 15 belong to the background area, and the pixel values are # 08 to B11, respectively.
- the first through ninth pixels from the left of frame ⁇ + 1 in FIG. 15 belong to the foreground area.
- the foreground component of the shirt time / V in the foreground area of the frame ⁇ + 1 is any of FO l / v to F 12 / v.
- the object corresponding to the foreground is rigid and moves at a constant speed. Since the scenery image moves so that it is displayed four pixels to the right in the next frame, the shutter is opened and the first shutter time / V of the ninth pixel from the left in frame # n + l in Figure 15
- the foreground component of the foreground is F12 / v
- the foreground component of the 10th pixel from the left in Fig. 15 at the second shirt time / V after the shirt is opened is also F12 / v.
- the foreground component of the fourth shirt time / V is F12 / v.
- the foreground component of the ninth pixel from the left of frame itn + 1 in Fig. 15 during the second shutter release time / V from when the shutter is opened is Fl l / v, which is 1 from the left in Fig. 15.
- the 0th pixel, the shirt is open and the 3rd shirt is the time / V foreground component
- the foreground component of the ninth pixel from the left of frame # n + 1 in Fig. 15 corresponding to the fourth shutter time / V after the shutter is opened is F09 / v.
- the background component of the 10th pixel from the left of frame ⁇ + 1 in Fig. 15 in Fig. 15 at the time of the first shirt time / V when the shirt is opened is V09 / ⁇ .
- the background components of the 11th pixel from the left of frame ttn + 1 in Fig. 15 corresponding to the first and second portions of the shutter time / V from when the shutter has opened are BIO / v.
- the background components of the first to third pixels from the left of the first pixel to the third pixel, and the background of the shirt time / V are Bl l / v.
- FIG. 16 is a model diagram of an image in which foreground components are extracted from the pixel values shown in FIG. Figure 17 shows the pixels that are adjacent to each other and are arranged in a single row in the three frames of the image that captures the foreground corresponding to the object moving to the right in the figure together with the stationary background.
- FIG. 5 is a model diagram in which pixel values of pixels at the same position are developed in the time direction. In FIG. 17, an uncovered background area is included. In Fig. 17, it can be assumed that the object corresponding to the foreground is a rigid body and moves at a constant speed.
- the motion amount V is 4, for example, the most significant amount of frame #nl in Figure 17
- the foreground component of the left pixel which is the first pixel when the shirt is open, is F13 / v
- the second pixel from the left in Figure 17 is the second pixel when the shirt is open.
- the time / V foreground component is also F13 / V.
- the fourth pixel from the left in Fig. 17 is the fourth pixel from the open pixel.
- the foreground component of the shirt time / V is F13 / v.
- the foreground component of the second pixel from the left of frame ttn-1 in Figure 17 at the first shutter release time / V after shutter release is F14 / v
- the third image from the left in Figure 17 The foreground component of the second shirt time / V, which is the first time that the shirt is open, is also F14 / v.
- the background component of the leftmost pixel of the leftmost pixel of frame # n-l in Fig. 17 and the second to fourth, and the background components of the shirtdown time / V of the shirtdown time are: , B25 / v.
- the background component of the second pixel from the left of frame # n-1 in FIG. 17, the third and fourth shutter shirt time / V after the shutter is opened, is B26 / v.
- the background component of the third pixel from the left of frame # n-l in Fig. 17 at the fourth shutter time / V after the shutter is opened is B27 / v.
- the leftmost pixel to the third pixel belong to the mixed area which is the covered background area.
- the fourth through 12th pixels from the left of frame ttn-1 in Fig. 17 are the foreground Belongs to the area.
- the foreground component of the frame is one of F13 / v to F24 / v.
- the leftmost pixel to the fourth pixel from the left of frame ⁇ in FIG. 17 belong to the background area, and the pixel values are ⁇ 25 to ⁇ 28, respectively.
- the object corresponding to the foreground is rigid and moves at a constant speed, and the image of the foreground moves so that it is displayed four pixels to the right in the next frame, so the left of the frame in Figure 17
- the foreground component of the fifth pixel from the shutter for the first shutter time / V after the shutter is opened is F13 / v
- the sixth pixel from the left in Fig. 17 is the second for the shutter open.
- the foreground component of the shirt time / V is also F13 / V.
- the foreground component of the seventh pixel from the left in Figure 17, the third pixel at the time of opening, and the foreground component of the seventh pixel from the left in Figure 17, the fourth pixel of the pixel at the eighth pixel from the left in Figure 17 The foreground component of the shirt time / V is F13 / v.
- the foreground component of the 6th pixel from the left of frame 7 ⁇ in Fig. 17 at the first shutter time / V when the shutter is open is F14 / v
- the 7th pixel from the left in Fig. 17 the foreground component of the shutter has opened is the second shutter time / V
- shutter is the first shutter time / V open
- the foreground component is F15 / V.
- the background component of the fifth pixel from the left of the frame in Figure 17 at the 2nd to 4th shutter time / V from when the shutter has opened is ⁇ 29 / ⁇ .
- the background component of the third and fourth shirt time / V when the shirt is open is B30 / v.
- the background component of the seventh pixel from the left of the frame in Fig. 17 corresponding to the fourth shutter time / V from when the shutter is opened is B31 / v.
- the eighth to 12th pixels from the left of frame #n in FIG. 17 belong to the foreground area.
- the value corresponding to the period of the shirt time / V in the foreground area of the frame ftn is any one of F13 / v to F20 / v.
- the leftmost pixel to the eighth pixel from the left of frame ⁇ + 1 in FIG. 17 belong to the background area, and their pixel values are ⁇ 25 to ⁇ 32, respectively.
- the object corresponding to the foreground is rigid and moves at a constant speed.
- the image of the foreground moves so that it is displayed four pixels to the right in the next frame, so the frame ttn + 1 in Fig. 17
- the foreground component of the first shutter time / V at the ninth pixel from the left of the pixel is F13 / V
- the shirt at the 10th pixel from the left in Figure 17 is open.
- the foreground component of the second Zeta time / V is also F13 / v.
- the foreground component of the fourth shirt time / V is F13 / v.
- the foreground component of the 10th pixel from the left of frame # n + l in Fig. 17 at the first shutter release time / V after the shutter is opened is F14 / v, which is 1 1 from the left in Fig. 17
- the foreground component of the second pixel, at the time when the shutter is opened and at the second shutter time / V, is also F14 / v.
- the foreground component of the second pixel from the left in Fig. 17 at the first shutter button time / V when the shutter button opens is F15 / v.
- the components of the ninth pixel from the left of frame # n + l in Fig. 17 and the background components of the 2nd to 4th, shirt time / v Becomes B33 / v.
- the background component of the 10th pixel from the left of frame # n + l in Fig. 17 corresponding to the 3rd and 4th shutter time / V from when the shutter has opened is B34 / v.
- the background component of the 11th pixel from the left of frame ttn + 1 in FIG. 17 corresponding to the fourth portion of the shutter time / V from when the shutter has opened is 35 / ⁇ .
- the 12th pixel from the left of frame ⁇ + l in FIG. 17 belongs to the foreground area.
- the foreground component of the shirt time / V in the foreground region of frame + 1 is any one of F13 / v to F16 / v.
- FIG. 18 is a model diagram of an image in which foreground components are extracted from the pixel values shown in FIG. You.
- the mixture ratio calculation unit 103 determines that the pixel belongs to the covered background area. Generates the estimated mixture ratio and the corresponding mixture ratio relationship information under the assumption, and the estimated mixture ratio, which is the estimated mixture ratio when the pixel belongs to the uncovered background area, and Generates mixing ratio relation information corresponding to.
- the mixture ratio calculation unit 103 supplies the generated two estimated mixture ratios and the mixture ratio relation information corresponding thereto to the region identification unit 104.
- the region specifying unit 104 determines each of the pixels of the input image based on the input image and the two estimated mixture ratios supplied from the mixture ratio calculation unit 103 and the mixture ratio relationship information corresponding thereto. , Foreground area, background area, or mixed area, and for each pixel the area information indicating whether the pixel belongs to the foreground area, background area, or mixed area. It is supplied to the blur adjusting unit 106.
- the foreground / background separation unit 105 extracts a foreground component image consisting of only foreground components based on the pixel values of a plurality of frames, area information, and the mixture ratio a, and supplies the extracted foreground component image to the motion blur adjustment unit 106. I do.
- the motion blur adjusting unit 106 is supplied from the foreground component image supplied from the foreground / background separating unit 105, the motion vector supplied from the motion detecting unit 102, and the region identifying unit 104. Based on the obtained region information, the amount of motion blur included in the foreground component image is adjusted, and a foreground component image in which the amount of motion blur is adjusted is output.
- step SI1 the mixture ratio calculation unit 103 determines that the pixel belongs to the covered background area based on the input image, the motion vector supplied from the motion detection unit 102, and its position information. Assuming that the estimated mixture ratio, which is the estimated mixture ratio, and the corresponding mixture ratio relation information are calculated, and that, assuming that the pixel belongs to the uncovered background area, the estimated mixture ratio, which is the estimated mixture ratio, The mixture ratio and the mixture ratio relation information corresponding thereto are calculated. Mixing ratio calculation The output unit 103 supplies the calculated two estimated mixture ratios and the mixture ratio relationship information corresponding thereto to the region identification unit 104. The details of the mixture ratio calculation process will be described later.
- the region identification unit 104 determines the input image based on the input image, the two estimated mixture ratios supplied from the mixture ratio calculation unit 103, and the mixture ratio relationship information corresponding thereto. For each pixel of, an area specifying process for generating area information indicating whether the pixel belongs to the foreground area, the background area, the covered background area, or the uncovered background area is executed. The details of the area specifying process will be described later.
- the region identification unit 104 generates a mixture ratio a based on the generated region information, the two estimated mixture ratios supplied from the mixture ratio calculation unit 103, and the mixture ratio relationship information corresponding thereto, The generated mixture ratio is supplied to the foreground / background separation unit 105.
- step S13 the foreground / background separation unit 105 extracts a foreground component from the input image based on the region information and the mixture ratio H, and supplies the foreground component image to the motion blur adjustment unit 106. I do.
- step S14 the motion-blur adjusting unit 106 determines, based on the motion vector and the area information, continuous pixels arranged in the motion direction, the uncovered background area, the foreground area, and the covered background area.
- a processing unit indicating the position of an image belonging to one of the regions on the image is generated, and the amount of motion blur included in the foreground component corresponding to the processing unit is adjusted. The details of the processing for adjusting the amount of motion blur will be described later.
- step S15 the image processing apparatus determines whether or not processing has been completed for the entire screen. If it is determined that processing has not been completed for the entire screen, the process proceeds to step S14, where the processing unit is determined. The process of adjusting the amount of motion blur for the foreground component corresponding to is repeated.
- step S15 If it is determined in step S15 that the processing has been completed for the entire screen, the processing ends.
- the image processing apparatus separates the foreground from the background, and The amount of ke can be adjusted. That is, the image processing apparatus can adjust the amount of motion blur included in the sample data that is the pixel value of the foreground pixel.
- FIG. 20 is a block diagram showing the configuration of the mixture ratio calculation unit 103.
- the estimated mixture ratio processing unit 201 executes an operation corresponding to the model of the covered background area.
- the estimated mixture ratio is calculated for each pixel, and the estimated mixture ratio is output together with the mixture ratio relationship information calculated with the calculation of the estimated mixture ratio.
- the mixture ratio relationship information output from the estimated mixture ratio processing unit 201 is, for example, a sum of foreground components.
- the estimated mixture ratio processing unit 202 Based on the motion vector and its position information supplied from the motion detection unit 102 and the input image, the estimated mixture ratio processing unit 202 performs an arithmetic operation corresponding to the model of the uncovered background area. The estimated mixture ratio is calculated for each pixel, and the estimated mixture ratio is output together with the mixture ratio relationship information calculated with the calculation of the estimated mixture ratio.
- the mixture ratio relationship information output by the estimated mixture ratio processing unit 202 is, for example, a sum of foreground components.
- the mixing ratio of the pixels belonging to the mixed region has the following properties. That is, the mixture ratio changes linearly in accordance with the change in the pixel position. If the change in pixel position is one-dimensional, the change in the mixture ratio can be represented by a straight line.If the change in pixel position is two-dimensional, the change in the mixture ratio can be represented by a plane. It is a thing.
- the slope of the mixture ratio is the inverse ratio of the amount of movement V in the foreground within the shutter time.
- Figure 21 shows an example of an ideal mixing ratio.
- the slope 1 can be expressed as the reciprocal of the motion amount V.
- the ideal mixture ratio H has a value of 1 in the background region, a value of 0 in the foreground region, and a value exceeding 0 and less than 1 in the mixture region.
- the pixel value C06 of the seventh pixel from the left of frame ttn can be expressed by equation (4) using the pixel value P06 of the seventh pixel from the left of frame #nl. it can.
- the pixel value C06 is expressed as the pixel value M of the pixel in the mixed area
- the pixel value P06 is expressed as the pixel value B of the pixel in the background area. That is, the pixel value M of the pixel in the mixed area and the pixel value B of the pixel in the background area can be expressed as Expressions (5) and (6), respectively.
- Equation (4) corresponds to the mixing ratio. Since the motion amount V is 4, the mixture ratio of the seventh pixel from the left of frame #n is 0.5.
- Equation (3) which represents hi, can be rewritten as equation (7).
- Equation (7) is the sum of the foreground components contained in the pixel of interest ⁇ iFi / v.
- the variables included in equation (7) are the mixture ratio H and the sum f of the foreground components.
- Fig. 23 shows a model in which pixel values are expanded in the time direction in which the amount of motion V is 4 and the number of virtual divisions in the time direction is 4 in the uncovered background area.
- the pixel value C of the frame #n of interest is regarded as the pixel value of the mixed area, and the frame # after the frame #n, as in the expression in the covered background area described above.
- Equation (3) showing the mixture ratio H can be expressed as Equation (8).
- the background object has been described as being stationary, even when the background object is moving, the pixel value of the pixel at the position corresponding to the background motion amount V can be used. Equations (4) to (8) can be applied. For example, in FIG. 22, when the motion amount V of the object corresponding to the background is 2 and the number of virtual divisions is 2, when the object corresponding to the background is moving to the right in the figure, the equation (6) The pixel value B of the pixel in the background area in) is the pixel value P04.
- Equations (7) and (8) each contain two variables, so the mixture ratio cannot be determined as it is.
- the motion vector supplied from the motion detection unit 102 and its position information are used.
- equation (9) holds for ⁇ 02 of frame ⁇ -1
- equation (10) holds for C06 of frame ⁇ .
- the motion of the foreground object V It is possible to select a plurality of sets of pixels belonging to a mixed region having the same sum of the mixture ratio and the foreground component, and pixels belonging to a corresponding background region. For example, a plurality of sets of pixels belonging to the mixed area and pixels belonging to the corresponding background area can be set to five sets.
- the pixels Mtl to Mt5 belonging to the mixed region and the pixels belonging to the corresponding background region are obtained from frame # n-3 to frame ⁇ + 2.
- Pixels Btl to Bt5 can be selected.
- white circles indicate pixels considered to belong to the background area
- black circles indicate pixels considered to belong to the mixed area.
- the five formulas (11) to (15) include the common variables such as the mixing ratio and the sum f of the foreground components, the formulas (11) to (15) By applying the least squares method, it is possible to obtain the mixing ratio f and the sum f of the foreground components.
- the estimated mixture ratio processing unit 201 previously stores a normal equation for calculating the sum f of the mixture ratio and the foreground component, and stores the pixel values belonging to the mixed region in the stored normal equation. Then, the pixel value belonging to the corresponding background area is set, and the mixture ratio Of and the sum f of the foreground components are calculated by a matrix solution.
- the estimated mixture ratio processing unit 201 associates the pixel value belonging to the mixed region with the normal equation in accordance with the background motion amount ⁇ ′, Then, the pixel values belonging to the corresponding background region are set, and the mixture ratio h and the sum f of the foreground components are calculated by the matrix solution.
- white circles indicate pixels considered to belong to the background area
- black circles indicate pixels considered to belong to the mixed area.
- the estimated mixture ratio processing unit 201 calculates the estimated mixture ratio using the motion amount V based on the model corresponding to the covered background area.
- the estimated mixture ratio processing unit 202 calculates the estimated mixture ratio using the motion amount V based on the model corresponding to the uncovered background area.
- the pixels belonging to the corresponding background area are selected from the frame after the frame of the pixel of interest.
- FIG. 26 is a block diagram illustrating a configuration of the estimated mixture ratio processing unit 201 that calculates the estimated mixture ratio using the motion amount V based on the model corresponding to the covered background area.
- the normal equation adding unit 222 Based on the motion vector supplied from the motion detection unit 102 and its position information, the normal equation adding unit 222 adds M images of the input image to the normal equation stored in advance. Set the pixel values belonging to the mixed area and the pixel values belonging to the corresponding background area included in. The normal equation adding unit 2 2 1 calculates the normal equation in which the pixel values belonging to the mixed area and the pixel values belonging to the corresponding background area are set, by an arithmetic unit 2 2 2 To supply.
- the arithmetic unit 222 solves the normal equation with the pixel values supplied from the normal equation adding unit 222 by applying a matrix solution method such as a sweeping-out method (Gauss-Jordan elimination method), and estimates Calculate the mixture ratio and output the calculated estimated mixture ratio.
- the calculation unit 222 outputs the calculated sum of the foreground components as the mixture ratio relationship information in accordance with the calculation of the estimated mixture ratio.
- the estimated mixture ratio processing unit 201 calculates the estimated mixture ratio using the motion amount V based on the model corresponding to the covered background area.
- the estimated mixture ratio processing unit 201 outputs the sum of foreground components as mixture ratio relationship information.
- the estimated mixture ratio processing unit 202 has the same configuration as the estimated mixture ratio processing unit 201, the description thereof is omitted.
- the mixture ratio calculation unit 103 assumes that the pixel belongs to the covered background area based on the input image, the motion vector supplied from the motion detection unit 102, and its position information. Of the estimated mixture ratio and the corresponding mixture ratio relationship information at the same time, and the estimated mixture ratio, which is the estimated mixture ratio when the pixel belongs to the uncovered background area, and the estimated mixture ratio. Corresponding mixture ratio relationship information can be generated.
- step S201 the estimated mixture ratio processing unit 201 converts the input background image into the covered background area based on the input image and the motion vector and its position information supplied from the motion detection unit 102. Execute the mixture ratio estimation process using the corresponding model. The details of the mixture ratio estimation process will be described later with reference to the flowchart in FIG.
- step S202 the estimated mixture ratio processing unit 202 responds to the anchor bird background area based on the input image and the motion vector and its position information supplied from the motion detection unit 102. Execute the mixture ratio estimation process using the model.
- step S203 the mixture ratio calculation unit 103 determines whether or not the mixture ratio has been estimated for the entire frame. If it is determined that the mixture ratio has not been estimated for the entire frame, the process proceeds to step S203. Returning to step S201, processing for estimating the mixture ratio for the next pixel is executed.
- step S203 If it is determined in step S203 that the mixture ratio has been estimated for the entire frame, the process ends.
- the mixture ratio calculation unit 103 calculates an estimated mixture ratio corresponding to each pixel based on the motion vector and its position information supplied from the motion detection unit 102 and the input image. be able to.
- step S221 the normal equation adding unit 221 reads the motion vector supplied from the motion detection unit 102 and the position information thereof, and acquires the motion amount V.
- step S222 the normal equation adding unit 221 selects a pixel from the input M frame images based on the motion amount V, and selects a pixel in a previously stored normal equation. Set the pixel value of the pixel.
- step S223 the normal equation adding unit 221 determines whether the setting of the pixel value has been completed for the target pixel, and has completed the setting of the pixel value for the target pixel. If it is determined that there is no pixel value, the process returns to step S222, and the process of setting the pixel value is repeated.
- step S224 the normal equation adding unit 2 21 calculates the normal equation in which the pixel value is set.
- the calculation unit 222 supplies the calculation unit 222.
- the calculation unit 222 solves the normal equation by applying a sweeping method (Gauss-Jordan elimination method) to the normal equation in which the pixel value is set, and obtains the estimated mixture ratio. Is calculated, and the process ends.
- the operation unit 2 2 2 Along with the calculation of the constant mixture ratio, the calculated sum of the foreground components corresponding to each pixel is output as mixture ratio relation information.
- the estimated mixture ratio processing unit 201 can calculate the estimated mixture ratio.
- step S202 in FIG. 27 the estimated mixture ratio processing unit 202 executes the uncovered
- the process of estimating the mixture ratio using the model corresponding to the background region is the same as the process shown in the flowchart of FIG. 28 using the normal equation corresponding to the model of the uncovered background region, and a description thereof will be omitted. .
- FIG. 29 is a block diagram showing another configuration of the mixture ratio calculation unit 103. As shown in FIG. The mixture ratio calculation unit 103 whose configuration is shown in FIG. 29 does not use a motion vector.
- the estimated mixture ratio processor 241 calculates an estimated mixture ratio for each pixel based on the input image by an operation corresponding to the model of the covered background area, and calculates the estimated mixture ratio with the calculation of the estimated mixture ratio. Outputs the estimated mixture ratio together with the ratio relation information.
- the mixture ratio relationship information output by the estimated mixture ratio processing unit 241 includes, for example, a sum of foreground components, a gradient of the mixture ratio, and the like.
- the estimated mixture ratio processing section 242 calculates the estimated mixture ratio for each pixel by an operation corresponding to the model of the uncovered background area based on the input image, and is calculated along with the calculation of the estimated mixture ratio.
- the estimated mixture ratio is output together with the mixture ratio relationship information.
- the mixture ratio relation information output by the estimated mixture ratio processing unit 242 includes, for example, the sum of foreground components, the gradient of the mixture ratio, and the like.
- the estimated mixture ratio processing unit 241 or the estimated mixture ratio processing unit 242 responds to the change in the position of the pixel by moving the object corresponding to the foreground at a constant speed within the shutter time.
- the mixing ratio ⁇ changes linearly
- an equation is established that approximates the mixture ratio h and the sum f of the foreground components.
- an equation that approximates the mixture ratio H and the sum f of the foreground components is solved.
- Equation (16) i is an index in the spatial direction where the position of the pixel of interest is 0. 1 is the slope of the straight line of the mixing ratio a. p is the intercept of the straight line of the mixture ratio Of and the mixture ratio a of the pixel of interest. In equation (16), index i is known, but slope 1 and intercept p are unknown.
- Figure 30 shows the relationship between the index, the slope, and the intercept p.
- a white circle indicates a target pixel
- a black circle indicates a neighboring pixel.
- Equation (16) a plurality of different mixture ratios for a plurality of pixels are represented by two variables.
- the five mixture ratios for the five pixels are expressed by two variables, a slope 1 and an intercept p.
- Equation (17) j is the horizontal index with the position of the pixel of interest set to 0, and k is the vertical index.
- m is the horizontal inclination of the surface of the mixture ratio H
- q is the vertical inclination of the surface of the mixture ratio H.
- p is the intercept of the plane of the mixture ratio Of.
- X represents a position in the spatial direction.
- equation (2 1) can be expressed as equation (2 2).
- Equation (22) j is the horizontal index with the position of the pixel of interest set to 0, and k is the vertical index.
- Equation (27) the unknown variables are the horizontal slope m of the surface of the mixture ratio, the vertical slope q of the surface of the mixture ratio, the intercept p, s, t, and the surface of the surface of the mixture ratio. u have six.
- a pixel value M and a pixel value B are set in Equation (27) in correspondence with a pixel in the vicinity of the pixel of interest, and for a plurality of expressions in which the pixel value M and the pixel value B are set, Calculate the mixture ratio by solving with the least squares method.
- Equation (28) The horizontal slope m, vertical slope q, intercept p, s, t, and u are expressed as variables w0, wl, w2, w3, w4, and w5, respectively, and jB, kB, B, j, k , And 1 are expressed as aO, al, a2, a3, a4 v and a5, respectively.
- equation (28) can be rewritten as equation (38).
- X is any one of integers from 0 to 8.
- equation (39) can be derived.
- equation (42) is obtained.
- wO is the horizontal slope m
- wl is the vertical slope q
- w2 is the intercept p
- w3 is s
- w4 is t
- w5 is u.
- the horizontal slope m, the vertical slope q, the intercept p, s, t, and u are obtained. be able to.
- Mc2 (0)-Bc2m + (-l)-Bc2-q + Bc2p + (0) s + (-1)-t + u (44
- Mc3 (+ 1) Bc3m + (-1) Bc3 q + Bc3p + (+ l) s + (_ l)-t + u (4 5
- Mc4 (-1)-Bc4-m + (0)-Bc4q + Bc4p + (_ l)-s + (0)-t + u (4 6
- Mc5 (0) Bc5-m + (0) Bc5-q + Bc5p + (0)-s + (0)-t + u (4 7
- Mc6 (+ 1) Bc6m + (0)-Bc6q + Bc6p + (+ l) s + (0) -t + u (4 8
- Mc7 (-1)-Bc7-m + (+ l)-Bc7-q + Bc7p + (-l)-s + (+ l)-t + u (4 9
- Mc8 (0)-Bc8m + (+ l) Bc8q + Bc8-p + (0) s + (+ l)-t + u
- Mc9 (+ 1) Bc9-m + (+ l ) ⁇ Bc9 ⁇ q + Bc9 ⁇ p + (+ l) ⁇ s + (+ l)-t + u
- the equation (4 3 In Equations (5 1) to (5 1) the pixel values Bel to Bc9 of the pixels in the background area of the pixels in frame - ⁇ -1 corresponding to the pixels in frame ttn are used ( uncovered background shown in FIG. 32 ) .
- the mixture ratio H of the pixels included in the region the following equations (52) to (60) are established: The pixel value of the pixel for which the mixture ratio is calculated is Mu5.
- mixture ratio estimation processing section 241 and the mixture ratio estimation processing section 242 have the same configuration as the mixture ratio processing section 201, the description thereof is omitted.
- the process of calculating the estimated mixture ratio by the mixture ratio calculation unit 103 having the configuration shown in FIG. 29 is the same as the process described with reference to the flowchart in FIG. 27, and a description thereof will be omitted.
- step S241 the estimated mixture ratio processing unit 241 sets a pixel value included in the input image to a normal equation corresponding to the model of the covered background area.
- step S224 the estimated mixture ratio processing unit 241 determines whether the setting for the target pixel has been completed, and determines that the setting for the target pixel has not been completed. If so, the process returns to step S241 to repeat the process of setting the pixel value in the normal equation.
- step S224 When it is determined in step S224 that the setting of the pixel value for the target pixel has been completed, the process proceeds to step S243, where the estimated mixture ratio processing unit 2411 sets the pixel value.
- the estimated mixture ratio is calculated, and the obtained estimated mixture ratio is output.
- the estimated mixture ratio processing unit 241 outputs the sum of the foreground components and the gradient of the estimated mixture ratio as the mixture ratio relationship information along with the calculation of the estimated mixture ratio.
- the estimated mixture ratio processing unit 241 can calculate the estimated mixture ratio based on the input image.
- the estimated mixture ratio processing unit 241 outputs the sum of the components of the foreground and the gradient of the estimated mixture ratio as the mixture ratio relation information.
- the processing of the mixture ratio estimation by the model corresponding to the uncovered background region by the estimated mixture ratio processing unit 242 is performed by the model of the uncovered background region.
- the process is the same as the process shown in the flowchart of FIG. 33 using the normal equation corresponding to the rule, and a description thereof will be omitted.
- FIG. 34 is a block diagram illustrating still another configuration of the mixture ratio calculation unit 103 that estimates a mixture ratio from an input image input as a component signal.
- a component refers to an individual signal such as a luminance signal and a color difference signal or an RGB (Red-green-blue) signal in a component signal.
- component 1 is assumed to be a luminance value Y
- component 2 is assumed to be a color difference U
- component 3 is assumed to be a color difference V.
- the estimated mixture ratio processing unit 241-1-1 calculates the estimated mixture ratio for each pixel based on the component 1 of the input image by calculation corresponding to the model of the covered background area, and calculates the estimated mixture ratio.
- the estimated mixture ratio is output together with the mixture ratio relation information calculated accordingly.
- the mixture ratio relationship information output by the estimated mixture ratio processing unit 24-1-1 is composed of, for example, the sum of foreground components and the gradient of the mixture ratio.
- the estimated mixture ratio processor 2 4 2-1-1 calculates the estimated mixture ratio for each pixel based on the component 1 of the input image by calculation corresponding to the model in the anchored background area, and calculates the estimated mixture ratio
- the estimated mixing ratio is output together with the mixing ratio relation information calculated in accordance with.
- the mixture ratio relation information output by the estimated mixture ratio processing unit 242-1 includes, for example, a sum of foreground components, a gradient of the mixture ratio, and the like.
- the estimated mixture ratio processor 2 4 1-2-2 calculates the estimated mixture ratio for each pixel based on the component 2 of the input image by calculation corresponding to the model of the covered background area, and calculates the estimated mixture ratio.
- the estimated mixture ratio is output together with the mixture ratio relation information calculated accordingly.
- the estimated mixture ratio processor 2 4 2-2-2 calculates the estimated mixture ratio for each pixel based on the component 2 of the input image by calculation corresponding to the model of the anchored background area, and calculates the estimated mixture ratio
- the estimated mixing ratio is output together with the mixing ratio relation information calculated in accordance with.
- the estimated mixture ratio processing unit 2 4 1-3 calculates the estimated mixture ratio for each pixel based on the component 3 of the input image by calculation corresponding to the model of the covered background area, and calculates the estimated mixture ratio.
- the estimated mixture ratio is output together with the mixture ratio relation information calculated accordingly.
- the estimated mixture ratio processing unit 2 4 2-3 calculates the estimated mixture ratio for each pixel based on the component 3 of the input image by calculation corresponding to the model of the anchored background area, and calculates the estimated mixture ratio
- the estimated mixing ratio is output together with the mixing ratio relation information calculated in accordance with.
- the mixture ratio calculation unit 103 whose configuration is shown in FIG. 34 is based on the input image input as the component signal, and for each component, the estimated mixture ratio based on the model corresponding to the covered back-drop region. , And an estimated mixture ratio by a model corresponding to the uncovered backdrop region.
- the mixture ratio calculation unit 103 calculates, for each component, the sum of the foreground components corresponding to the estimated mixture ratio by the model corresponding to the covered background region and the estimated mixture ratio by the model corresponding to the uncovered background region. And output the mixture ratio relation information including the inclination of the estimated mixture ratio.
- the mixture ratio calculation unit 103 uses the estimated mixture ratio processing unit 201 and the estimated mixture ratio processing unit 2.
- the estimated mixture ratio of the model corresponding to the covered background region and the estimated mixture ratio of the model corresponding to the uncovered background region are calculated for each component, and the estimated mixture ratio is calculated together with the estimated mixture ratio.
- the mixing ratio relationship information may be output.
- the above-described processing for obtaining the mixture ratio can be applied even when the image corresponding to the background area includes motion.
- the mixture ratio calculator 103 shifts the entire image in response to this movement, and the object corresponding to the background is stationary.
- the processing is performed in the same manner as in the case. Also, it corresponds to the background area
- the mixture ratio calculator 103 selects the pixel corresponding to the motion as the pixel corresponding to the pixel belonging to the mixed area, and performs the above-described processing. .
- FIG. 35 is a block diagram showing a configuration of the area specifying unit 104.
- the prediction error calculation unit 301 corresponds to each pixel based on the estimated mixture ratio based on the model corresponding to the covered background region and the mixture ratio relationship information corresponding to the estimated mixture ratio based on the model corresponding to the covered background region.
- the calculated error value (a prediction error corresponding to the estimated mixture ratio) is calculated, and the calculated error value is supplied to the uncovered background area determination unit 303.
- the mixture ratio calculation unit 103 estimates the mixture ratio by applying the least squares method to the equations (11) to (15), the prediction error calculation unit 301 (61) based on the mixture ratio information, which is the sum of the components of the foreground corresponding to the estimated mixture ratio of the model corresponding to the background region and the estimated mixture ratio of the model corresponding to the bird background region.
- the least square error sum which is the error value S is calculated.
- the prediction error calculation unit 302 calculates each of the estimated mixing ratios based on the model corresponding to the uncovered background area and the mixing ratio relation information corresponding to the estimated mixing ratio based on the model corresponding to the uncovered background area. An error value corresponding to the pixel is calculated, and the calculated error value is supplied to the covered background area determination unit 304.
- the prediction error calculation unit 301 and the prediction error calculation unit 302 calculate, for example, an error value corresponding to each pixel for each block including a plurality of pixels such as a block including 5 pixels ⁇ 5 pixels. It may be.
- the prediction error calculation unit 301 and the prediction error calculation unit 302 may set t in advance, or obtain t from the mixture ratio calculation unit 103 as mixture ratio relation information. It may be.
- the uncovered background area determination unit 303 determines, for each pixel, whether or not the error value supplied from the prediction error calculation unit 301 is equal to or greater than a threshold value Th stored in advance. Is determined to be greater than or equal to the threshold Th, it is determined that the pixel of interest belongs to the uncovered background area, and a flag indicating the uncovered background area is set in association with the pixel of interest.
- the uncovered background area determination unit 303 supplies a flag indicating the uncovered background area set for each pixel to the synthesis unit 303.
- the covered background area determination unit 304 determines, for each pixel, whether or not the error value supplied from the prediction error calculation unit 302 is equal to or greater than a threshold value Th stored in advance. If is determined to be greater than or equal to the threshold Th, it is determined that the pixel of interest belongs to the covered background area, and a flag indicating the covered background area is set for the pixel of interest.
- the covered background area determination unit 304 supplies a flag indicating the covered background area set for each pixel to the combining unit 303.
- the foreground / background area determination unit 3005 determines a foreground area and a background area based on the input image.
- the foreground / background region determination unit 3005 calculates the difference between the pixel value of the pixel of interest in the frame of interest ttn and the pixel corresponding to the pixel of interest in frame # n-1. The motion of the pixel of interest is determined. The foreground / background region determination unit 3005 calculates the difference between the pixel value of the pixel of interest in the frame of interest ttn and the pixel corresponding to the pixel of interest in frame # n + l. The motion is determined for the pixel of interest. The foreground / background area determination unit 300 determines that the pixel of interest is determined to be moving from frame # n-l to frame # ⁇ , and is determined to be moving from frame to frame # n + l. It is determined that they belong.
- the foreground background area determination unit 3 0 when the pixel of interest, is determined from the frame # n -l for frame over arm stationary, is determined stationary for frame #n force al frame # n + l, attention It is determined that the pixel belongs to the background area.
- the foreground / background area determination unit 3005 supplies a flag indicating the foreground area and a flag indicating the background area, set for each pixel, to the synthesizing unit 303.
- the synthesizing unit 303 sets the flag indicating the uncovered background area supplied from the uncovered background area determination unit 303 and the covered background area supplied from the covered background area determination unit 304. On the basis of the flag indicating the foreground area supplied from the foreground / background area determination unit 3005, and the flag indicating the background area, for each pixel, the uncovered background area, the covered background area, the foreground area, And area information indicating one of the background areas.
- the combining section 303 supplies the combined area information to the mixture ratio determining section 307, and outputs the area information.
- the mixture ratio determination unit 307 determines the mixture ratio H based on the area information supplied from the synthesis unit 306. More specifically, the mixture ratio determination unit 307 sets 0 to the mixture ratio when the target pixel belongs to the foreground region, and sets 1 when the target pixel belongs to the background region. Set the mixing ratio to. If the target pixel belongs to the covered background region, the mixture ratio determination unit 307 sets the mixture ratio estimated by the model corresponding to the covered covered backdrop region to the mixture ratio H, and the target pixel If it belongs to the uncovered background region, set the mixture ratio estimated by the model corresponding to the uncovered background region to the mixture ratio. The mixture ratio determining unit 307 outputs the mixture ratio determined based on the area information.
- the determination process of the uncovered background area determination unit 303 will be described with reference to FIGS. 36 and 37.
- the pixel value of the pixel of interest belonging to the background area of the frame of interest ⁇ is the same as that of the next frame # ⁇ + 1 of frame ⁇ ⁇ . It is equal to the pixel value of the pixel corresponding to the target pixel.
- W indicates a background area.
- the pixel value of the pixel of frame # ⁇ + 1 indicated by the motion vector based on the target pixel is the pixel value of the corresponding pixel of frame ttn + 2. Equal to pixel value.
- the estimated mixture ratio of the model corresponding to the covered coverback region corresponding to the pixels belonging to the background region is substantially 1, and the sum of the components of the foreground is substantially zero.
- the pixel value of the pixel of interest in frame #n of interest is indicated by a motion vector based on the pixel of interest. Equal to the pixel value of the pixel in frame # n + l. Also, the foreground object has a strong correlation in the spatial direction, so that the values of the foreground components are almost equal.
- the foreground component included in the pixel value of the pixel of interest of the frame of interest belonging to the uncovered background area indicated by Z in the figure is the frame ttn Is different from the foreground component included in the pixel value of the pixel corresponding to the target pixel of the next frame lln + 1.
- the foreground component included in the pixel value of the pixel in frame # n + l, indicated by the motion vector based on the pixel of interest is It is different from the foreground component included in the pixel value of the pixel corresponding to the time ⁇ + 2.
- the least-squares error sum of the estimated mixture ratio and the sum of the foreground components calculated by the least squares method is calculated when the pixel of interest belongs to the uncovered background area. , which is a large value.
- the uncovered background area determination unit 303 calculates the estimated mixture ratio calculated by the least squares method based on the model corresponding to the covered covered background area and the error value of the sum of the foreground components. By determining whether or not S is equal to or greater than the threshold Th, it can be determined whether or not the corresponding pixel belongs to the uncovered background area.
- the covered background area determination unit 304 calculates the estimated mixture ratio and the error value S of the sum of the foreground components calculated by the least squares method based on the model corresponding to the uncovered background area. By determining whether or not the pixel is equal to or greater than the threshold Th, it is possible to determine whether or not the corresponding pixel belongs to the covered background area.
- FIGS. 38 to 43 are diagrams illustrating an example of an input image and a result of region determination performed by the region identifying unit 104 corresponding to the input image.
- FIG. 38 is a diagram showing an input image.
- the foreground object moves from the left side to the right side in the figure.
- FIG. 39 is based on the input image shown in FIG. 38, in which the mixture ratio calculation unit 103 identifies the region based on the estimated mixture ratio calculated using seven frames and the sum of the foreground components.
- FIG. 11 is a diagram illustrating a result of a determination made by a unit 104 with a threshold Th set to 70 for each pixel.
- FIG. 40 is based on the estimated mixture ratio and the sum of the components of the foreground calculated by the mixture ratio calculator 103 using the three frames for the input image shown in FIG. 38.
- the area specifying unit 104 sets the threshold value Th of the sum of the error values S for the block to 7500, assuming that the mixing ratio is constant for each block of 5 pixels ⁇ 5 pixels.
- FIG. 10 is a diagram illustrating a result of a determination made by setting a threshold value Th of an error value S of one pixel to 10;
- FIG. 41 is a diagram showing an input image.
- the foreground object moves from the left side to the right side in the figure.
- FIG. 10 is a diagram illustrating a result of a determination made by a unit 104 with a threshold Th set to 70 for each pixel.
- FIG. 43 the area is identified based on the estimated mixture ratio calculated using the three frames and the sum of the foreground components by the mixture ratio calculator 103 with respect to the input image shown in FIG. 41.
- the unit 104 sets the threshold value Th of the sum of the error values S for the block to 755, assuming that the mixing ratio is constant for each block of 5 pixels ⁇ 5 pixels, and sets one of the blocks in the block.
- FIG. 10 is a diagram illustrating a result of determination that a threshold value Th of a pixel error value S is set to 10;
- the covered background area is located at the front end of the foreground object in the traveling direction of the foreground object, and the uncovered background area is located at the foreground area with respect to the foreground area. It is located at the rear end of the object in the direction of travel.
- the area specifying unit 104 can determine the area almost accurately.
- the mixture ratio calculating unit 103 estimates the mixture ratio by applying the least squares method to the equations (28) to (36), and the data indicating the slope of the estimated mixture ratio together with the estimated mixture ratio.
- the process performed by the area specifying unit 104 when outputting as the mixture ratio relationship information will be described.
- the pixel value M of the target pixel is expressed by Expression (6 2) from Expression (27).
- Equation (62) j is a horizontal index with the position of the pixel of interest set to 0, and k is a vertical index.
- m is the horizontal inclination of the plane of the mixture ratio, and q is the vertical inclination of the plane of the mixture ratio.
- p is the intercept of the surface of the mixture ratio.
- s, t, and u are variables indicating the relationship between m, q, and P and the foreground components, as shown in equations (23) to (25).
- the prediction error calculation unit 301 calculates the estimated mixture ratio of the model corresponding to the covered background region and the estimated mixture ratio of the model corresponding to the covered background region, j, k, m, q, p, Based on the mixing ratio relationship information including S , t, and u, the least square error sum that is the error value S can be calculated by the operation corresponding to the equation (63).
- the prediction error calculation unit 302 uses a model corresponding to the uncovered covered background area Based on the estimated mixture ratio and the mixture ratio relationship information corresponding to the estimated mixture ratio based on the model corresponding to the uncovered backround region, the error value corresponding to each pixel is calculated by the operation corresponding to Equation (63). Is calculated, and the calculated error value is supplied to the covered background area determination unit 304.
- the uncovered background area determination unit 303 previously stores, for each pixel, the error value supplied from the prediction error calculation unit 301 and determines whether or not the error value is equal to or greater than a threshold Th. If the value is determined to be equal to or greater than the threshold Th, it is determined that the pixel of interest belongs to the uncovered background area, and a flag indicating the uncovered background area is set in association with the pixel of interest. .
- the uncovered background area determination unit 303 supplies a flag indicating an uncovered background area set for each pixel to the synthesis unit 303.
- the covered background area determination unit 304 determines, for each pixel, whether or not the error value supplied from the prediction error calculation unit 302 is equal to or greater than a threshold value Th stored in advance. If the value is determined to be equal to or greater than the threshold Th, the pixel of interest is covered. It determines that the pixel belongs to one background area, and sets a flag indicating the covered background area to the pixel of interest. The covered background area determination unit 304 supplies a flag indicating the covered background area set for each pixel to the combining unit 303.
- the foreground / background region determination unit 3005 determines the foreground region and the background region based on the input image, and combines the flag indicating the foreground region and the flag indicating the background region set for each pixel. 0 to 6
- the synthesizing section 303 sets the flag indicating the uncovered background area supplied from the uncovered background area determination section 303 and the covered background area supplied from the covered background area determination section 304. On the basis of the flag indicating the foreground area supplied from the foreground / background area determination unit 3005, and the flag indicating the background area, the uncovered background area, the covered background area, the foreground area, and Area information indicating one of the background areas is synthesized.
- the synthesizing unit 303 supplies the synthesized region information to the mixture ratio determining unit 307, and outputs the region information.
- the mixing ratio determining unit 307 determines the mixing ratio a based on the area information supplied from the synthesizing unit 306.
- the region identification unit 104 can generate the region information based on the estimated mixture ratio and the mixture ratio relationship information in accordance with the mixture ratio estimation method of the mixture ratio calculation unit 103. it can.
- FIG. 44 shows the configuration of the region specifying unit 104 that specifies a region based on the mixture ratio estimated for each component signal and the corresponding mixture ratio relationship information, and the input image input as the component signal. It is a block diagram shown. The same portions as those shown in FIG. 35 are denoted by the same reference numerals, and description thereof will be omitted.
- the prediction error calculator 3 0 1—1 calculates the estimated mixture ratio of the model corresponding to the covered background region and the estimated mixture ratio of the model corresponding to the covered background region, calculated from the component 1 of the input image. Corresponding mixing ratio An error value corresponding to each pixel is calculated based on the relationship information, and the calculated error value is supplied to the adder 3221.
- the prediction error calculation unit 310_1_2 corresponds to the estimated mixture ratio of the model corresponding to the covered background region and the estimated mixture ratio of the model corresponding to the covered background region, calculated from the component 2 of the input image. An error value corresponding to each pixel is calculated based on the mixing ratio relationship information to be calculated, and the calculated error value is supplied to the adding unit 3221.
- the prediction error calculator 3 0 1-3 supports the estimated mixture ratio of the model corresponding to the covered background region and the estimated mixture ratio of the model corresponding to the covered background region, calculated from the component 3 of the input image. An error value corresponding to each pixel is calculated based on the mixing ratio relationship information to be calculated, and the calculated error value is supplied to the adding unit 3221.
- the prediction error calculator 3 0 2 _ 1 calculates the estimated mixture ratio of the model corresponding to the uncovered background region and the estimated mixture ratio of the model corresponding to the uncovered background region, calculated from the component 1 of the input image. An error value corresponding to each pixel is calculated based on the corresponding mixture ratio relationship information, and the calculated error value is supplied to the adder 322.
- the prediction error calculator 3 0 2-2 calculates the estimated mixture ratio of the model corresponding to the uncovered background region and the estimated mixture ratio of the model corresponding to the uncovered background region, calculated from the component 2 of the input image. An error value corresponding to each pixel is calculated based on the corresponding mixture ratio relationship information, and the calculated error value is supplied to the adder 322.
- the prediction error calculator 3 0 2-3 calculates the estimated mixture ratio of the model corresponding to the uncovered background region and the estimated mixture ratio of the model corresponding to the uncovered background region, calculated from the component 3 of the input image. An error value corresponding to each pixel is calculated based on the corresponding mixture ratio relationship information, and the calculated error value is supplied to the adder 322.
- the adder 3 2 1 receives the error value supplied from the prediction error calculator 3 0 1-1, the error value supplied from the prediction error calculator 3 0 1 1 1 1, and the prediction error calculator 3 0 1 _ 3 The supplied error values are added, and the added error value is supplied to the anchored background area determination unit 303.
- the adder 3 2 2 receives the error value supplied from the prediction error calculator 3 0 2-1, the error value supplied from the prediction error calculator 3 0 2-2, and the prediction error calculator 3 0 2-3 The supplied error values are added, and the added error value is supplied to the covered background area determination unit 304.
- the addition unit 3 2 3 adds the component 1 of the input image, the component 2 of the input image, and the component 3 of the input image, and adds the added component 1, component 2, and component 3 to the foreground / background region determination unit 3 0 5 the c uncovered-background-area determining unit 3 0 3 supplies, for each pixel, the error value supplied from the adder 3 2 1, it is determined whether the threshold Th or stored in advance. If the error value is determined to be equal to or greater than the threshold Th, it is determined that the pixel of interest belongs to the uncovered background area, and a flag indicating the uncovered background area is set in association with the pixel of interest. I do.
- the uncovered background area determination unit 303 supplies a flag indicating the uncovered background area set for each pixel to the synthesis unit 303.
- the covered background area determination unit 304 determines, for each pixel, whether or not the error value supplied from the addition unit 3222 is equal to or greater than a threshold value Th stored in advance. If it is determined to be Th or more, it is determined that the pixel of interest belongs to the covered background area, and a flag indicating the covered background area is set for the pixel of interest.
- the covered background area determination unit 304 supplies a flag indicating the covered background area set for each pixel to the synthesizing unit 303.
- the foreground / background region determination unit 3005 determines a foreground region and a background region based on the added component 1, component 2, and component 3, and The flag indicating the foreground area and the flag indicating the background area, which have been set in, are supplied to the synthesizing unit 310.
- the combining unit 303 includes a flag indicating the uncovered background area supplied from the uncovered background area determination unit 303, and a covered background area supplied from the power bird background area determination unit 304. Based on the flag indicating the foreground area and the flag indicating the background area supplied from the foreground / background area determination unit 305 for each pixel based on the uncovered background area, the covered background area, and the foreground.
- the area information indicating either the area or the background area is synthesized.
- the combining section 303 supplies the combined area information to the mixture ratio determining section 307, and outputs the area information.
- the region specifying unit 104 shown in FIG. 44 uses the mixture ratio estimated for each component signal and the mixture ratio relation information corresponding thereto, and the input image input as the component signal, based on the input image.
- the area can be specified.
- the area specifying unit 104 having the configuration shown in FIG. 44 can specify the area more accurately than the area specifying unit 104 having the configuration shown in FIG.
- FIG. 45 shows another example of the region specifying unit 104 that specifies a region based on the mixture ratio estimated for each component signal and the corresponding mixture ratio relation information, and the input image input as the component signal. It is a block diagram which shows a structure.
- the prediction error calculator 3 0 1—1 corresponds to the estimated mixture ratio of the model corresponding to the covered background region and the estimated mixture ratio of the model corresponding to the covered background region, calculated from the component 1 of the input image. An error value corresponding to each pixel is calculated based on the mixing ratio relationship information to be calculated, and the calculated error value is supplied to the anchored background area determination unit 303-3-1.
- the prediction error calculation unit 310_1_2 calculates the estimated mixture ratio of the model corresponding to the covered background area, calculated from the component 2 of the input image, and the coverage. An error value corresponding to each pixel is calculated based on the mixture ratio relationship information corresponding to the estimated mixture ratio by the model corresponding to the background region, and the calculated error value is used as an anchor background region determination unit. Supply 2
- the prediction error calculator 3 0 1-3 corresponds to the estimated mixture ratio of the model corresponding to the covered background region and the estimated mixture ratio of the model corresponding to the covered background region, calculated from the component 3 of the input image. An error value corresponding to each pixel is calculated based on the mixing ratio relationship information to be calculated, and the calculated error value is supplied to the anchored background area determination unit 303-3-3.
- the prediction error calculator 3 0 2—1 calculates the estimated mixture ratio of the model corresponding to the uncovered background region and the estimated mixture ratio of the model corresponding to the uncovered background region, calculated from the component 1 of the input image. An error value corresponding to each pixel is calculated based on the corresponding mixture ratio relationship information, and the calculated error value is supplied to the covered background area determination unit 3044-1.
- the prediction error calculator 3 0 2-2 calculates the estimated mixture ratio of the model corresponding to the uncovered background region and the estimated mixture ratio of the model corresponding to the uncovered background region, calculated from the component 2 of the input image. An error value corresponding to each pixel is calculated based on the corresponding mixture ratio relationship information, and the calculated error value is supplied to the covered background area determination unit 3044-2.
- the prediction error calculator 3 0 2-3 calculates the estimated mixture ratio of the model corresponding to the uncovered background region and the estimated mixture ratio of the model corresponding to the uncovered background region, calculated from the component 3 of the input image. An error value corresponding to each pixel is calculated based on the corresponding mixture ratio relationship information, and the calculated error value is supplied to the covered background area determination unit 304_3.
- the covered background area determination unit 303-1 determines whether the error value supplied from the prediction error calculation unit 310-1 for each pixel is equal to or greater than a threshold value Th stored in advance. If the error value is determined to be equal to or greater than the threshold Th, it is determined that the pixel of interest belongs to the uncovered background area, and the image of interest is determined. A flag indicating the uncovered background area is set corresponding to the element. The covered background area determination unit 303-1 supplies a flag indicating the covered background area set for each pixel to the logical sum operation unit 341.
- ⁇ Uncovered background area determination unit 3 0 3-2 determines, for each pixel, whether or not the error value supplied from the prediction error calculation unit 3 0 1-2 is equal to or greater than a threshold value Th stored in advance, and If it is determined to be Th or more, it is determined that the pixel of interest belongs to the uncovered background area, and a flag indicating the uncovered background area is set in association with the pixel of interest.
- the covered background area determination unit 303-3-2 supplies the flag indicating the covered background area set for each pixel to the logical sum operation unit 341 (the uncovered background area determination unit).
- 3 0 3-3 determines, for each pixel, whether or not the error value supplied from the prediction error calculation unit 3 0 1 _ 3 is equal to or greater than a threshold value Th stored in advance, and If it is determined to be greater than or equal to Th, it is determined that the pixel of interest belongs to the uncovered background area, and a flag indicating the uncovered background area is set corresponding to the pixel of interest.
- the covered background area determination unit 303-3-3 supplies the flag indicating the covered background area set for each pixel to the logical sum operation unit 341.
- the region determining unit 304-1 determines, for each pixel, whether or not the error value supplied from the prediction error calculating unit 302-1 is equal to or greater than a threshold value Th stored in advance.
- the background area determination unit 3044-1 supplies a flag indicating a covered backdrop region set for each pixel to the logical sum operation unit 342.
- the covered background area determination unit 3044-2 determines, for each pixel, whether the error value supplied from the prediction error calculation unit 3002-2 is equal to or greater than a threshold value Th stored in advance. If it is determined that the error value is equal to or greater than the threshold Th, the image of interest is It determines that the element belongs to the covered background area, and sets a flag indicating the covered background area corresponding to the pixel of interest.
- the covered knock ground area determination unit 304-2 supplies the flag indicating the covered back dull area set for each pixel to the logical sum operation unit 3422.
- the covered background area determination unit 304_3 determines, for each pixel, whether the error value supplied from the prediction error calculation unit 302-3 is equal to or greater than a threshold value Th stored in advance. If it is determined that the error value is equal to or greater than the threshold Th, it is determined that the pixel of interest belongs to the covered background area, and a flag indicating the covered background area is set corresponding to the pixel of interest. I do.
- the covered knock area determination section 304-4-3 supplies a flag indicating a covered background area set for each pixel to the logical sum operation section 342.
- the foreground / background area determination unit 3005-1 determines the foreground area and the background area based on the component 1, and performs a logical OR of the flag indicating the foreground area and the flag indicating the background area set for each pixel. Supplied to the operation unit 3 4 3.
- the foreground / background area determination unit 3005-2 determines the foreground area and the background area based on the component 2, and ORs the flag indicating the foreground area and the flag indicating the background area set for each pixel. Supplied to the operation unit 3 4 3.
- the foreground / background area determination unit 300-5-3 determines the foreground area and the background area based on the component 3, and logically ORs the flag indicating the foreground area and the flag indicating the background area set for each pixel. Supplied to the operation unit 3 4 3.
- the OR operation unit 341 is a flag indicating the uncovered background area supplied from the uncovered background area determination unit 303-3, and is supplied from the uncovered background area determination unit 303-3-2. Based on the flag indicating the uncovered background area and the flag indicating the uncovered background area supplied from the covered background area determination unit 303, the covered background area determination unit 30 is provided. 3—Uncovered background area indicated by flag supplied from 1 and uncovered background area judgment The logical sum of the covered background area indicated by the flag supplied from part 3 03-3 and the uncovered background area indicated by the flag supplied from the covered background area determination section 303-3 Is calculated to generate a flag indicating the uncovered background area calculated by the logical sum. The logical sum operation unit 341 supplies the generated flag indicating the uncovered background area to the synthesis unit 306.
- the OR operation unit 342 includes a flag indicating the covered background area supplied from the covered background area determination unit 3044-1 and the covered background area determination unit 3044-2 supplied from the covered background area determination unit. Based on the flag indicating the ground area and the flag indicating the covered background area supplied from the covered background area determination section 304-4-3, the flag supplied from the covered background area determination section 3044-1 The flag supplied from the covered background area indicated by the flag and the flag supplied from the covered background area judging section 304-1-2, and the flag supplied from the force bird background area judging section 304-4-3 Calculate the logical sum of the indicated covered background area and calculate with the logical sum Generating a flag indicating the covered background area. The OR operation unit 342 supplies the generated flag indicating the covered background area to the synthesis unit 306.
- the OR operation unit 343 includes a flag indicating the foreground region supplied from the foreground / background region determination unit 3005-1, a flag indicating the foreground region supplied from the foreground / background region determination unit 3005-2, And the foreground / background area determination unit 3005--3 determines the foreground area and foreground background area indicated by the flag supplied from the foreground / background area determination unit 3005-1 based on the flag indicating the foreground area. Calculates the logical sum of the foreground area indicated by the flag supplied from the section 3055-2 and the foreground area indicated by the flag supplied from the foreground / background area determination section 305-5-3, and calculates the logical sum. A flag indicating the foreground area is generated.
- the OR operation unit 343 supplies the generated flag indicating the foreground area to the synthesis unit 306.
- the OR operation unit 3 4 3 outputs the background region supplied from the foreground / background region determination unit 3 05—1. Based on the flag indicating the background area, the flag indicating the background area supplied from the foreground / background area determination unit 300-5-2, and the flag indicating the background area supplied from the foreground / background area determination unit 300-5-3.
- the synthesizing unit 303 includes a flag indicating the uncovered backdrop region supplied from the logical sum calculating unit 341, a flag indicating the covered background region supplied from the logical sum calculating unit 342, and a logical sum.
- the combining section 303 outputs the combined area information.
- the area specifying unit 104 having the configuration shown in FIG. 45 can output the area information specifying the foreground area, the background area, the covered background area, and the uncovered background area without leaking.
- FIG. 46 is a diagram illustrating a mixture ratio estimated for each component signal and mixture ratio relation information corresponding thereto, and an area specifying unit 104 for specifying a region based on an input image input as a component signal.
- FIG. 3 is a block diagram showing the configuration of FIG. Portions similar to those in the case shown in FIG. 45 are denoted by the same reference numerals, and description thereof will be omitted.
- the AND operation unit 36 1 is provided with a flag indicating the uncovered background area supplied from the uncovered background area determination unit 3 0 3-1, and supplied from the uncovered background area determination unit 3 0 3-2. Based on a flag indicating the uncovered background area and a flag indicating the uncovered background area supplied from the uncovered background area determination unit 303_3. Uncovered background area judging unit 3 03-1 Uncovered background area indicated by the flag supplied from 3-1, Uncovered background area judging unit 3 0 3—Uncovered background indicated by the flag supplied from 3-2 Calculates the logical product of the uncovered background region indicated by the flag supplied from the ground region and the uncovered background region determination unit 303 to indicate the uncovered background region calculated by the logical product Generate a flag. The AND operation unit 361 supplies the generated flag indicating the uncovered background area to the synthesis unit 306.
- the AND operation unit 365 includes a flag indicating the covered background area supplied from the covered background area determination unit 304-1, and the covered back area determination unit 304-4 supplied from the covered background area determination unit. Based on the flag indicating the ground area and the flag indicating the covered background area supplied from the covered background area determination unit 304-4-3, the flag supplied from the covered background area determination unit 3044-1 is The flag supplied from the covered background area indicated by the flag provided by the covered background area judgment unit 304-4-2 and the flag supplied by the force bird background area judgment unit 304_3 Calculates the logical product of the indicated covered background areas and calculates the logical product. Generate a flag indicating the covered background area. The logical product operation unit 365 supplies the generated flag indicating the covered background area to the synthesis unit 360.
- the AND operation unit 365 includes a flag indicating the foreground region supplied from the foreground / background region determination unit 3005-1, a flag indicating the foreground region supplied from the foreground / background region determination unit 3005-2, And the foreground / background area determination unit 300-5-3, based on the flag indicating the foreground area supplied from the foreground / background area determination unit 300-5—determination of the foreground area and foreground background area indicated by the flag supplied from the Calculates the logical product of the foreground area indicated by the flag supplied from the section 3 05-2 and the foreground area indicated by the flag supplied from the foreground / background area determination section 3 2005-3, and calculates the logical product A flag indicating the foreground area is generated.
- the calculation unit 363 supplies the generated flag indicating the foreground area to the synthesis unit 306.
- the AND operation unit 365 includes a flag indicating the background area supplied from the foreground / background area determination unit 3005-11, a flag indicating the background area supplied from the foreground / background area determination unit 3005-2, And the foreground / background area determination unit 3005--3 determines the background area and foreground background area indicated by the flag supplied from the foreground / background area determination unit 3005-1 based on the flag indicating the background area. Calculates the logical product of the background area indicated by the flag supplied from section 3 05-2 and the background area indicated by the flag supplied from foreground background area determination section 3 05-3, and calculates the logical product.
- a flag is generated that indicates the background area that has been set.
- the logical product operation unit 366 supplies a flag indicating the generated background area to the synthesis unit 306.
- the synthesizing unit 360 includes a flag indicating the uncovered backdrop region supplied from the logical product calculating unit 361, a flag indicating the covered background region supplied from the logical product calculating unit 362, and logical product. Indicates one of the uncovered background area, covered background area, foreground area, and background area for each pixel based on the flag indicating the foreground area and the flag indicating the background area supplied from the calculation unit 365. Synthesize area information.
- the combining section 303 outputs the combined area information.
- the area specifying unit 104 whose configuration is shown in FIG. 46 can output area information with fewer errors.
- step S301 the area specifying unit 104 executes specific processing of the covered background area for each pixel based on the error value. Details of the specific processing of the covered background area will be described later.
- step S302 the area specifying unit 104 executes a specific process of the uncovered background area for each pixel based on the error value. Details of the specific processing of the uncovered background area will be described later.
- step S303 the foreground / background area determination unit 3005 specifies a foreground area and a background area for each pixel based on the input image, and the process ends.
- the specific processing of the covered background area corresponding to step S301 will be described.
- step S321 the prediction error calculation section 302 calculates an error value by a model corresponding to the uncovered background region.
- the prediction error calculation unit 304 supplies the calculated error value to the covered background area determination unit 304.
- step S322 the covered background area determination unit 304 determines whether the error value supplied from the prediction error calculation unit 302 is equal to or larger than the threshold Th, and determines whether the error value is equal to the threshold Th. If it is determined that it is equal to or greater than Th, the flow proceeds to step S323, and a flag indicating the covered background area is set to the pixel corresponding to the error value. The procedure proceeds to step S324.
- step S322 If it is determined in step S322 that the error value is not equal to or greater than the threshold Th, the processing in step S332 is skipped because the area is not in the covered background area, and the procedure proceeds to step S322. Proceed to.
- step S3224 the area specifying unit 104 determines whether or not processing has been performed on the entire screen, and when it is determined that processing has not been performed on the entire screen, the area specifying unit 104 determines in step S321. Return, and repeat the process of determining the covered background area. If it is determined in step S324 that the entire screen has been processed, the process proceeds to step S325, in which the covered background area determination unit 304 outputs a flag indicating the covered background area. The process ends.
- step S302 With reference to the flowchart in FIG. 49, the specific processing of the anchored background area corresponding to step S302 will be described.
- step S3401 the prediction error calculation unit 310 calculates an error value based on a model corresponding to the covered background area.
- the prediction error calculation unit 301 supplies the calculated error value to the uncovered background area determination unit 303.
- step S3342 the uncovered background area determination unit 303 determines whether or not the error value supplied from the prediction error calculation unit 301 is equal to or greater than the threshold Th, and determines whether the error value is equal to the threshold Th. If it is determined to be equal to or greater than Th, go to step S 3 4 3 Then, a flag indicating the uncovered background area is set to the pixel corresponding to the error value, and the procedure proceeds to step S344.
- step S3342 If it is determined in step S3342 that the error value is not greater than or equal to the threshold value Th, the processing is not skipped in the uncovered background area, and the process in step S3343 is skipped, and the procedure proceeds to step S3344. move on.
- step S3344 the area specifying unit 104 determines whether or not processing has been performed for the entire screen, and when it is determined that processing has not been performed for the entire screen, the area specifying unit 104 determines in step S3341. Return and repeat the process of determining the uncovered background area.
- step S344 If it is determined in step S344 that the entire screen has been processed, the process proceeds to step S345, where the uncovered background area determination unit 303 outputs a flag indicating the uncovered background area. Then, the process ends.
- the region specifying unit 104 calculates the error value based on the estimated mixture ratio and the mixture ratio relationship information corresponding to the estimated mixture ratio, and based on the calculated error value, The uncovered background area can be specified.
- the region specifying unit 104 can determine the mixture ratio H based on the generated region information, and can output the determined mixture ratio H.
- the area specifying unit 104 shown in FIG. 44 has a configuration similar to that of the flow charts of FIGS. 47 to 49 except that a covered background area or an uncovered background area is specified based on the added error value. Since the area is specified by the same processing as that described with reference to FIG. 1, detailed description of the processing is omitted.
- the area specifying unit 104 shown in FIG. 45 specifies the area for each component, except that the final area is obtained by the logical sum of the specified areas. Since the area is specified by the same processing as described with reference to the flowchart of FIG. 49, detailed description of the processing is omitted.
- the area specifying unit 104 shown in FIG. 46 specifies the area for each component, except that the final area is obtained by the logical product of the specified areas. Since the area is specified by the same processing as the processing described with reference to the flowchart of FIG. 7, detailed description of the processing is omitted.
- FIG. 50 is a block diagram illustrating an example of the configuration of the foreground / background separation unit 105.
- the input image supplied to the foreground / background separation unit 105 is supplied to the separation unit 601, switch 602, and switch 604.
- the information indicating the covered background area and the area information indicating the uncovered background area supplied from the area specifying unit 104 are supplied to the separating unit 601.
- the area information indicating the foreground area is supplied to the switch 602.
- the area information indicating the background area is supplied to the switch 604.
- the mixing ratio supplied from the mixing ratio calculation unit 103 is supplied to the separation unit 601.
- the separation unit 601 is an area information indicating a covered background area, and an area indicating an uncovered background area. Based on the information and the mixture ratio ⁇ , the foreground components are separated from the input image, and the separated foreground components are supplied to the synthesis unit 603, and the background components are separated from the input image.
- the separated background component is supplied to the synthesis unit 605.
- the switch 602 is closed when the pixel corresponding to the foreground is input based on the area information indicating the foreground area, and supplies only the pixel corresponding to the foreground included in the input image to the synthesizing unit 603 I do.
- the switch 604 is closed when a pixel corresponding to the background is input based on the area information indicating the background area, and supplies only the pixel corresponding to the background included in the input image to the combining unit 605. I do.
- the combining unit 603 combines the foreground component image based on the foreground component supplied from the separation unit 601 and the foreground pixel supplied from the switch 602, and combines The foreground component image is output. Since the foreground area and the mixed area do not overlap, the combining unit 603 combines the foreground component image by applying a logical product operation to the foreground component and the foreground pixel, for example. .
- the synthesizing unit 603 stores an image in which all pixel values are 0 in a built-in frame memory in an initialization process executed at the beginning of the process of synthesizing the foreground component image, and In the synthesis process, the foreground component image is stored (overwritten). Therefore, in the foreground component image output from the combining unit 603, 0 is stored as a pixel value in a pixel corresponding to the background region.
- the synthesis unit 605 synthesizes a background component image based on the component corresponding to the background supplied from the separation unit 601 and the pixel corresponding to the background supplied from the switch 604, and synthesizes them. Outputs a background component image. Since the background region and the mixed region do not overlap, the combining unit 605 combines the background component image by applying a logical product operation to the component corresponding to the background and the pixel corresponding to the background, for example.
- the synthesizing unit 605 stores an image in which all pixel values are 0 in a built-in frame memory in an initialization process executed at the beginning of the background component image synthesizing process.
- the background component image is stored (overwritten) in the composition process. Accordingly, in the background component image output from the combining unit 605, 0 is stored as a pixel value in a pixel corresponding to the foreground area.
- FIG. 51A is a diagram illustrating an input image input to the foreground / background separation unit 105, and a foreground component image and a background component image output from the foreground / background separation unit 105.
- Figure 5 1 B is, c Figure 5 1 illustrates the input image input into the foreground background separator 1 0 5, and the model of the foreground component image and the background component image output from the foreground background separator 1 0 5 A is a schematic diagram of a displayed image
- FIG. 51B is a one-line image including pixels belonging to the foreground area, background area, and mixed area corresponding to FIG. 51A.
- FIG. 4 shows a model diagram in which pixels are developed in the time direction.
- the background component image output from the foreground / background separation unit 105 includes pixels belonging to the background area and the background Is composed of
- the foreground component image output from the foreground / background separator 105 is composed of the pixels belonging to the foreground area and the foreground components included in the pixels of the mixed area. You.
- the pixel values of the pixels in the mixed area are separated into a background component and a foreground component by the foreground / background separation unit 105.
- the separated background component forms a background component image together with the pixels belonging to the background region.
- the separated foreground component forms a foreground component image together with the pixels belonging to the foreground area.
- the pixel value of the pixel corresponding to the background region is set to 0, and meaningful pixel values are set to the pixel corresponding to the foreground region and the pixel corresponding to the mixed region.
- the pixel value of the pixel corresponding to the foreground area is set to 0, and a significant pixel value is set to the pixel corresponding to the background area and the pixel corresponding to the mixed area.
- FIG. 52 is a model of an image showing foreground components and background components of two frames including a foreground corresponding to an object moving from left to right in the figure.
- the foreground motion amount V is 4, and the number of virtual divisions is 4.
- the leftmost pixel and the 14th to 18th pixels from the left consist only of background components and belong to the background area.
- the second through fourth pixels from the left include the background component and the foreground component, and belong to the anchor bird background area.
- the 11th to 13th pixels from the left include the background component and the foreground component, and belong to the covered background region.
- the fifth through tenth pixels from the left consist only of foreground components and belong to the foreground area.
- the first through fifth pixels from the left, and 18 from the left The ith pixel consists of only background components and belongs to the background area.
- the sixth through eighth pixels from the left include the background component and the foreground component, and belong to the uncovered background area.
- the fifteenth through seventeenth pixels from the left include the background component and the foreground component, and belong to the covered background area.
- the ninth to 14th pixels from the left consist only of foreground components and belong to the foreground area.
- FIG. 53 is a diagram illustrating a process of separating a foreground component from pixels belonging to a covered background area.
- 1 to 18 are mixing ratios corresponding to each of the pixels in frame #n.
- the 15th to 17th pixels from the left belong to the covered background area.
- h 15 is the mixture ratio of the 15th pixel from the left of frame #n.
- P15 is the pixel value of the 15th pixel from the left of frame # ⁇ -1.
- Equation (65) the sum ⁇ 5 of the foreground components of the 15th pixel from the left of frame ttn is expressed by Equation (65).
- Equation (6 6) the sum ⁇ 6 of the foreground components of the 16th pixel from the left of frame ⁇ is expressed by Equation (6 6), and the sum ⁇ 7 of the foreground components of the 17th pixel from the left of frame # ⁇ is , which is expressed by equation (67).
- P is the pixel value of the corresponding pixel in the previous frame.
- FIG. 54 is a diagram illustrating a process of separating a foreground component from pixels belonging to an uncovered background area.
- hi 1 through hi 18 are the mixing ratios corresponding to each of the pixels in the frame tin.
- the second through fourth pixels from the left belong to the uncovered background area.
- the pixel value C02 of the second pixel from the left of frame #n is represented by Expression (69).
- 2 is the mixture ratio of the second pixel from the left of frame ⁇ .
- ⁇ 02 is the pixel value of the second pixel from the left of frame # ⁇ + 1.
- Equation (70) the sum f02 of the foreground components of the second pixel from the left of frame # ⁇ is expressed by Equation (70).
- equation (7 1) the sum f03 of the foreground components of the third pixel from the left of frame # ⁇ is expressed by equation (7 1)
- equation f04 of the foreground components of the fourth pixel from the left of frame ⁇ ⁇ is expressed by equation (7 2).
- the foreground component fu included in the pixel value C of the pixel belonging to the uncovered background area is calculated by Expression (73).
- ⁇ is the pixel value of the corresponding pixel in the next frame.
- the separation unit 601 includes information indicating the covered background area and information indicating the uncovered background area included in the area information.
- the foreground component and the background component can be separated from the pixels belonging to the mixed region based on the mixing ratio for each pixel.
- FIG. 55 is a block diagram illustrating an example of a configuration of the separation unit 6001 that performs the above-described processing.
- the image input to the separation section 601 is supplied to the frame memory 621, and area information indicating the covered background area and the uncovered background area supplied from the mixture ratio calculation section 103. , And the mixing ratio H are input to the separation processing block 62.
- the frame memory 6 21 stores the input image in frame units.
- the frame memory 6 2 1 stores the frame 1, which is the frame immediately before the frame tfn, the frame # ⁇ , and the frame ifn + 1 which is the frame after the frame #n. Is stored.
- the frame memory 6 21 supplies the corresponding pixels of the frame # n ⁇ 1, the frame #n, and the frame # n + l to the separation processing block 6 22.
- the separation processing block 62 2 includes the frame #nl, the frame ttn, and the frame supplied from the frame memory 62 1 based on the area information indicating the covered background area and the uncovered background area, and the mixing ratio. Apply the operation described with reference to Figs. 53 and 54 to the pixel value of the corresponding pixel of # n + l. Separate the foreground and background components from the pixels belonging to the mixed area of frame tin. And supplies it to the frame memory 6 23.
- the separation processing block 6 22 includes an uncovered area processing section 6 31, a covered area processing section 6 32, a combining section 6 33, and a combining section 6 3 4.
- the multiplier 6 4 1 of the uncovered area processing unit 6 3 1 multiplies the mixture ratio H by the pixel value of the pixel of the frame ttn + 1 supplied from the frame memory 6 2 1 and outputs the result to the switch 6 4 2 .
- the switch 642 is closed when the pixel of the frame supplied from the frame memory 62 1 (corresponding to the pixel of the frame ⁇ + 1) is in the uncovered background area, and the switch 642 is closed.
- the pixel value multiplied by the mixture ratio supplied from is supplied to the computing unit 643 and the synthesizing unit 634. From Switch 6 4 2
- the value obtained by multiplying the pixel value of the pixel of the input frame ttn + 1 by the mixture ratio H is equal to the background component of the pixel value of the corresponding pixel of the frame ttn + 1.
- Arithmetic unit 643 subtracts the background component supplied from switch 642 from the pixel value of the pixel of frame #n supplied from frame memory 621, and obtains the foreground component.
- the arithmetic unit 634 supplies the foreground component of the pixel of the frame ttn belonging to the anchored background area to the synthesizing unit 633.
- the multiplier 651 of the covered area processing unit 632 multiplies the mixture ratio H by the pixel value of the pixel of the frame ttn-1 supplied from the frame memory 621 and outputs the result to the switch 652.
- the switch 652 is closed when the pixel of the frame ttn (corresponding to the pixel of the frame #nl) supplied from the frame memory 6 21 is a covered background area, and is supplied from the multiplier 651.
- the pixel value multiplied by the mixture ratio a is supplied to the computing unit 653 and the synthesizing unit 6334.
- the value obtained by multiplying the pixel value of the pixel of the frame ttn_l output from the switch 652 by the mixture ratio is equal to the background component of the pixel value of the corresponding pixel of the frame.
- the computing unit 653 subtracts the background component supplied from the switch 652 from the pixel value of the pixel of the frame ⁇ supplied from the frame memory 621, thereby obtaining a foreground component.
- the computing unit 653 supplies the foreground component of the pixel of frame # ⁇ belonging to the covered background area to the combining unit 633.
- the synthesizing unit 633 includes, in the frame, the foreground components of the pixels belonging to the uncovered background area supplied from the computing unit 643, and the covered background area supplied from the computing unit 653.
- the foreground components of the pixels belonging to are synthesized and supplied to the frame memory 6 23.
- the synthesizing unit 634 includes a background component of a pixel belonging to the anchored background area supplied from the switch 642 and a pixel belonging to the covered background area supplied from the switch 652 of the frame.
- the background components of are synthesized and supplied to the frame memory 623.
- the frame memory 6 2 3 stores the frame supplied from the separation processing block 6 2 2 «The foreground component and the background component of the pixel in the mixed region of n are stored respectively.
- the frame memory 623 outputs the stored foreground components of the pixels of the mixed region of the frame # ⁇ and the stored background components of the pixels of the mixed region of the frame ttn.
- the synthesizing unit 603 generates a foreground component image by synthesizing the foreground components of the pixels in the mixed area of the frame ttn and the pixels belonging to the foreground area, which are output from the separating unit 601.
- the reference numeral 605 generates a background component image by synthesizing the background component of the pixel in the mixed region of the frame ⁇ and the pixel belonging to the background region, which are output from the separation unit 601.
- FIG. 56 shows an example of a foreground component image corresponding to frame ⁇ of FIG. Since the leftmost pixel and the 14th pixel from the left consisted of only the background component before the foreground and background were separated, the pixel value was set to 0.
- the second to fourth pixels from the left belong to the covered background area before the foreground and background are separated, the background component is set to 0, and the foreground component is left as it is.
- the 11th to 13th pixels from the left belong to the covered background area, the background component is set to 0, and the foreground component is left as it is.
- the fifth to tenth pixels from the left are left as they are because they consist only of foreground components.
- 'FIG. 56 ⁇ shows an example of a background component image corresponding to frame # ⁇ in FIG. The leftmost pixel and the 14th pixel from the left are left alone because they consisted only of the background component before the foreground and background were separated.
- the second to fourth pixels from the left belong to the covered background area before the foreground and the background are separated, the foreground component is set to 0, and the background component is left as it is.
- the 11th to 13th pixels from the left belong to the covered background area before the foreground and background are separated, the foreground component is set to 0, and the background component is left as it is. 5th to 10th pixels from the left Before the foreground and the background are separated, the pixel consists of only the foreground components, so the pixel value is set to 0.
- step S601 the frame memory 621 of the separation unit 601 acquires the input image, and sets the frame ttn to be separated from the foreground and background into the previous frame ⁇ -l and the subsequent frame tt ⁇ -l. With frame ttn + 1
- step S602 the separation processing block 622 of the separation unit 601 acquires the area information supplied from the mixture ratio calculation unit 103.
- Separating section 6 0 1 of the separation processing block 6 2 2 acquires the supplied mixing ratio a from the mixing ratio calculating unit 1 0 3.
- step S604 the uncovered area processing unit 631 determines the pixel value of the pixel belonging to the uncovered background area supplied from the frame memory 621, based on the area information and the mixture ratio. , The background components are extracted.
- step S605 the uncovered area processing unit 631 executes the processing of the pixels belonging to the uncovered background area supplied from the frame memory 621, based on the area information and the mixture ratio ⁇ . Extract the foreground component from the value.
- step S606 the covered area processing unit 632 calculates the pixel value of the pixel belonging to the covered background area supplied from the frame memory 621, based on the area information and the mixing ratio. Extract background components.
- step S607 the covered area processing unit 632 calculates the pixel value of the pixel belonging to the covered background area supplied from the frame memory 621, based on the area information and the mixing ratio. Extract foreground components.
- step S 608 the synthesizing unit 633 3 extracts the foreground components of the pixels belonging to the anchored background area extracted in the processing in step S 605 and the pixels in the processing in step S 607.
- the pixels belonging to the covered background area are combined with the foreground components.
- the synthesized foreground component is supplied to the synthesis unit 603. Further, the synthesizing unit 603 synthesizes the pixels belonging to the foreground area supplied via the switch 602 with the foreground components supplied from the separating unit 601 to generate a foreground component image. To achieve.
- step S609 the synthesizing unit 634 extracts the background components of the pixels belonging to the uncovered background area extracted in the processing of step S604 and the background components of the pixels in the processing of step S606. And the background components of the pixels belonging to the covered background area.
- the synthesized background component is supplied to the synthesis unit 605. Further, the synthesizing unit 605 synthesizes the pixels belonging to the background area supplied via the switch 604 and the background component supplied from the separating unit 601 to generate a background component image. To achieve.
- step S610 the synthesizing unit 603 outputs a foreground component image.
- step S611 the synthesizing unit 605 outputs the background component image, and the process ends.
- the foreground / background separating unit 105 separates the foreground component and the background component from the input image based on the region information and the mixture ratio, and outputs a foreground component image including only the foreground component, and A background component image consisting of only background components can be output. Next, adjustment of the amount of motion blur in the foreground component image will be described.
- FIG. 58 is a block diagram illustrating an example of the configuration of the motion-blur adjusting unit 106.
- the motion vector and its position information supplied from the motion detecting unit 102 are supplied to the processing unit determining unit 801, the modeling unit 802, and the arithmetic unit 805.
- the area information supplied from the area specifying unit 104 is supplied to the processing unit determining unit 801.
- the foreground component image supplied from the foreground / background separation unit 105 is supplied to the adding unit 804.
- the processing unit determination unit 8001 generates a processing unit based on the motion vector, its position information, and area information, and sends the generated processing unit to the modeling unit 8002 and the adding unit 800. Supply.
- the processing unit generated by the processing unit determination unit 801 starts with a pixel corresponding to the covered background area of the foreground component image, as shown in A in FIG. Consecutive pixels arranged in the motion direction up to the pixel corresponding to the covered background area, or consecutive pixels arranged in the motion direction starting from the pixel corresponding to the uncovered background area to the pixel corresponding to the covered background area Is shown.
- the processing unit is composed of, for example, two data: an upper left point (a pixel specified by the processing unit and a position of a pixel located at the leftmost or uppermost position on the image) and a lower right point.
- the modeling unit 802 performs modeling based on the motion vector and the input processing unit. More specifically, for example, the modeling unit 802 includes a plurality of models corresponding to the number of pixels included in a processing unit, the number of virtual divisions of pixel values in the time direction, and the number of foreground components for each pixel. Is stored in advance, and a model that specifies the correspondence between the pixel value and the foreground component as shown in FIG. 60 is selected based on the processing unit and the number of virtual divisions of the pixel value in the time direction. .
- the modeling unit 8002 sets the number of virtual divisions to 5 and is positioned at the leftmost position.
- the pixel contains one foreground component
- the second pixel from the left contains two foreground components
- the third pixel from the left contains three foreground components
- the fourth pixel from the left contains four foreground components.
- the fifth pixel from the left contains five foreground components
- the sixth pixel from the left contains five foreground components
- the seventh pixel from the left contains five foreground components.
- the eighth pixel from the left contains five foreground components
- the ninth pixel from the left contains four foreground components
- the tenth pixel from the left contains three foreground components
- 1 The first pixel contains two foreground components
- the second pixel from the left contains one foreground component
- the modeling unit 8002 determines the model based on the motion vector and the processing unit. May be generated.
- the modeling unit 802 supplies the selected model to the equation generation unit 803.
- the equation generation unit 803 generates an equation based on the model supplied from the modeling unit 802. Referring to the foreground component image model shown in FIG. 60, the number of foreground components is 8, the number of pixels corresponding to the processing unit is 12, the amount of motion V is 5, and the number of virtual divisions is The equation generated by the equation generator 803 when the value is 5 will be described.
- C05 F05 / v + F04 / v + F03 / v + F02 / v + F01 / v (7 8)
- C06 F06 / v + F05 / v + F04 / v + F03 / v + F02 / v (7 9)
- C07 F07 / v + F06 / v + F05 / v + F04 / v + F03 / v (80)
- C08 F08 / v + F07 / v + F06 / v + F05 / v + F04 / v (8 1)
- C09 F08 / v + F07 / v + F06 / v + F05 / v (8 2)
- Equations (86) to (97) show the equations generated by the equation generator 803.
- C03 lFOl / v + 1F02 / v + lF03 / v + 0F04 / v + 0-F05 / v + 0F06 / v + 0F07 / v + 0F08 / v 88)
- C04 l-FOl / v + 1-F02 / v + l-F03 / v + lF04 / v + 0-F05 / v
- C05 lFO1 / v + 1F02 / v + l-F03 / v + lF04 / v + l-F05 / v
- Equations (86) to (97) can also be expressed as Equation (98)
- Ci .aij'Filv (98)
- j indicates the position of the pixel.
- j has a value of any one of 1 to 12.
- I indicates the position of the foreground value.
- i has a value of any one of 1 to 8.
- aij has a value of 0 or 1 corresponding to the values of i and j.
- equation (98) can be expressed as equation (99).
- ej is an error included in the target pixel Cj.
- Equation (99) can be rewritten as equation (100). '
- equation (103) can be derived.
- equation (105) Based on the least squares method, the normal equation generated by the equation generator 8 03 An example is shown in equation (105).
- the error contained in the pixel C can be dispersed.
- the equation generating section 803 supplies the normal equation generated in this way to the adding section 804.
- the adding unit 804 converts the pixel value C included in the foreground component image into an equation of a matrix supplied from the equation generating unit 803 based on the processing unit supplied from the processing unit determining unit 801. Set.
- the adding unit 804 supplies the matrix in which the pixel values C are set to the arithmetic unit 805.
- the arithmetic unit 805 calculates the foreground component Fi / v from which the motion blur has been removed by a processing based on a solution method such as a sweeping-out method (Gauss-Jordan elimination), and calculates the foreground pixel from which the motion blur has been removed. Calculate Fi corresponding to any one of the integers from 0 to 8, i.e., the value of Fi, which is a pixel value from which motion blur has been removed, as shown in FIG. 61.
- the foreground component image is output to the motion blur adding unit 806 and the selecting unit 807.
- each of F01 to F08 is set for each of C03 to C10 in order not to change the position of the foreground component image with respect to the screen. Yes, and can correspond to any position.
- the motion blur adding unit 806 includes a motion blur adjustment amount v ′ having a value different from the motion amount V, for example, a motion blur adjustment amount V ′ having a half value of the motion amount V, and a value irrelevant to the motion amount V.
- the motion blur adding unit 806 divides the foreground pixel value Fi from which the motion blur has been removed by the motion blur adjustment amount v ′ to obtain a foreground component Fi / v ′. Is calculated, and the sum of the foreground components Fi / v 'is calculated to generate a pixel value in which the amount of motion blur is adjusted.
- the pixel value C02 is (F01) / ⁇ ′
- the pixel value C03 is (F01 + F02) / ⁇ ′
- the pixel value C04 is ( FO 1 + F02 + F03) / ⁇ ′
- the pixel: C05 is (F02 + F03 + F04) / v ′.
- the motion blur adding unit 806 supplies the foreground component image in which the amount of motion blur has been adjusted to the selecting unit 807.
- the selection unit 807 is, for example, a calculation unit 807 based on a selection signal corresponding to the user's selection.
- the motion blur adjustment unit 106 can adjust the amount of motion blur based on the selection signal and the motion blur adjustment amount v ′.
- the motion-blur adjusting unit 106 calculates the equation (10 6) Generate an expression for the matrix shown.
- the motion-blur adjusting unit 106 sets an expression corresponding to the length of the processing unit in this way, and calculates Fi, which is a pixel value in which the amount of motion blur has been adjusted. Similarly, for example, when the number of pixels included in the processing unit is 100, an equation corresponding to 100 pixels is generated, and Fi is calculated.
- FIG. 64 is a diagram illustrating another configuration of the motion-blur adjusting unit 106.
- the same parts as those shown in FIG. 58 are denoted by the same reference numerals, and description thereof will be omitted.
- the selection unit 821 based on the selection signal, supplies the input motion vector and its position signal to the processing unit determination unit 8001 and the modeling unit 8002 as they are, or the motion vector. Is replaced by the motion-blur adjustment amount v ', and the motion vector and its position signal whose size has been replaced by the motion-blur adjustment amount v' are processed by the processing unit determination unit 81 and the modeling unit 80.
- Supply 2
- the processing unit determination unit 801 to the calculation unit 805 of the motion blur adjustment unit 106 in FIG. 64 correspond to the values of the motion amount V and the motion blur adjustment amount v ′.
- the motion-blur adjusting unit 106 generates an equation corresponding to the motion amount V and the processing unit, sets the pixel value of the foreground component image in the generated equation, and reduces the amount of motion blur. An adjusted foreground component image is calculated.
- step S801 the processing unit determination unit 801 of the motion blur adjustment unit 106 generates a processing unit based on the motion vector and the area information, and the generated processing unit is modeled by the modeling unit. 802.
- step S802 the modeling unit 802 of the motion-blur adjusting unit 106 selects or generates a model according to the motion amount V and the processing unit.
- step S803 the equation generator 803 creates a normal equation based on the selected model.
- step S804 the adding unit 804 sets the pixel value of the foreground component image in the created normal equation.
- step S805 the adding unit 804 determines whether or not the pixel values of all the pixels corresponding to the processing unit have been set, and determines whether the pixel values of all the pixels corresponding to the processing unit have been set. If it is determined that the setting has not been performed, the process returns to step S804, and the process of setting the pixel value to the normal equation is repeated.
- step S805 If it is determined in step S805 that the pixel values of all the pixels in the processing unit have been set, the process proceeds to step S806, and the arithmetic unit 805 is supplied from the adding unit 804. Based on the normal equation in which the pixel value is set, the pixel value of the foreground with the amount of motion blur adjusted is calculated, and the process ends.
- the motion blur adjusting unit 106 can adjust the amount of motion blur from the foreground image including motion blur based on the motion vector and the area information.
- FIG. 66 is a block diagram illustrating another example of the configuration of the motion-blur adjusting unit 106.
- the motion vector and its position information supplied from the motion detection unit 102 are supplied to the processing unit determination unit 901 and the correction unit 905, and the region information supplied from the region identification unit 104 is , And are supplied to the processing unit determination unit 901.
- the foreground component image supplied from the foreground / background separator 105 is supplied to the calculator 904.
- the processing unit determination unit 901 generates a processing unit based on the motion vector, its position information, and area information, and supplies the generated processing unit to the modeling unit 902 together with the motion vector. I do.
- the modeling unit 902 executes modeling based on the motion vector and the input processing unit. More specifically, for example, the modeling unit 92 includes a plurality of models corresponding to the number of pixels included in the processing unit, the number of virtual divisions of pixel values in the time direction, and the number of foreground components for each pixel. Is stored in advance, and a model that specifies the correspondence between pixel values and foreground components as shown in Fig. 67 is selected based on the processing unit and the number of virtual divisions of pixel values in the time direction. .
- the modeling unit 902 sets the number of virtual divisions to 5, and the leftmost pixel is one. Contains the foreground component, the second pixel from the left contains two foreground components, the third pixel from the left contains three foreground components, and the fourth pixel from the left contains four foreground components. The fifth pixel from the left contains five foreground components, the sixth pixel from the left contains five foreground components, the seventh pixel from the left contains five foreground components, and the eighth from the left.
- Pixel contains five foreground components
- the ninth pixel from the left contains four foreground components
- the 10th pixel from the left contains three foreground components
- the 11th pixel from the left Contains two foreground components
- the second pixel from the left contains one foreground component, for a total of eight foreground components.
- the modeling unit 902 uses the model based on the motion vector and the processing unit. May be generated.
- the equation generator 903 uses the model supplied from the modeling unit 902 to Generate an expression.
- the number of foreground components is 8
- the number of pixels corresponding to the processing unit is 12, and the motion amount V is 5.
- an example of an equation generated by the equation generating unit 93 will be described.
- the pixel value C12 includes only the foreground component F08 / v, as shown in equation (107), and the pixel value C11 includes the foreground component F08 / v and the foreground component Consists of the sum of products of F07 / v. Therefore, the foreground component F07 / v can be obtained by equation (108).
- the foreground components F06 / v to FOl / v can be obtained by Expressions (109) to (114). .
- the equation generation unit 903 generates an equation for calculating a foreground component based on a difference between pixel values, as shown in Expressions (107) to (114).
- the equation generator 903 supplies the generated equation to the calculator 904.
- the calculation unit 904 sets the pixel value of the foreground component image to the equation supplied from the equation generation unit 903, and calculates the foreground component based on the equation in which the pixel value is set. Performance For example, when the equations (107) to (111) are supplied from the equation generator 903, the arithmetic unit 904 calculates the equations (107) to (114) as follows. Set the pixel values C05 to C12.
- the calculation unit 904 calculates a foreground component based on the equation in which the pixel value is set. For example, the arithmetic unit 904 calculates the foreground component FO l as shown in FIG. 68 by an arithmetic operation based on equations (107) to (114) in which pixel values C05 to C12 are set. Calculate / v to F08 / v. The calculation unit 904 supplies the foreground components FO l / v to F08 / v to the correction unit 905.
- the correction unit 905 removes motion blur by multiplying the foreground component supplied from the calculation unit 904 by the motion amount V included in the motion vector supplied from the processing unit determination unit 901. Calculate the pixel value of the foreground. For example, when the foreground components FO l / v to FOS / v supplied from the arithmetic unit 904 are supplied, the correction unit 905 provides, for each of the foreground components FO l / v to F08 / v, By multiplying by the motion amount V of 5, foreground pixel values F01 to F08 from which motion blur has been removed are calculated as shown in FIG.
- the correcting unit 905 supplies the foreground component image composed of the pixel values of the foreground, from which the motion blur has been removed, calculated as described above, to the motion blur adding unit 906 and the selecting unit 907.
- the motion-blur adding unit 9 06 generates a motion-blur adjustment amount v ′ having a value different from the motion amount V, for example, a motion-blur adjustment amount v ′ having a half value of the motion amount V,
- the amount of motion blur can be adjusted with the motion blur adjustment amount v '. For example, as shown in FIG.
- the motion-blur adding unit 906 divides the foreground pixel value Fi from which the motion blur has been removed by the motion-blur adjustment amount v ′ to obtain a foreground component Fi / v ′. Is calculated, and the sum of the foreground components Fi / v 'is calculated to generate a pixel value in which the amount of motion blur is adjusted.
- the pixel value C02 is (F01) / v'
- the pixel value C03 is (F01 + F02) / v '
- the pixel value C04 is (F01 + F02 + F03) / v '
- the pixel value C05 is (F02 + F03 + F04) / v'.
- the motion blur adding unit 906 supplies the foreground component image in which the amount of motion blur has been adjusted to the selecting unit 907.
- the selection unit 907 supplies the foreground component image from which the motion blur has been removed supplied from the correction unit 905 and the motion blur addition unit 906 based on a selection signal corresponding to the user's selection, for example.
- One of the selected foreground component images with the adjusted amount of motion blur is selected, and the selected foreground component image is output.
- the motion blur adjustment unit 106 can adjust the amount of motion blur based on the selection signal and the motion blur adjustment amount v ′.
- step S901 the processing unit determination unit 901 of the motion blur adjustment unit 106 generates a processing unit based on the motion vector and the area information, and converts the generated processing unit into a modeling unit. It is supplied to 902 and the collection section 905.
- step S902 the modeling unit 902 of the motion-blur adjusting unit 106 selects or generates a model according to the motion amount V and the processing unit.
- step S903 the equation generation unit 903 generates an equation for calculating a foreground component from a difference between pixel values of the foreground component image based on the selected or generated model.
- step S904 the arithmetic unit 904 sets the pixel value of the foreground component image in the created equation, and calculates the foreground component from the pixel value difference based on the equation in which the pixel value is set. Extract.
- step S905 the arithmetic unit 904 determines whether all foreground components corresponding to the processing unit have been extracted, and has not extracted all foreground components corresponding to the processing unit. If the judgment is made, the process returns to step S904, and the process of extracting the foreground component is repeated.
- step S905 If it is determined in step S905 that all the foreground components corresponding to the processing unit have been extracted, the process proceeds to step S906, where the correction unit 905 determines the operation unit based on the motion amount V.
- the correction unit 905 determines the operation unit based on the motion amount V.
- Each of the foreground components FO l / v to FOS / v supplied from 904 is corrected to calculate foreground pixel values F01 to F08 from which motion blur has been removed.
- step S907 the motion-blur adding unit 906 calculates the pixel value of the foreground with the amount of the motion blur adjusted, and the selecting unit 907 selects the image or the motion-blurred image. One of the images in which the amount of motion blur has been adjusted is selected, the selected image is output, and the process ends.
- the motion-blur adjusting unit 106 shown in FIG. 66 can adjust the motion blur from the foreground image including the motion blur with a simpler operation and more quickly.
- the image processing apparatus having the configuration shown in FIG. 2 can adjust the amount of motion blur included in the input image.
- the mixing ratio H has been described as the ratio of the background component included in the pixel value, it may be the ratio of the foreground component included in the pixel value.
- an example is given in which an image of the real space having the three-dimensional space and the time axis information is projected onto the two-dimensional space and the time and space having the time axis information using the video camera.
- Is not limited to this example and corrects the distortion caused by the projection of more first information in the first dimension onto the second information in the smaller second dimension. It can be used to extract significant information, or to synthesize images more naturally.
- the sensor is not limited to a CCD, but may be a solid-state imaging device.
- the sensor may be a sensor in which the detection elements are arranged in a matrix, but may be a sensor in which the detection elements are arranged in a line.
- the recording medium on which the program for performing the signal processing of the present invention is recorded as shown in FIG.
- a magnetic disk 51 including a floppy (registered trademark) disk
- an optical disk 52 CD-ROM (Compact Di- sc-Read Only Memory)
- DVD including Digital Versatile Disk
- magneto-optical disk 53 including MD (Mini-Disc) (trademark)
- semiconductor media 54 instead, it is provided with the R0M 22 in which the program is recorded, which is provided to the user in a state in which it is previously installed in the computer, and the hard disk included in the storage unit 28.
- steps for describing a program to be recorded on a recording medium are not only performed in chronological order according to the order described, but are not necessarily performed in chronological order. Alternatively, it also includes processes that are individually executed.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
- Studio Circuits (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Description
Claims
Priority Applications (5)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| KR1020037002804A KR100894923B1 (ko) | 2001-06-27 | 2002-06-26 | 화상 처리 장치, 방법 및 기록 매체와 촬상 장치 |
| CA002420069A CA2420069C (en) | 2001-06-27 | 2002-06-26 | Image processing apparatus and method, and image pickup apparatus |
| US10/362,558 US7181080B2 (en) | 2001-06-27 | 2002-06-26 | Image processing apparatus and method, and image pickup apparatus |
| EP02741315A EP1400925A4 (en) | 2001-06-27 | 2002-06-26 | IMAGE PROCESSING APPARATUS AND METHOD AND IMAGING APPARATUS |
| MXPA03001576A MXPA03001576A (es) | 2001-06-27 | 2002-06-26 | Aparato y metodo de procesamiento de imagenes y aparato de captura de imagenes. |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2001-194611 | 2001-06-27 | ||
| JP2001194611A JP4840630B2 (ja) | 2001-06-27 | 2001-06-27 | 画像処理装置および方法、記録媒体、並びにプログラム |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2003003307A1 true WO2003003307A1 (en) | 2003-01-09 |
Family
ID=19032715
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2002/006385 Ceased WO2003003307A1 (en) | 2001-06-27 | 2002-06-26 | Image processing apparatus and method, and image pickup apparatus |
Country Status (8)
| Country | Link |
|---|---|
| US (1) | US7181080B2 (ja) |
| EP (1) | EP1400925A4 (ja) |
| JP (1) | JP4840630B2 (ja) |
| KR (1) | KR100894923B1 (ja) |
| CN (1) | CN1313975C (ja) |
| CA (1) | CA2420069C (ja) |
| MX (1) | MXPA03001576A (ja) |
| WO (1) | WO2003003307A1 (ja) |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4596220B2 (ja) * | 2001-06-26 | 2010-12-08 | ソニー株式会社 | 画像処理装置および方法、記録媒体、並びにプログラム |
| JP4192719B2 (ja) * | 2003-08-22 | 2008-12-10 | ソニー株式会社 | 画像処理装置および方法、並びにプログラム |
| CN100423557C (zh) * | 2004-02-13 | 2008-10-01 | 索尼株式会社 | 图像处理装置以及图像处理方法 |
| WO2005079061A1 (ja) * | 2004-02-13 | 2005-08-25 | Sony Corporation | 画像処理装置と画像処理方法およびプログラム |
| US7710498B2 (en) | 2004-02-13 | 2010-05-04 | Sony Corporation | Image processing apparatus, image processing method and program |
| US8265392B2 (en) * | 2006-02-07 | 2012-09-11 | Qualcomm Incorporated | Inter-mode region-of-interest video object segmentation |
| US8265349B2 (en) * | 2006-02-07 | 2012-09-11 | Qualcomm Incorporated | Intra-mode region-of-interest video object segmentation |
| US8150155B2 (en) | 2006-02-07 | 2012-04-03 | Qualcomm Incorporated | Multi-mode region-of-interest video object segmentation |
| KR101225062B1 (ko) * | 2006-03-10 | 2013-01-23 | 삼성전자주식회사 | 이미지 프레임의 선택적 출력 장치 및 방법 |
| KR101299249B1 (ko) * | 2008-08-29 | 2013-08-22 | 삼성테크윈 주식회사 | 디지털 촬영장치, 그 제어방법 및 제어방법을 실행시키기 위한 프로그램을 저장한 기록매체 |
| CN101916449A (zh) * | 2010-08-21 | 2010-12-15 | 上海交通大学 | 建立图像处理中基于运动信息的背景模型的方法 |
| US10262421B2 (en) * | 2014-08-04 | 2019-04-16 | Nec Corporation | Image processing system for detecting stationary state of moving object from image, image processing method, and recording medium |
| CN113223043B (zh) * | 2021-03-26 | 2024-07-19 | 西安闻泰信息技术有限公司 | 一种移动目标的检测方法、装置、设备及介质 |
| CN113129227A (zh) * | 2021-03-29 | 2021-07-16 | 影石创新科技股份有限公司 | 图像处理方法、装置、计算机设备和存储介质 |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH06113338A (ja) * | 1992-09-30 | 1994-04-22 | Fujitsu Ltd | 多眼式立体映像の動き補償予測方法 |
| JPH07336688A (ja) * | 1994-06-06 | 1995-12-22 | Nippon Hoso Kyokai <Nhk> | アンカバー領域の検出方法 |
| JPH10164436A (ja) * | 1996-12-04 | 1998-06-19 | Sony Corp | 輪郭抽出装置、輪郭抽出方法、キー信号生成装置及びキー信号生成方法 |
| EP0933727A2 (en) * | 1998-01-29 | 1999-08-04 | Canon Kabushiki Kaisha | Image information processing apparatus and its method |
| JP2001250119A (ja) * | 1999-12-28 | 2001-09-14 | Sony Corp | 信号処理装置および方法、並びに記録媒体 |
| JP2002190028A (ja) * | 2000-12-21 | 2002-07-05 | Sony Corp | 信号処理装置および方法、並びに記録媒体 |
| JP2002190015A (ja) * | 2000-12-21 | 2002-07-05 | Sony Corp | 画像処理装置および方法、並びに記録媒体 |
| JP2002190016A (ja) * | 2000-12-21 | 2002-07-05 | Sony Corp | 信号処理装置および方法、並びに記録媒体 |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6054999A (en) * | 1988-03-22 | 2000-04-25 | Strandberg; Oerjan | Method and apparatus for computer supported animation |
| GB2231752B (en) * | 1989-04-27 | 1993-08-04 | Sony Corp | Motion dependent video signal processing |
| JPH05153493A (ja) * | 1991-11-27 | 1993-06-18 | Matsushita Electric Ind Co Ltd | 映像信号合成装置 |
| US5812787A (en) * | 1995-06-30 | 1998-09-22 | Intel Corporation | Video coding scheme with foreground/background separation |
| US6259827B1 (en) * | 1996-03-21 | 2001-07-10 | Cognex Corporation | Machine vision methods for enhancing the contrast between an object and its background using multiple on-axis images |
| JP2952226B2 (ja) * | 1997-02-14 | 1999-09-20 | 日本電信電話株式会社 | 動画像の予測符号化方法および復号方法、動画像予測符号化または復号プログラムを記録した記録媒体、および、動画像予測符号化データを記録した記録媒体 |
| US5933151A (en) * | 1997-03-26 | 1999-08-03 | Lucent Technologies Inc. | Simulated natural movement of a computer-generated synthesized talking head |
| JP2000030040A (ja) * | 1998-07-14 | 2000-01-28 | Canon Inc | 画像処理装置及びコンピュータ読み取り可能な記憶媒体 |
| US6771834B1 (en) * | 1999-07-02 | 2004-08-03 | Intel Corporation | Method for segmenting a digital image |
-
2001
- 2001-06-27 JP JP2001194611A patent/JP4840630B2/ja not_active Expired - Fee Related
-
2002
- 2002-06-26 CN CNB028027655A patent/CN1313975C/zh not_active Expired - Fee Related
- 2002-06-26 WO PCT/JP2002/006385 patent/WO2003003307A1/ja not_active Ceased
- 2002-06-26 US US10/362,558 patent/US7181080B2/en not_active Expired - Fee Related
- 2002-06-26 MX MXPA03001576A patent/MXPA03001576A/es active IP Right Grant
- 2002-06-26 KR KR1020037002804A patent/KR100894923B1/ko not_active Expired - Fee Related
- 2002-06-26 CA CA002420069A patent/CA2420069C/en not_active Expired - Fee Related
- 2002-06-26 EP EP02741315A patent/EP1400925A4/en not_active Withdrawn
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH06113338A (ja) * | 1992-09-30 | 1994-04-22 | Fujitsu Ltd | 多眼式立体映像の動き補償予測方法 |
| JPH07336688A (ja) * | 1994-06-06 | 1995-12-22 | Nippon Hoso Kyokai <Nhk> | アンカバー領域の検出方法 |
| JPH10164436A (ja) * | 1996-12-04 | 1998-06-19 | Sony Corp | 輪郭抽出装置、輪郭抽出方法、キー信号生成装置及びキー信号生成方法 |
| EP0933727A2 (en) * | 1998-01-29 | 1999-08-04 | Canon Kabushiki Kaisha | Image information processing apparatus and its method |
| JP2001250119A (ja) * | 1999-12-28 | 2001-09-14 | Sony Corp | 信号処理装置および方法、並びに記録媒体 |
| JP2002190028A (ja) * | 2000-12-21 | 2002-07-05 | Sony Corp | 信号処理装置および方法、並びに記録媒体 |
| JP2002190015A (ja) * | 2000-12-21 | 2002-07-05 | Sony Corp | 画像処理装置および方法、並びに記録媒体 |
| JP2002190016A (ja) * | 2000-12-21 | 2002-07-05 | Sony Corp | 信号処理装置および方法、並びに記録媒体 |
Non-Patent Citations (1)
| Title |
|---|
| See also references of EP1400925A4 * |
Also Published As
| Publication number | Publication date |
|---|---|
| EP1400925A1 (en) | 2004-03-24 |
| US20040028259A1 (en) | 2004-02-12 |
| CN1471694A (zh) | 2004-01-28 |
| CA2420069C (en) | 2010-03-09 |
| JP4840630B2 (ja) | 2011-12-21 |
| US7181080B2 (en) | 2007-02-20 |
| CN1313975C (zh) | 2007-05-02 |
| EP1400925A4 (en) | 2007-05-09 |
| JP2003008991A (ja) | 2003-01-10 |
| MXPA03001576A (es) | 2003-06-04 |
| KR100894923B1 (ko) | 2009-04-27 |
| KR20040014951A (ko) | 2004-02-18 |
| CA2420069A1 (en) | 2003-02-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP4596222B2 (ja) | 画像処理装置および方法、記録媒体、並びにプログラム | |
| JP4596202B2 (ja) | 画像処理装置および方法、並びに記録媒体 | |
| JP4596220B2 (ja) | 画像処理装置および方法、記録媒体、並びにプログラム | |
| JP4596216B2 (ja) | 画像処理装置および方法、記録媒体、並びにプログラム | |
| JP4596221B2 (ja) | 画像処理装置および方法、記録媒体、並びにプログラム | |
| JP4106874B2 (ja) | 画像処理装置および方法、並びに記録媒体 | |
| JP4596203B2 (ja) | 画像処理装置および方法、記録媒体、並びにプログラム | |
| WO2003003306A1 (en) | Image processing apparatus and method | |
| JP4674408B2 (ja) | 画像処理装置および方法、記録媒体、並びにプログラム | |
| WO2003003307A1 (en) | Image processing apparatus and method, and image pickup apparatus | |
| JP4596223B2 (ja) | 画像処理装置および方法、記録媒体、並びにプログラム | |
| JP4596219B2 (ja) | 画像処理装置および方法、記録媒体、並びにプログラム | |
| JP4596215B2 (ja) | 画像処理装置および方法、記録媒体、並びにプログラム | |
| JP4596217B2 (ja) | 画像処理装置および方法、記録媒体、並びにプログラム | |
| JP4596209B2 (ja) | 画像処理装置および方法、記録媒体、並びにプログラム | |
| JP4596218B2 (ja) | 画像処理装置および方法、記録媒体、並びにプログラム | |
| JP4596205B2 (ja) | 画像処理装置および方法、並びにプログラム | |
| JP4596225B2 (ja) | 画像処理装置および方法、記録媒体、並びにプログラム | |
| WO2003001456A1 (en) | Image processing apparatus and method, and image pickup apparatus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| AK | Designated states |
Kind code of ref document: A1 Designated state(s): CA CN IN KR MX SG US |
|
| AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 210/MUMNP/2003 Country of ref document: IN |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2002741315 Country of ref document: EP |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2420069 Country of ref document: CA |
|
| WWE | Wipo information: entry into national phase |
Ref document number: PA/a/2003/001576 Country of ref document: MX |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 1020037002804 Country of ref document: KR |
|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
| WWE | Wipo information: entry into national phase |
Ref document number: 028027655 Country of ref document: CN |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 10362558 Country of ref document: US |
|
| WWP | Wipo information: published in national office |
Ref document number: 1020037002804 Country of ref document: KR |
|
| WWP | Wipo information: published in national office |
Ref document number: 2002741315 Country of ref document: EP |

