WO2003063082A1 - Moving picture search apparatus - Google Patents
Moving picture search apparatus Download PDFInfo
- Publication number
- WO2003063082A1 WO2003063082A1 PCT/JP2003/000709 JP0300709W WO03063082A1 WO 2003063082 A1 WO2003063082 A1 WO 2003063082A1 JP 0300709 W JP0300709 W JP 0300709W WO 03063082 A1 WO03063082 A1 WO 03063082A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- moving image
- database
- feature amount
- search
- input
- 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
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/783—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/7847—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using low-level visual features of the video content
- G06F16/785—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using low-level visual features of the video content using colour or luminescence
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
Definitions
- the present invention relates to a moving image search method for efficiently searching for moving image information in an information use environment, and a lighting device therefor.
- an image search method a search using a still image is performed, and a feature amount such as a luminance value or a variance value for each frame may be used.
- a feature amount such as a luminance value or a variance value for each frame.
- a threshold that is a reference for matching is set according to the input data in order to perform appropriate matching. It had to be changed and the process was complicated.
- the conventional image matching is a search method for the entire image, and has a problem to be solved that requires a large amount of calculation processing.
- the present invention realizes a moving image search method that performs high-speed processing with a small amount of information by calculating a feature amount of a moving image.
- means for calculating a feature amount of a moving image in units of local regions means for calculating a histogram of local region feature amounts of a moving image, and calculating a correlation value between before and after a frame.
- the threshold value can be easily set without depending on the input data as compared with the average value and the variance value of the luminance value of each frame. Disclosure of the invention
- the present invention provides a search moving image input unit for inputting search target moving image data for search; calculating a feature amount for the input search target moving image Area extraction unit that extracts one or more rectangular or arbitrary-shaped local areas to perform the search, and the brightness signal, color difference signal, and RGB color of the feature calculation area extracted from the input video to be searched.
- a signal extraction unit that extracts image signal components such as spatial components, XYZ color space components, uniform color space components, and Munsell color space components, or converts and extracts these signals.
- a feature amount table creation processing unit that creates a feature amount table to be used in a search process from feature amounts calculated based on a moving image; and a feature amount calculation unit including: moving image data for a database for a database Enter A moving image database registration unit for registering moving image data for the input database in the moving image database; and a moving image for registering moving image data for the input database.
- a local region extracting unit for extracting one or more rectangular or arbitrary-shaped local regions for calculating a feature amount with respect to the inputted database registration moving image; and a database for input database registration. Extract the luminance signal, color difference signal, RGB color space components, XYZ color space components, uniform color space components, Munsell color space components, etc.
- a signal extraction unit for converting and extracting a feature amount, and a feature amount table used for a search process from a feature amount calculated based on the input moving image for database registration.
- a table database registration processing unit for registering the created database feature table in the feature table database;
- a database for registering a feature amount table for the database; a process of reading out a feature amount table for a database used in a search process after receiving a process of a feature amount calculating unit of the input moving image to be searched;
- a database read-out processing unit for performing a search, a feature amount table created based on the input video to be searched, and a feature amount table data.
- the feature amount used for matching judgment is extracted and output to the matching result judging unit, and the matching result from the matching result judging unit is stored.
- a matching determination based on a certain threshold is performed.
- a search processing unit including a matching result determination unit that outputs a matching result to the matching processing unit; and receives a matching result of the matching result determination unit, and reads out a search result moving image corresponding to the result from the moving image database.
- a search result output unit that outputs a search result moving image, and a moving image search method and a device thereof.
- a local part is extracted from the image frame, and a histogram of luminance information calculated from the local part is calculated. Then, the calculated histogram is compared between before and after the frame, and the correlation value is obtained. By grouping the correlation values according to a certain threshold value, faster moving image matching can be realized.
- the correlation value does not depend on the input data as compared with the average value or the variance value of the luminance value of each frame, and does not need to be changed with respect to the threshold value input when performing grouping
- the correlation value is used as a feature value.
- An appropriate threshold value can be set.
- the grouping feature amount as a search parameter, the number of parameters used for searching for a moving image scene can be suppressed, so that the matching processing time can be shortened.
- FIG. 1 is a block diagram showing the basic principle of the present invention.
- FIG. 2 is a schematic block diagram of a hardware configuration by a single terminal for realizing the present invention.
- FIG. 3 is a schematic block diagram of a hardware configuration of a plurality of terminals for realizing the present invention.
- FIG. 4 is a diagram showing a local region when there is an image of the n-th frame and an image of the (n + 1) -th frame.
- FIG. 5 is an example showing a region in which one or more feature values are calculated for a moving image in units of local regions.
- FIG. 6 is an explanatory diagram for actually extracting a local region.
- FIG. 7 is a schematic diagram illustrating signal extraction in a color space.
- FIG. 8 is a schematic diagram in which the correlation value calculated from the signal information is subjected to grouping processing as a temporal feature amount.
- FIG. 9 is a schematic diagram for creating a feature amount table.
- FIG. 10 is a block diagram showing the search processing of the present invention.
- FIG. 11 is a flowchart showing the input side to be searched and the processing procedure of the matching process.
- FIG. 12 is a flowchart showing a processing procedure on the database side serving as feature amount information for comparison.
- FIG. 13 is a block diagram illustrating a search process according to the embodiment in which the order of the local region extraction process and the signal extraction process is changed.
- FIG. 14 is a flowchart showing the input side to be searched and the processing procedure of the matching process.
- FIG. 15 is a flowchart showing a processing procedure on the database side as feature amount information for comparison.
- FIG. 1 is a block diagram showing the basic principle of the present invention, which will be described.
- a feature amount is calculated based on this moving image information, and a feature amount table 2 for input is created.
- a feature amount table 2 for input is created for a moving image for a database.
- a plurality of feature amount tables 3 are created in advance for registration and stored.
- the created input feature quantity table 2 is compared with the accumulated registration feature quantity table 3 (matching processing âmatching judgmentâ), and the reference is made.
- the search result list is sent to the moving image database 4.
- a moving image is output from the moving image database 4 based on this list, and a desired moving image 5 is obtained.
- the input / output interface 8 is mainly recorded in the external recording device 11.
- the moving image information is designated by a mouse (input device) 7, and the feature amount is calculated from the main moving image information by the memory (internal recording device) 12 and the central processing unit 9.
- This feature data is recorded by the internal 12 or the external recording device 11 via the input / output interface 8.
- the moving image information recorded in the external recording device 11, centering on the input / output interface 8 is stored in the memory (internal recording device) 12 and the central processing unit 9 by the feature amount. Is calculated.
- This feature data is recorded by the internal 12 or the external recording device 11 via the input / output interface 8.
- the calculation data is read from the recording device 11 to the memory 12, a search process is performed by the central processing unit 9, and the result is sent to the internal recording device 12 or the external recording device 11 via the input / output interface 8. send.
- the central processing unit 9 reads the search result image list from the internal recording device 12 or the external recording device 11 via the input / output interface 8 and outputs the search result image to the external recording device 1 1 via the input / output interface 8.
- the search result image is recorded on the external recording device 11 via the input / output interface 8 or output to the display device 10.
- search target moving image about the input and output interface 1 7, in the moving image information recorded in the external recording device 2 0, or on the network output interface 1 7 (code 2 1, 2 2)
- the moving image recorded on the external device 23 of the connected server 14 is specified by the mouse (input device) 16 of the client 13, and the memory (internal recording device) of the client 13 is specified.
- the feature amount is calculated from the main moving image information by the central processing unit 18 and the central processing unit 18.
- This feature data is Internal 24 via the input / output interface 17 on the client side 13 or the external recording device 11, or internal 29 or the external recording device 2 via the input / output interface 26 on the server 14 3 or send it to the search server side 14 and record it in the internal 29 or the external recording device 23 via the input / output interface 26.
- the moving image information recorded in the external recording device 23 centering on the input / output interface 26 of the server side 14 is centrally processed by the memory (internal recording device) 29.
- the device 27 calculates a feature value.
- the characteristic amount data is recorded by the internal unit 29 or the external recording unit 23 via the input / output interface 26.
- the calculation data is read from the recording device 23 to the memory 29, a search process is performed by the central processing unit 27, and the result is sent to the internal recording device 29 or the external recording device 23 via the input / output interface 26. send.
- the central processing unit 27 reads the search result image list from the internal recording device 29 or the external recording device 23 via the input / output interface 26, and externally records the search result images via the input / output interface 26. Output from the device 23, the search result image is returned to the client terminal 13 via the input / output interface 26 of the search server 14 (reference numeral 22), and the external recording of the client terminal 13 is performed. The information is recorded on the device 20 or output to the display device 19.
- FIGS. 4 to 7 are schematic diagrams showing calculation of a feature amount which is a main part of the present invention.
- a local part is extracted from an image frame, and a histogram of luminance information calculated from the local part is calculated.
- the calculated histogram is compared before and after the frame, and the correlation value is obtained.
- the purpose is to realize faster moving image matching by grouping correlation values by a certain threshold.
- the correlation value does not depend on the input as compared with the average value or the variance value of the luminance value of each frame, and it is not necessary to change the threshold when performing grouping with respect to the input. By using, search can be performed more effectively.
- a change in the histogram is obtained for a moving image before and after a frame in the time direction, so that at present, a search can be performed using feature amounts from the entire screen.
- the search can be performed using the local histogram correlation value even in the scene of the minute fluctuation that could not be detected, the scene of the camera burn, the zoom scene, and the like.
- FIG. 4 shows a case where there is an image of the n-th frame and an image of the (n + 1) -th frame. For example, between a local region within the image xl, x2, x 3, x 4, x5 and yl, y2, y 3, y 4, y5, and calculates a luminance value histogram between local regions, respectively Find the correlation value.
- the correlation value distribution is calculated from the input video information.
- the obtained correlation value distribution is classified into two with a certain threshold, and grouping is performed using continuity in the time direction, and the number of structures is used as a matching parameter.
- the moving image search method using this matching parameter and its processing device are superior in terms of processing speed and search accuracy.
- FIG. 5 is an example of a region in which one or more feature values are calculated for a moving image in units of local regions. As described above, it is possible to calculate the feature amount in local area units without depending on the aspect ratio of the image frame.
- FIG. 6 is an explanatory diagram when a local region is actually extracted.
- one 8 â 8 pixel block (only X3) is extracted, or a plurality of 8 â 8 pixel blocks, for example, 5 It can also be extracted at three (xl to x5) positions.
- the shape of this pixel block is rectangular in this example, but is not limited to this. Any shape such as a circle or polygon may be used, and N pixels may be extracted from the entire area. It is also possible to move five positions in the local area in the figure.
- Fig. 8 shows an example of using a certain threshold based on a signal, a histogram, or a correlation value calculated from the input moving image information, or a feature amount such as a grouping process thereof.
- the process is divided into eight loops, and the time length (the number of frames) of the group, the total number of groups, threshold values, etc. are created as a feature amount table.
- FIG. 9 is a grouping of FIG. 8 by a plurality of threshold values. Perform processing to divide into two or more groups by two or more thresholds (Th1, Th2, Th3 in Fig. 9), and characterize the time length (number of frames), total number of groups, thresholds, etc. of the groups It is created as a quantity table.
- the matching between the feature information of the input video to be searched and the feature information extracted from the database for comparison corresponding to the feature information is performed by, for example, Equation (1) or Equation (2).
- the determination is made according to the following determination formula.
- FIG. 10 is a block diagram showing the search processing of the present invention.
- FIG. 11 is a flowchart showing the input side to be searched and the processing procedure of the matching processing.
- FIG. 12 is feature amount information for comparison.
- 9 is a flowchart illustrating a processing procedure on the database side.
- step 101 search moving image information is input. With respect to the search moving image information input in step 102, it is determined whether or not the extraction processing is performed with a part of the frame of the moving image as a local region. If the local region extraction is skipped (no extraction is performed), the process proceeds to step 104. If the local region extraction is not skipped (extraction is performed), the process proceeds to step 103.
- step 103 one or more rectangular or arbitrary local shapes are extracted in order to extract a part of a frame of the moving image as a local area in order to calculate a feature amount with respect to the input moving image information for search.
- a region extraction process is performed.
- step 104 the luminance signal, color difference signal, each component of the RGB color space, and the like of the entire local region extracted in step 103 or the input moving image information for search are extracted.
- step 105 it is determined whether or not to calculate a histogram of the signal extracted from the input moving image information for search. If the histogram calculation is skipped (not calculated), the process proceeds to step 107. If the histogram calculation is not skipped (calculated), the process proceeds to step 106.
- step 106 a histogram of the signal extracted in step 104 is calculated.
- step 107 it is determined whether the signal extracted in step 104 or the histogram calculated in step 106 is capable of calculating a correlation value. If the calculation of the correlation value is skipped (not calculated), the process proceeds to step 109. If the calculation of the correlation value is not skipped (calculated), the process proceeds to step 108.
- the correlation value is calculated between adjacent frames or between arbitrary frames along the time direction.
- step 109 it is determined whether or not to perform the grouping process on the correlation value in step 108 or the histogram in step 106. If the grouping process is skipped (not performed), the process proceeds to step 111. If the grouping process is not skipped (processed), the process proceeds to step 110.
- step 110 the correlation value of step 108 is subjected to a grouping process in which the number of frames in the time direction is 1 or more based on a certain threshold value, and the number of frames, the number of groups, etc. are calculated as feature amounts. I do.
- step 111 a feature amount table used for search processing is created from the feature amounts calculated based on the input moving image information for search.
- step 112 the process of calculating the feature amount of the input moving image information is performed, and the process of reading out the feature amount table for the database used for the search process is performed.
- step 113 feature values used for matching judgment are extracted from the feature table registered in the database based on the feature table used in the search process created in step 111.
- step 114 matching is determined based on the feature amount of the moving image to be searched and the feature amount of the moving image in the database. If they match, the process proceeds to step 115. If they do not match, the process proceeds to step 116. In steps 1 1 and 5, Save the matching result. In step 116, it is determined whether or not to end the matching process. If the processing is to be ended, the processing proceeds to step 1 17. If not, the processing returns to the processing in step 113.
- step 117 upon receiving the matching result, a process of reading a search result moving image corresponding to the result from the moving image database is performed.
- step 118 the search result moving image read by the read processing unit is output. It should be noted that all of the four skip processes of step 102, step 105, step 107 and step 109 in FIG. 11 are not executed.
- step 201 the moving image information for database registration is input, and in step 202, the input moving image information for database registration is registered in the moving image database.
- step 203 it is determined whether or not extraction processing is to be performed on the database registration moving image information by using a part of the 'frame of the moving image as a local region. If the local region extraction is skipped (no extraction), the process proceeds to step 205. If the local region extraction is not skipped (extraction), the process proceeds to step 204.
- one or more rectangular shapes or arbitrary shapes are required to extract a part of the frame of the moving image as a local region in order to calculate a feature amount of the moving image information for database registration. Performs a local region extraction process.
- step 205 the luminance signal, color difference signal, RGB color space components, and the like of the local area extracted in step 204 or the entire moving image information for database registration are extracted.
- step 206 it is determined whether or not to calculate a histogram of the signal extracted from the moving image information for database registration. If the histogram calculation is skipped (not calculated), the process proceeds to step 208. If the histogram calculation is not skipped (calculated), the process proceeds to step 207.
- step 205 a histogram of the signal extracted in step 205 is calculated.
- step 208 it is determined whether the signal extracted in step 205 or the histogram calculated in step 207 is capable of calculating a correlation value. If the calculation of the correlation value is skipped (not calculated), the process proceeds to step 210. If the calculation of the correlation value is not skipped (calculated), the process proceeds to step 209. This correlation value is The correlation value is calculated between the frames obtained or between arbitrary frames along the time direction.
- step 210 it is determined whether the correlation value in step 209 or the histogram in step 207 is to be subjected to the grouping process. If the grouping processing is skipped (not performed), the process proceeds to step 211. If the grouping process is not skipped (processed), the process proceeds to step 211.
- step 211 the correlation value of step 209 is subjected to a grouping process in units of one or more frames in the time direction based on a certain threshold, and the number of frames, the number of groups, etc. Is calculated as
- a feature amount table calculated based on the moving image information for database registration and a feature amount table used for the search processing are created.
- the feature amount template for the database used for the search processing is registered in the feature amount template database. It should be noted that all of the four skip processes of step 203, step 206, step 208 and step 210 in FIG. 12 are not executed.
- FIG. 13 is a block diagram showing a search process according to an embodiment in which the order of the local region extraction process and the signal extraction process is changed.
- FIG. 14 is a flowchart showing a processing procedure of the input side to be searched and the matching process.
- FIG. 15 is a flowchart showing a processing procedure on the database side as feature amount information for comparison. The processing procedure on the input side will be described with reference to FIG. 14.
- step 301 the moving image information for search is input, and the moving image information for search input in step 302 is input. Extracts luminance signal, color difference signal, each component of RGB color space, etc. of video information.
- step 303 it is determined whether or not the extraction process is performed with a part of the frame of the input moving image as a local region. If the local region extraction is skipped (no extraction), the process proceeds to step 304. If the local region extraction is not skipped (extraction), the process proceeds to step 304. 9
- step 304 one or more rectangular shapes or arbitrary shapes are extracted to extract a part of the frame of the moving image as a local region in order to calculate the feature amount for the input moving image information for search.
- a local area extraction process is performed on the shape.
- step 300 the histogram of the extracted signal is obtained for the local region extracted in step 304 or the entire luminance signal, color difference signal, and RGB color space components of the search moving image information in step 302. It is determined whether to calculate the force. If the histogram calculation is skipped (not calculated), the process proceeds to step 307. If the histogram calculation is not skipped (calculated), the process proceeds to step 306.
- step 304 a histogram of the signal extracted in step 304 is calculated.
- step 307 whether or not to calculate a correlation value between adjacent frames or an arbitrary frame in the time direction for the signal extracted in step 304 or the histogram calculated in step 306 Is determined. If the correlation value calculation is skipped (not calculated), the process proceeds to step 309. If the correlation value calculation is not skipped (calculated), the process proceeds to step 308.
- step 309 it is determined whether or not the correlation value of step 308 can be subjected to the group operation. If the grouping process is skipped (no processing), the process proceeds to step 311. If the grouping process is not skipped (processed), the process proceeds to step 310.
- step 310 the correlation value of step 308 is subjected to grouping processing of correlation values in units of one or more frames in the time direction at a certain threshold value, and the number of frames, the number of groups, etc. Is calculated as
- a feature amount table used for search processing is created from the feature amounts calculated based on the input moving image information for search.
- the process of calculating the feature amount of the input moving image information is performed, and the process of reading out the feature amount table for the database used for the search process is performed.
- the feature amount used for matching judgment is determined from the feature amount table registered in the database. Extract.
- step 314 matching judgment is performed based on the feature amount of the moving image to be searched and the feature amount of the moving image in the database. If they match, the process proceeds to step 3 15 If they do not match, the process proceeds to step 316. Step 3 15 saves the matching result. In step 316, it is determined whether to end the matching process. If the processing is to be terminated, the process proceeds to step 317. If not, the process returns to step 313.
- step 317 upon receiving the matching result, a process of reading out a search result moving image corresponding to the result from the moving image database is performed.
- step 318 the search result moving image read by the read processing unit is output. It should be noted that none of the four skip processes of step 303, step 305, step 307 and step 309 in FIG. 14 are performed.
- step 410 the moving image information for database registration is input.
- step 402 the input moving image information for database registration is registered in the moving image database.
- step 4003 the luminance signal, color difference signal, each component of the RGB color space, etc. of the input moving image information for database registration are extracted.
- step 404 it is determined whether or not extraction processing is to be performed on the moving image information for database registration with a part of the frame of the moving image as a local region. If the local region extraction is skipped (not extracted), the process proceeds to step 406. If the local region extraction is not skipped (extracted), the process proceeds to step 405.
- one or more rectangular shapes or arbitrary shapes should be extracted in order to extract a part of the frame of the moving image as a local region in order to calculate a feature amount for the moving image information for database registration. Performs a local region extraction process.
- step 406 the power of calculating the histogram of the luminance signal, the color difference signal, and the components of the RGB color space, etc. of the local area extracted in step 405 or the entire moving image information for database registration extracted in step 403 It determines whether or not. If the histogram calculation is skipped (not calculated), the process proceeds to step 408. If the histogram calculation is not skipped (calculated), the process proceeds to step 407.
- step 407 a histogram of the signal extracted in step 405 is calculated.
- step 408 it is determined whether the signal extracted in step 405 or the histogram calculated in step 407 is capable of calculating a correlation value. Correlation calculation If the output is skipped (not calculated), the process proceeds to step 410. If the correlation value calculation is not skipped (calculated), the process proceeds to step 409. The correlation value is calculated between adjacent frames or between arbitrary frames along the time direction.
- step 410 it is determined whether or not to perform grouping processing on the correlation value in step 409. If the grouping processing is skipped (no processing), the process proceeds to step 412. If the grouping processing is not skipped (processed), the process proceeds to step 411.
- step 411 the correlation values of step 409 are grouped into correlation values in units of one or more frames in the time direction at a certain threshold, and the number of frames, the number of groups, etc. are used as feature amounts. calculate.
- step 412 a feature amount table to be used in the search processing is created from the feature amounts calculated based on the moving image information for database registration.
- step 4 13 the feature amount table for the database used for the search process is registered in the feature amount table database. It should be noted that all of the four skip processes of step 404, step 406, step 408 and step 410 in FIG. 15 are not executed.
- a search moving image input unit for inputting search target moving image data; and 1 for calculating a feature amount with respect to the input search target moving image.
- a local region extraction unit for extracting the above-mentioned rectangular or arbitrary-shaped local region, and a luminance signal, a color difference signal, components of an RGB color space, and XY of a feature amount calculation region extracted from an input search target moving image.
- a signal extraction unit that extracts image signal components such as each component of the Z color space, each component of the uniform color space, and each component of the Munsell color space, or converts and extracts those signals, based on the input moving image to be searched
- a feature amount calculation unit including a feature amount table creation processing unit that creates a feature amount table for use in a search process from the feature amounts calculated in the above step; and moving image data for a database is input as a database.
- data A moving image input unit; a moving image database registration processing unit for registering inputted moving image data in a moving image database; and a moving image for registering inputted database moving image data.
- Database and; input A local region extraction unit that extracts one or more rectangular or arbitrary-shaped local regions for calculating feature values from the input database registration moving image, Luminance signal, color difference signal, RG of extracted feature amount calculation area
- a signal extraction unit that extracts each component of the B color space, each component of the XYZ color space, each component of the uniform color space, each component of the Munsell color space, or converts and extracts these signals, and the input moving image for database registration
- a feature amount calculation unit comprising a feature amount table creation processing unit for creating a feature amount table used in a search process from the feature amounts calculated based on the feature amount; and the created database feature amount table as a feature amount table.
- a table database registration processing unit for registering in the database for the database; a table database for registering the created feature amount table for the database; and a process performed by the feature amount calculation unit of the input moving image to be searched and used for the search process.
- a table database read processing unit that reads the feature amount table for the database, and the input moving image to be searched
- a feature value used for matching determination is extracted from the feature value table created based on the feature value table and a feature value table registered in the feature value table database, and is output to the matching result determination unit.
- the matching processing unit that saves the matching results from the database and outputs the matching results to the processing unit after the matching process is completed, and the features of the search target moving image and the database moving image extracted by the matching processing unit.
- a matching processing based on a certain threshold value is performed, and a matching result determining unit that outputs a matching result to a matching processing unit; and a moving image database that receives the matching result of the matching result determining unit Video data for reading the search result moving image corresponding to the result from the And a search result output unit for outputting the moving image read search result, the moving image retrieval method and apparatus consisting of the read processor; base read processing unit and.
- the relative value does not depend on the input data as compared with the average value or the variance value of the luminance value of each frame, and does not need to be varied with respect to the threshold value input when performing grouping.
- an appropriate threshold value can be set.
- the grouping feature amount as a search parameter, the Since the number of parameters used for the search can be reduced, the matching processing time can be shortened.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Library & Information Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Color Television Systems (AREA)
Description
Claims
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US10/502,370 US20060001833A1 (en) | 2002-01-25 | 2003-01-23 | Moving picture search apparatus |
| EP03731840A EP1477933A1 (en) | 2002-01-25 | 2003-01-27 | Moving picture search apparatus |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2002-17625 | 2002-01-25 | ||
| JP2002017625A JP2003216954A (ja) | 2002-01-25 | 2002-01-25 | åç»åæ€çŽ¢ææ³åã³ãã®è£ 眮 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2003063082A1 true WO2003063082A1 (en) | 2003-07-31 |
Family
ID=27606176
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2003/000709 Ceased WO2003063082A1 (en) | 2002-01-25 | 2003-01-27 | Moving picture search apparatus |
Country Status (4)
| Country | Link |
|---|---|
| US (1) | US20060001833A1 (ja) |
| EP (1) | EP1477933A1 (ja) |
| JP (1) | JP2003216954A (ja) |
| WO (1) | WO2003063082A1 (ja) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8600113B2 (en) | 2004-11-12 | 2013-12-03 | The University Court Of The University Of St. Andrews | System, method and computer program product for video fingerprinting |
| JP4951521B2 (ja) * | 2004-11-30 | 2012-06-13 | ãŠãããŒã·ãã£ãŒ ã³ãŒã ãªã ã¶ ãŠãããŒã·ãã£ãŒ ãªã ã»ã³ã ã¢ã³ããªã¥ãŒãº | ãããªãã£ã³ã¬ãŒããªã³ãã®ã·ã¹ãã ãæ¹æ³ãåã³ã³ã³ãã¥ãŒã¿ããã°ã©ã 補å |
| US8196045B2 (en) | 2006-10-05 | 2012-06-05 | Blinkx Uk Limited | Various methods and apparatus for moving thumbnails with metadata |
| US8078603B1 (en) | 2006-10-05 | 2011-12-13 | Blinkx Uk Ltd | Various methods and apparatuses for moving thumbnails |
| AU2006252090A1 (en) * | 2006-12-18 | 2008-07-03 | Canon Kabushiki Kaisha | Dynamic Layouts |
| US20090327272A1 (en) * | 2008-06-30 | 2009-12-31 | Rami Koivunen | Method and System for Searching Multiple Data Types |
| JP2011529293A (ja) * | 2008-07-23 | 2011-12-01 | ãšã«ãã£ãŒãŠãŒ ãã¯ãããžãŒãº ãšã¹ãšãŒãšã¹ | ãã¬ãŒã ã«åºã¥ããããªãããã³ã° |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002117407A (ja) * | 2000-10-10 | 2002-04-19 | Satake Corp | åç»åæ€çŽ¢æ¹æ³åã³ãã®è£ 眮 |
-
2002
- 2002-01-25 JP JP2002017625A patent/JP2003216954A/ja active Pending
-
2003
- 2003-01-23 US US10/502,370 patent/US20060001833A1/en not_active Abandoned
- 2003-01-27 WO PCT/JP2003/000709 patent/WO2003063082A1/ja not_active Ceased
- 2003-01-27 EP EP03731840A patent/EP1477933A1/en not_active Withdrawn
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002117407A (ja) * | 2000-10-10 | 2002-04-19 | Satake Corp | åç»åæ€çŽ¢æ¹æ³åã³ãã®è£ 眮 |
Non-Patent Citations (3)
| Title |
|---|
| AKIRA KODAMA ET AL.: "Kido bunpu no sokan o mochiita dogazo kensaku shuho no kento", ITE TECHNICAL REPORT, vol. 23, no. 80, 17 December 1999 (1999-12-17), pages 73 - 78, XP002967059 * |
| HIDEKAZU TAKAHASHI ET AL.: "Jikan tokuchoryo ni yoru dogazo kensaku shuho no kaiseki", ITE TECHNICAL REPORT, vol. 24, no. 79, 15 December 2000 (2000-12-15), pages 61 - 65, XP002967057 * |
| HIDEKAZU TAKAHASHI ET AL.: "Kido bunpu no sokan to iro kukan joho o riyo shita dogazo kensaku hoshiki no kento", INFORMATION PROCESSING SOCIETY OF JAPAN KENKYU HOKOKU, vol. 2000, no. 24, 3 March 2000 (2000-03-03), pages 7 - 12, XP002967058 * |
Also Published As
| Publication number | Publication date |
|---|---|
| JP2003216954A (ja) | 2003-07-31 |
| EP1477933A1 (en) | 2004-11-17 |
| US20060001833A1 (en) | 2006-01-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8965067B2 (en) | Face data acquirer, end user video conference device, server, method, computer program and computer program product for extracting face data | |
| KR101942987B1 (ko) | ëìì 배겜 ì ê±° ë°©ë², ìì€í ë° ì»Žíší° íë ê°ë¥ ì ì¥ì¥ì¹ | |
| KR100860988B1 (ko) | ìíì€ë€ììì ê°ì²Ž íì§ë¥Œ ìí ë°©ë² ë° ì¥ì¹ | |
| JP5097280B2 (ja) | ç»ååã³ç»å矀ã衚çŸãæ¯èŒåã³æ€çŽ¢ããæ¹æ³åã³è£ 眮ãããã°ã©ã ãã³ã³ãã¥ãŒã¿èªã¿åãå¯èœãªèšæ¶åªäœ | |
| CN116363554B (zh) | äžç§çæ§è§é¢å ³é®åž§æåæ¹æ³ãç³»ç»ãä»èŽšã讟å€åç»ç«¯ | |
| KR101812103B1 (ko) | ìžë€ìŒìŽë¯žì§ ì€ì ë°©ë² ë° ì€ì íë¡ê·žëš | |
| US7760249B2 (en) | Image recording and playing system and image recording and playing method | |
| KR20130120175A (ko) | ìºëŠ¬ì»€ì² ìë ìì± ì¥ì¹, ë°©ë² ë° ì»Žíší° íë ê°ë¥í êž°ë¡ ë§€ì²Ž | |
| JP2004310475A (ja) | ç»ååŠçè£ çœ®ãç»ååŠçãè¡ãæºåž¯é»è©±ãããã³ç»ååŠçããã°ã©ã | |
| JP2012105205A (ja) | ããŒãã¬ãŒã æœåºè£ 眮ãããŒãã¬ãŒã æœåºããã°ã©ã ãããŒãã¬ãŒã æœåºæ¹æ³ãæ®åè£ çœ®ãããã³ãµãŒãè£ çœ® | |
| CN111080683B (zh) | åŸåå€çæ¹æ³ãè£ çœ®ãååšä»èŽšåçµåè®Ÿå€ | |
| KR20070009174A (ko) | ìŽë¯žì§ ë¶ëЬ í©ì± ì¥ì¹ ë° ë°©ë², ê·žëŠ¬ê³ ê·žê²ì ìŽì©íìŽë¯žì§ ê²ìë°©ë² | |
| WO2003063082A1 (en) | Moving picture search apparatus | |
| EP2890108A1 (en) | Method for sorting a group of images of a database and method for color correcting an image, corresponding devices, computer program and non-transitory computer-readable medium | |
| US20050243180A1 (en) | Electronic camera and image processing program | |
| JP3358692B2 (ja) | æ åãããã¯å顿¹æ³åã³è£ 眮 | |
| CN104851114B (zh) | äžç§å®ç°åŸåå±éšåè²çæ¹æ³åç»ç«¯ | |
| WO2005048133A1 (ja) | ç»ååŠçæ¹æ³ãç»ååŠçã·ã¹ãã ãããã³ç»ååŠçããã°ã©ã | |
| JP2007122101A (ja) | ç»ååŠçè£ çœ®åã³ç»ååŠçæ¹æ³ | |
| JP2005339076A (ja) | ãªããžã§ã¯ãé åæœåºè£ çœ®ãæœåºæ¹æ³ããã®æ¹æ³ã®ããã°ã©ã åã³ãã®ããã°ã©ã ãèšé²ããèšé²åªäœ | |
| CN115018706B (zh) | 暪å±è§é¢çŽ æèœ¬å䞺ç«å±è§é¢çŽ æçæ¹æ³ | |
| CN110517252B (zh) | äžç§è§é¢æ£æµæ¹æ³åè£ çœ® | |
| JP2000137820A (ja) | ç»ååŠçè£ çœ®ããã³æ¹æ³ã䞊ã³ã«æäŸåªäœ | |
| KR20050024134A (ko) | ìì묌ìì ìëìŒë¡ ìì ë°°ì¹ ì 볎륌 ì¶ì¶íë ìì€í | |
| TWI736060B (zh) | é«è§£æåºŠèŠé »åœ±åèçæ¹æ³ãè£çœ®åé»åèšå |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| AK | Designated states |
Kind code of ref document: A1 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE ES FI GB GD GE GH GM HR HU ID IL IN IS KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ OM PH PL PT RO RU SC SD SE SG SK SL TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW |
|
| AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IT LU MC NL PT SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
| 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: 2003731840 Country of ref document: EP |
|
| WWP | Wipo information: published in national office |
Ref document number: 2003731840 Country of ref document: EP |
|
| ENP | Entry into the national phase |
Ref document number: 2006001833 Country of ref document: US Kind code of ref document: A1 |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 10502370 Country of ref document: US |
|
| WWP | Wipo information: published in national office |
Ref document number: 10502370 Country of ref document: US |
|
| WWW | Wipo information: withdrawn in national office |
Ref document number: 2003731840 Country of ref document: EP |