JPH04335481A - Method and device for extracting area for color image - Google Patents

Abstract

PURPOSE:To extract the designated area without excess or shortage, with regard to an area extracting method and an area extracting device for extracting image data of the designated area from a color image. CONSTITUTION:By this method, a picture element belonging to a designated area is selected and extracted, based on a relative position in a color space of a representative color for showing a representative color and a color of each picture element of a color image in a designated area of the color image, and from an obtained extraction result, an isolated point whose color is different from a peripheral picture element is eliminated. Also, the area extracting device is constituted by providing an area extracting means 111 for selecting and extracting the picture element belonging to the designated area, based on the relative position in the color space of the representative color for showing the representative color and the color of each picture element of the color image in the designated area of the color image, and an isolated point eliminating means 112 for smoothing the obtained extraction result and eliminating the isolated point whose color is different from the peripheral picture element.

Description

[Detailed description of the invention]

0001

[Field of Industrial Application] The present invention relates to a region extraction method and a region extraction device for extracting image data of a designated region from a color image, and more particularly to a region extraction method and device for extracting image data by focusing on the color of a designated region. and a region extraction device. When designing posters, products, etc. using a computer-based design system, or when creating computer graphics, it is necessary to assign appropriate colors to specified areas of the color image obtained by scanning the created color image or photograph. , you may need to adjust the assigned colors. Therefore, there is a need for a function to extract an area specified by an operator from a color image.

0002

2. Description of the Related Art Conventionally, region extraction processing has been performed by focusing on the position on an image of the region to be extracted. For example, an operator uses a position input device such as a light pen to trace and specify the outline of the area to be extracted on a color monitor, and the area surrounded by the specified outline is then designated as the designated area. There is a way to extract it as

[0003] Furthermore, contour extraction processing is performed on the entire color image or a part thereof, a contour surrounding a desired area is specified from among the extracted contours, and image data surrounded by this contour is extracted. There is a way.

0004

[Problems to be Solved by the Invention] Incidentally, since tracing the outline with a light pen is very complicated, when this method is used, the work burden on the operator during area extraction processing is heavy. In addition, since the positional accuracy of the light pen is not very high, it is difficult to accurately trace the outline of the specified area to be extracted, and in order to extract the specified area without too much or too little, it is necessary to trace the outline many times. I had to.

[0005] Furthermore, the process of extracting contours from a color image involves time-consuming processes such as masking and thinning processes for the entire or part of the color image, and it also involves time-consuming processes such as masking and thinning processes for the entire or part of the color image. If the contour is not extracted, the contour extraction process has to be performed again, so the time required for the region extraction process is long. on the other hand,
Another possible method is to focus on the color of the area to be extracted and specify that color to extract the corresponding area.

[0006] However, in this case, taking into consideration that the colors that humans perceive as similar colors have a wide range in color space, the color range is defined by three parameters, for example, hue, saturation, and brightness. Must be specified as a range. However, the operator needs sufficient experience to understand the correspondence between the range of these parameters and the range of colors perceived by humans and to specify an appropriate range.

Furthermore, if the area extracted as described above is different from the specified area intended by the operator, it is necessary to change the ranges of the hue, saturation, and brightness parameters. However, by specifying each of the three parameters independently in this way, it is difficult for the operator to intuitively grasp changes in the color range caused by changes in the range of each of these components (parameters). It is difficult to feed back the result to the next extraction process and extract exactly the specified area.

As a technique for solving these difficulties, the present applicant has already filed Japanese Patent Application No. 3-49981 titled ``Color Extraction Method and Color Extraction Apparatus''. In this method, a sphere or a cube centered on a representative color of a specified area is set as an extraction range in a color space, and image data distributed within this extraction range is extracted as image data belonging to the specified area.

For example, in the HSV color system, as the saturation decreases, the distance in the color space between colors belonging to different hues decreases. Therefore,
If you try to extract all the pixels in the specified area, including colors with low saturation, the extraction range described above will become too large, and the pixels outside the specified area will be whitish or blackish pixels of a different hue. , the pixels are also extracted together, resulting in too many pixels being extracted outside the specified area. On the other hand, if an attempt is made to exclude pixels outside the specified area, the above-mentioned extraction range becomes too small, and some of the pixels in the specified area remain unextracted. Therefore, in order to remove the above-mentioned over-extracted pixels and unextracted pixels, complicated correction work is required.

SUMMARY OF THE INVENTION An object of the present invention is to provide a region extraction method and a region extraction device that extract a specified region in just the right amount.

[0011]

[Means for Solving the Problems] FIG. 1 is a diagram showing the principle of the region extraction method of the present invention. The invention of claim 1 provides for selecting and extracting pixels belonging to a specified area based on the relative position in a color space between a representative color indicating a representative color in a specified area of a color image and the color of each pixel of a color image. The method is characterized in that isolated points whose color is different from surrounding pixels are removed from the obtained extraction results.

FIG. 2 is a block diagram of the principle of the region extraction device of the present invention. The invention of claim 2 provides for selecting and extracting pixels belonging to a designated area based on the relative position in a color space between a representative color indicating a representative color in a designated area of a color image and the color of each pixel of the color image. area extraction means 111
and isolated point removing means 112 for smoothing the obtained extraction result and removing isolated points whose color is different from surrounding pixels.

According to a third aspect of the invention, in the area extracting device according to the second aspect, the isolated point removing means 112 stores in advance image data representing a predetermined color corresponding to all pixels of the color image. A holding means 121 holds the image data of each pixel obtained by the area extraction means 111 according to the position in the color image, and a holding means 121 stores the image data of each pixel of the color image held in the holding means 121 as that pixel. It is characterized by comprising an intermediate value filter 122 that replaces image data in a block consisting of a plurality of surrounding pixels with an intermediate value and outputs the image data as the image data of that pixel.

FIG. 3 is a diagram showing the configuration of a region extracting device according to a fourth aspect of the present invention. According to a fourth aspect of the invention, in the area extracting device according to the second aspect, the isolated point removing means 112 may perform the following steps in advance:
holding means 121 which stores image data representing a predetermined color corresponding to all pixels of a color image, and holds image data of each pixel obtained by the area extraction means 111 according to its position in the color image; , a detection means 131 for detecting a block including a plurality of pixels including at least one pixel having a color different from surrounding pixels from the color image held in the holding means 121 as a block including an isolated point; and a detection result by the detection means 131. an intermediate value filter 132 that replaces the image data of the center pixel of the corresponding block held in the holding means 121 with the intermediate value of the image data in that block in response to the input of the image data of the center pixel; It is characterized by having the following.

[0015]

[Operation] The present invention extracts pixels belonging to a specified area based on the relative position in the color space with the representative color, and removes isolated points from this extraction result. By removing pixels left behind inside the designated area, the designated area can be extracted without excess or deficiency.

The invention of claim 2 provides an isolated point removing means 112.
However, by smoothing the extraction result obtained by the area extraction means 111, isolated points whose color is different from surrounding pixels are removed, so that pixels extracted from outside the specified area or left behind inside the specified area are removed. By removing the pixels that have been detected, it is possible to extract just the right amount of the specified area. Furthermore, since the intermediate value filter 122 replaces the image data of each pixel held in the holding unit 121 with the intermediate value of the image data of the corresponding block, deterioration in image quality due to isolated point removal processing can be reduced.

Furthermore, the detection means 131 is connected to the holding means 121.
Blocks containing isolated points are detected from the color image held in the color image, and the intermediate value filter 132 performs the above-mentioned replacement processing on the detected blocks, thereby minimizing deterioration in image quality due to isolated point removal processing. be able to.

[0018]

Embodiments Hereinafter, embodiments of the present invention will be described in detail based on the drawings. FIG. 4 shows the configuration of an embodiment of an image processing system to which the region extraction device of the present invention is applied. In FIG. 4, the image processing system stores image data representing the color of each pixel of a color image read from a document by an image reading device 201 or a color image created by an image creation device 202 in an RGB color system in an image holding memory 203. The image data is temporarily stored in the image storage memory 203 and then transferred from the image storage memory 203 to the display memory 205 and displayed on the color monitor 206.

The area extraction device 210 also includes a coordinate conversion section 211 and a memory 212, and the coordinate conversion section 211 converts image data in the RGB color system stored in the image holding memory 203, for example. The image data is converted into HSV color system image data (hereinafter simply referred to as image data), and the obtained image data is held in the memory 212 and subjected to area extraction processing by the area extraction means 111. Further, the image data extracted by the area extraction means 111 is subjected to isolated point removal processing by the isolated point removal means 112, and then converted to image data in the RGB color system by the coordinate conversion section 213, and is converted into image data in the RGB color system for display as described above. memory 2
05, the image is combined with the original color image held in the image holding memory 203 and displayed on the color monitor 206.

The coordinate conversion units 211 and 213 described above may be configured using, for example, a lookup table, and the coordinate conversion unit 211 converts the HSV table into an HSV table corresponding to each image data expressed in the RGB color system. A lookup table storing color system image data is prepared, and the coordinate conversion unit 213 prepares a lookup table storing RGB color system image data corresponding to the HSV color system image data. Bye. In addition, calculation processing is performed using a conversion formula for converting image data in the RGB color system to image data in the HSV color system, and a conversion formula for converting image data in the HSV color system to image data in the RGB color system. The coordinate system of the image data may also be converted.

Further, in FIG. 4, the area extracting means 112
is formed from a conditional expression creation section 221, an arithmetic processing section 222, and a determination processing section 223, and the conditional expression creation section 221 creates a color indicating an extraction range corresponding to a designated area in response to an instruction from an operator. A conditional expression representing a solid in space is created, and based on this conditional expression, the arithmetic processing unit 222 and the determination processing unit 223
is configured to select and extract image data distributed within the extraction range.

The operator inputs the position of at least one pixel to indicate the specified region and a numerical value (hereinafter referred to as index r) indicating the size of the extraction region as an instruction to the conditional expression creation unit 221. Bye. In response to this, the conditional expression creation unit 221 sets the average value of the image data of at least one specified pixel as the representative color of the specified area, and creates a conditional expression representing a sphere or a cube centered on the representative color. good. Also,
The configuration may be such that the radius of the sphere and the length of the sides of the cube in the above-mentioned conditional expression are determined according to the value of the index r. As a result, when redoing the region extraction process, the size of the extraction range can be controlled by changing the value of the index r, so the results of the previous extraction process can be fed back to specify the optimal extraction range. becomes possible.

Further, the conditional expression creation unit 221 described above limits the range in the hue direction particularly narrowly, for example, in the hue direction expressed by a rectangular parallelepiped or sphere long in the saturation-brightness direction and two planes with a constant hue. The extraction range may be a solid that expands more in the saturation-brightness direction than in the hue direction, such as the intersection with the limit. The arithmetic processing unit 222 performs calculations by substituting each image data stored in the memory 212 into the conditional expression obtained in this way, and the determination processing unit 223 uses the calculation results to determine the above-mentioned extraction range. The configuration is such that the image data found to be included is extracted and sent to the isolated point removal means 112.

In FIG. 4, the isolated point removing means 112 is
The image data extracted by the above-mentioned area extraction means 111 is held in the holding means 121 according to the position in the color image, and the content of this holding means 121 is On the other hand, the intermediate value filter 122 is configured to perform replacement processing using an intermediate value.

Image data indicating white may be stored in advance in the holding means 121 in correspondence with all pixels of a color image, and the image data extracted by the area extraction means 111 may be overwritten. In addition, the intermediate value filter 122
is 3 pixels centered on each pixel stored in the holding means 121.
The configuration is such that the image data of blocks consisting of ×3 pixels are arranged in order of size, paying attention to the value of the saturation parameter, and the fifth image data is set as the median value, and the image data of the central pixel of this minute area is replaced. Bye.

In this case, if 5 or more of the pixels included in the block are white, the median value becomes image data representing white, so the central pixel of this block is replaced with white, and vice versa. , when five or more pixels have a chromatic color, the median value becomes image data indicating the chromatic color, and the image data of the central pixel is replaced with this image data. For example, as shown in FIG. 5(a), a small number of chromatic pixels (shown with diagonal lines in the figure) surrounded by white pixels are each replaced with white pixels;
As shown in (b), a small number of white pixels surrounded by chromatic colored pixels are each replaced by the surrounding chromatic colored pixels.

In this way, by smoothing the image data stored in the holding means 121 using the intermediate value filter 122, a small number of pixels extracted from outside the specified area or left behind inside the specified area can be smoothed. A small number of pixels can be removed as isolated points. Also, the median filter is
Since the image quality is less degraded than other smoothing filters such as averaging filters, when the isolated point removing means 112 is configured using the intermediate value filter 122, the deterioration of image quality due to the isolated point removing process is small.

[0028] Here, since most of the overtaken portion and the leftover portion are island-like portions consisting of a small number of pixels,
By removing isolated points using the median filter as described above, most of the excess and leftover portions can be removed. This makes it possible to extract just the right amount of the specified area, thereby improving the accuracy of area extraction processing.

[0029] Furthermore, the overexcessed areas and unremoved areas that could not be removed by the above-mentioned isolated point removal process are extremely small, and since they are areas where a relatively large number of pixels are gathered,
Processing to remove these portions is easy, and the burden on the operator during correction processing to remove excessively removed portions and unremoved portions can be significantly reduced. By the way, the intermediate value filter is a filter that causes less deterioration in image quality, but
It is inevitable that image quality deterioration such as blurring of the outline of the designated area will occur.

A method for removing isolated points while minimizing such deterioration in image quality will be described below. FIG. 6 shows a configuration diagram of an embodiment of the isolated point removing means of the region extracting device according to the fourth aspect. In FIG. 6, the detection means 131 is formed of a differential processing section 311 and an isolated point discriminating section 312. pixels (
For example, the differential of the image data is obtained for each block consisting of 8×8 pixels), and based on the result of this differentiation, the isolated point determination unit 312 determines whether or not each block contains an isolated point. ing. The above-mentioned differential processing section 311
For example, the differential of the saturation parameter of the image data of each row and each column may be calculated sequentially for each block. For example, if at least one of the differential results in each row and each column includes a pair of a positive extreme value and a negative extreme value, the isolated point determination unit 312 determines that the block includes an isolated point. It is only necessary to determine that there is one and send the determination result to the readout circuit 321.

The readout circuit 321 includes the detection means 13 described above.
1, the image data is sequentially read out block by block from the holding means 121, and in accordance with the input of the determination result from the isolated point discriminating section 312, blocks determined to include isolated points are filtered through the intermediate value filter 132. The blocks that are determined not to include isolated points are sent to the coordinate conversion unit 213 as they are.

The intermediate value filter 132 is constructed in the same manner as the intermediate value filter 122 described above, and is an intermediate value of the image data of a minute portion of 3×3 pixels centered on each pixel of the input block. All that is required is to perform a process of replacing the image data. In this case, like the blocks shown by arrows a and b in FIG. 7, the extracted chromatic pixels are surrounded by white pixels, or the block shown by arrow c,
A block in which a white pixel is surrounded by extracted chromatic pixels is detected as a block containing an isolated point by the above-mentioned detection means 131, and an intermediate value filter is applied to these blocks. On the other hand, like the block indicated by arrow d in FIG. 7, a block that includes an outline of a designated area (indicated by a dotted line in the figure) is not considered to include an isolated point.

In this way, by detecting blocks containing isolated points and applying an intermediate value filter only to the detected blocks, it is possible to selectively remove isolated points. Deterioration in image quality can be minimized. Note that filters for smoothing image data are not limited to the above-mentioned intermediate value filter;
A simple local average filter, local weighted average filter, etc. may also be used. For example, a local average filter simply calculates the average value of image data for each block consisting of a plurality of pixels, smoothes this average value as image data of pixels included in that block, and removes isolated points. This method is simple and requires a small amount of calculation, so it has the advantage of speeding up isolated point removal processing, but it may significantly degrade image quality.

[0034]

Effects of the Invention As explained above, the present invention extracts pixels belonging to a specified area based on the position of each pixel of a color image in the color space, and from the obtained extraction results, it is possible to determine whether the pixels are different in color from the surrounding area. By removing isolated points, it becomes possible to extract the specified area without excess or deficiency, and the burden of correction processing on the operator can be reduced.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the principle of the region extraction method of the present invention.

FIG. 2 is a principle block diagram of the region extraction device of the present invention.

FIG. 3 is a diagram showing the configuration of a region extracting device according to a fourth aspect of the present invention.

FIG. 4 is a configuration diagram of an embodiment of an image processing system to which the region extraction device of the present invention is applied.

FIG. 5 is an explanatory diagram of isolated point removal processing using an intermediate value filter.

FIG. 6 is a configuration diagram of an embodiment of isolated point removing means of the region extracting device according to claim 4;

FIG. 7 is an explanatory diagram of isolated point removal processing of the region extraction device according to the fourth aspect.

[Explanation of symbols]

111 Area extraction means 112 Isolated point removal means 121 Holding means 122,132 Intermediate value filter 131 Detection means 201 Image reading device 202 Image creation device 203 Image holding memory 205 Display memory 206 Color monitor 210 Area extraction device 211, 213 Coordinate conversion section 212 Memory 221 Conditional expression creation part 222 Arithmetic processing unit 223 Judgment processing unit 311 Differential processing section 312 Isolated point determination unit 321 Readout circuit

Claims (4)

[Claims] 1. Selecting and extracting pixels belonging to the specified area based on the relative position in a color space between a representative color indicating a representative color in the specified area of the color image and the color of each pixel of the color image. An area extraction method is characterized in that isolated points having a color different from surrounding pixels are removed from the obtained extraction results. 2. Selecting and extracting pixels belonging to the specified area based on the relative position in a color space between a representative color indicating a representative color in the specified area of the color image and the color of each pixel of the color image. area extraction means (111) for
An area extracting device characterized by comprising: isolated point removing means (112) for smoothing the obtained extraction result and removing isolated points whose color is different from surrounding pixels. 3. The area extracting device according to claim 2, wherein the isolated point removing means (112) stores in advance image data representing a predetermined color corresponding to all pixels of the color image. , a holding means (121) for holding the image data of each pixel obtained by the area extraction means (111) according to the position in the color image; and the holding means (12).
An intermediate value filter (122) that replaces the image data of each pixel of the color image held in 1) with the intermediate value of image data in a block consisting of that pixel and a plurality of surrounding pixels, and outputs the result as image data of that pixel. A region extraction device characterized by comprising: 4. The area extracting device according to claim 2, wherein the isolated point removing means (112) stores in advance image data representing a predetermined color corresponding to all pixels of the color image. , a holding means (121) for holding the image data of each pixel obtained by the area extraction means (111) according to the position in the color image; and the holding means (12).
Detection means (1) for detecting a block consisting of a plurality of pixels including at least one pixel whose color is different from surrounding pixels from the color image held in (1) as a block containing an isolated point.
31), and in response to the input of the detection result by the detection means (131), the image data of the central pixel of the corresponding block held in the holding means (121) is replaced with the intermediate value of the image data in that block. and an intermediate value filter (13) that outputs the image data of the central pixel.
2) A region extraction device comprising: