JP5104528B2 - Image processing apparatus and image processing program - Google Patents

Description

  The present invention relates to an image processing apparatus and an image processing program.

  A technique for extracting an area of an image constituent element (also called an image object) from an image is known.

  In Patent Document 1, a background color of a document image is specified, pixels other than the background region are extracted from the document image using the background color, and the connected components are generated by integrating the pixels. A method is described in which at least a shape feature is used to classify a predetermined region and a result of region identification of the document image is obtained.

  Patent Document 2 discloses a method for extracting the same color region in an image by performing a raster scan on a search block having a certain size in the image, obtaining a variance of pixels in the search block, and a threshold having a variance. A method is described in which the search block is used as a seed block when the value is below the threshold, and whether or not surrounding pixels are the same color area as the seed block based on the seed block is determined in pixel units, and the same color area is extracted. .

  Japanese Patent Laid-Open No. 2004-228688 sets the color range that is regarded as the same color of the input image based on the read information of the input image, and labels the connected pixels in the set color range, thereby labeling the input image. Techniques for performing are described. Here, the reading information of the input image includes the number of colors of the input image, reading resolution, luminance value, saturation, color difference between adjacent pixels, color dispersion, and the like.

  In Patent Document 4, the entire image is divided into a plurality of parts with a uniform amount of feature and a large area, and a non-uniform part is reduced into a plurality of areas. A large region is selected as a starting region, and regions that are around the starting region and whose features are similar to those of the starting region are integrated into the starting region to form one divided region. Describes a method of forming a plurality of divided regions by repeatedly performing the above operation on an unintegrated region.

JP 2001-297303 A JP 2000-132680 A JP 2000-172849 A JP 2002-032763 A

  By the way, when each area divided by the background color area is extracted from the image as an image component area, a plurality of areas to be handled as one area may be extracted as separate image component areas. is there.

  An object of the present invention is to provide an image processing apparatus and an image processing program capable of extracting a plurality of areas which are divided by a background area but should be handled as one area as one image component area. .

An image processing apparatus according to the present invention includes an image receiving unit that receives an image, a background color specifying unit that specifies a background color of the image, and each region partitioned by the region of the background color from the image. A region extracting means for extracting as an image component region and a relationship between at least two of the extracted image component regions satisfy a predetermined condition to be treated as one image component region in, that having a, a region integrating means for integrating the at least two image components area in one image component area.

In the present invention , the region integration unit may further include a circumscribed rectangle of an image component region (hereinafter referred to as “included region”) among the extracted image component regions as another image component region (hereinafter referred to as “image component region”). The inclusion region and the inclusion region are integrated into one image component region when the inclusion region is included in a circumscribed rectangle of the “inclusion region”) and the inclusion region is close to the inclusion region by a predetermined degree or more. To do.

Further , in the present invention , the region integration means determines, for each of the four sides of the circumscribed rectangle of the inclusion region, whether or not the side is close to the inclusion region, and a predetermined number of sides is included in the inclusion region. When it is determined that the area is close to the area, it is determined that the inclusion area is close to the inclusion area.

  In the aspect of the invention, the region integration unit may determine whether or not the interval between the pixel and the inclusion region is equal to or less than a predetermined interval for each pixel on the side in a direction perpendicular to the side. Inspection is performed, and it is determined based on the inspection result whether the side is close to the inclusion region.

  In the aspect of the invention, the region integration unit may be configured such that, when the number of pixels having the interval equal to or less than the predetermined interval among the pixels on the side is equal to or greater than the predetermined number, the side is close to the inclusion region. Judge that.

  In the aspect of the invention, the region integration unit may be configured such that, for each pixel on the side, an interval between the pixel and the inclusion region is equal to or less than a predetermined interval in a direction perpendicular to the side. It is inspected whether the width of the inclusion area opposite to is greater than or equal to a predetermined width, and based on the inspection result, it is determined whether the side is close to the inclusion area.

  Also, in one aspect of the present invention, the region integration unit is configured such that, among the pixels on the side, the number of pixels in which the interval is equal to or smaller than a predetermined interval and the width is equal to or greater than a predetermined width It is determined that the side is close to the inclusion area.

  In the aspect of the invention, the region integration unit may determine the predetermined interval according to a size of the circumscribed rectangle of the included region such that the larger the circumscribed rectangle of the included region, the larger the predetermined interval. To change.

  In the aspect of the invention, the region integration unit may be configured such that the predetermined width is increased according to a size of the circumscribed rectangle of the included region so that the predetermined width increases as the circumscribed rectangle of the included region increases. To change.

Further, in the reference example , the region integration unit includes a circumscribed rectangle of two image component regions of the extracted image component regions that overlap each other, and features of the two image component regions are predetermined to each other. If they are more or less similar, the two image component regions are integrated into one image component region.

Further, in the reference example , the region integration unit is characterized in that circumscribed rectangles of two image component regions of the extracted image component regions are close to each other by a predetermined degree or more, and the characteristics of the two image component regions Are similar to each other by a predetermined degree or more, the two image component regions are integrated into one image component region.

An image processing program according to the present invention includes: a procedure for accepting an image to a computer; a procedure for specifying a background color of the image; and each region partitioned by the background color region from the image. When at least two of the extracted image component regions and the relationship between at least two image component regions among the extracted image component regions satisfy a predetermined condition to be treated as one image component region, the at least a step of integrating the two image component area in one image component area, Ru is running.

In the image processing program according to the present invention , in the integration step, a circumscribed rectangle of an image component region (hereinafter referred to as “included region”) among the extracted image component regions is another image configuration. When the element region (hereinafter referred to as “inclusion region”) is included in a circumscribed rectangle and the inclusion region is close to the inclusion region by a predetermined degree or more, the inclusion region and the inclusion region are combined into one image. For each of the four sides of the circumscribed rectangle of the inclusion region, it is determined whether or not the side is close to the inclusion region, and a predetermined number of sides is determined to be close to the inclusion region If it is determined that the inclusion area is close to the inclusion area .

In the reference example , in the integration step, circumscribed rectangles of two image component regions of the extracted image component regions overlap each other, and characteristics of the two image component regions are predetermined to each other. If they are more or less similar, the two image component regions are integrated into one image component region.

In the reference example , in the integration step, the circumscribed rectangles of the two image component regions out of the extracted image component regions are close to each other by a predetermined degree and the features of the two image component regions are Are similar to each other by a predetermined degree or more, the two image component regions are integrated into one image component region.

  According to the first aspect of the present invention, it is possible to extract a plurality of areas that are divided by the background area but should be handled as one area as one image component area.

Furthermore , according to the first aspect of the present invention, it is possible to extract an outer peripheral area and an inner area that are separated from each other by a background area as one image component area.

Further , according to the invention described in claim 1 , it is easily determined whether or not the inclusion area is close to the inclusion area by determining whether or not each side of the circumscribed rectangle of the inclusion area is close to the inclusion area. It becomes possible to do.

According to the invention described in claim 2 , it is easily determined whether or not the side is close to the inclusion region by checking whether or not the interval between each pixel on the side and the inclusion region is a predetermined interval. It becomes possible to do.

According to the third aspect of the present invention, it is possible to determine whether or not the side is close to the inclusion region according to the number of pixels on the side where the interval is equal to or less than the predetermined interval. .

According to the invention of claim 4 , whether the side is close to the inclusion region in consideration of the distance between each pixel on the side and the inclusion region and the width of the inclusion region facing each pixel on the side. It becomes possible to determine whether or not.

According to the fifth aspect of the present invention, whether or not the side is close to the inclusion region according to the number of pixels on the side where the interval is equal to or smaller than the predetermined interval and the width is equal to or larger than the predetermined width. Can be determined.

According to the invention described in claim 6 , it is possible to set an appropriate predetermined interval according to the size of the inclusion region.

According to the seventh aspect of the present invention, it is possible to set an appropriate predetermined width according to the size of the inclusion region.

According to the reference example , it is possible to extract, as one image component area, two areas that are separated by a background area and that have circumscribed rectangles that overlap each other and that have similar characteristics.

According to the reference example , it is possible to extract, as one image component area, two areas that are separated by a background area and that have circumscribed rectangles close to each other and that have similar characteristics.

According to the eighth aspect of the present invention, it is possible to extract a plurality of areas that are divided by the background area but should be handled as one area as one image component area.

Furthermore , according to the eighth aspect of the present invention, it is possible to extract an outer peripheral area and an inner area adjacent to each other divided by the background area as one image constituent element area, and a circumscribed rectangle of the inclusion area. By determining whether each side is close to the inclusion region, it is possible to easily determine whether the inclusion region is close to the inclusion region .

According to the reference example , it is possible to extract, as one image component area, two areas that are separated by a background area and that have circumscribed rectangles that overlap each other and that have similar characteristics.

According to the reference example , it is possible to extract, as one image component area, two areas that are separated by a background area and that have circumscribed rectangles close to each other and that have similar characteristics.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram illustrating an example of a configuration of an image processing apparatus 10 according to the present embodiment. The image processing apparatus 10 is an apparatus that extracts an image component area (that is, an image component area) from an image to be processed. In the following description, an image component is referred to as an “image object”, and an image component region is referred to as an “image object region”.

  In one embodiment, the image processing apparatus 10 is realized by cooperation of hardware resources and software, and is, for example, a computer. Specifically, the functions of the image processing apparatus 10 are realized by reading an image processing program recorded on a recording medium into a main memory and executing it by a CPU (Central Processing Unit). The image processing program can be provided by being recorded on a computer-readable recording medium such as a CD-ROM, or can be provided by communication as a data signal. However, the image processing apparatus 10 may be realized only by hardware. Further, the image processing apparatus 10 may be physically realized by one apparatus or may be realized by a plurality of apparatuses.

  As illustrated in FIG. 1, the image processing apparatus 10 includes an image receiving unit 11, a background color specifying unit 12, a region extracting unit 13, a region integrating unit 14, and an image object output unit 15.

  The image receiving unit 11 receives an image to be processed (hereinafter referred to as “input image”). In one aspect, the input image is a color image having a plurality of color components, and includes, for example, three color components of red (R), green (G), and blue (B). The input image is, for example, image data in which a plurality of pixels are arranged in a matrix in the horizontal direction and the vertical direction, and each pixel has a gradation value of each color. In a specific aspect, the input image is bitmap image data of 8 bits for each color of RGB. The input image is, for example, image data generated by an image reading apparatus that optically reads a document image. However, the input image may be image data recorded on various recording media.

  The background color specifying unit 12 specifies the background color of the input image received by the image receiving unit 11. The background color specifying unit 12 may specify one specific color as the background color, or may specify a specific range of colors in the color space as the background color. The background color specifying unit 12 can specify the background color by various methods. For example, the background color specifying unit 12 specifies a color having the highest appearance frequency in the input image or a color within a predetermined range from the color as the background color.

  The area extraction unit 13 extracts each area partitioned by the background color area (that is, the background area) from the input image as an image object area. In one aspect, the region extraction unit 13 converts the input image into a binary image by classifying the pixels of the input image into pixels of the background color and pixels other than the background color. For example, the region extraction unit 13 sets the pixel value of the background color pixel to “0” and sets the pixel value of the pixel other than the background color to “1”, and then displays the full color input image shown in FIG. A binary image is generated. In FIG. 3, the region with the pixel value “0” is expressed in white, and the region with the pixel value “1” is expressed in black. Next, the area extraction unit 13 extracts the area of each connected component of pixels other than the background color in the binary image as an image object area. Here, the connected component is a binary image in which pixels having the same value are connected to each other to form one lump. In a specific aspect, the region extraction unit 13 performs a labeling process on the binary image to generate a labeling image. Here, the labeling process is a process of assigning a different label for each connected component to each connected component pixel in the binary image. For example, the region extraction unit 13 performs a labeling process on the binary image in FIG. 3 to generate a labeling image shown in FIG. In FIG. 4, different label numbers 1, 2, 3, and 4 are assigned to the pixels of each connected component, and regions with different label numbers are represented by different hatching. Yes. Each pixel group region to which the same label number is assigned corresponds to an image object region. That is, in the example of FIG. 4, four image object regions with label numbers 1 to 4 are extracted. In FIG. 4, the outer peripheral area and the inner area of the white character “D” should be treated as one image object area originally, but are separated by the background area and separated into separate image objects with label numbers 3 and 4. It is extracted as a region. As described above, in the extraction processing of the area extraction unit 13, pixels other than the background color are not connected to the area that is divided by the background area, such as the outer peripheral area and the inner area of the white character, but should be treated as one area. Are not extracted as the same region.

  The region integration unit 14 performs the above-described at least 2 when the relationship between at least two image object regions among the image object regions extracted by the region extraction unit 13 satisfies a predetermined condition to be handled as one image object region. One image object area is integrated into one image object area.

  The predetermined condition is set in advance, for example, depending on when the image object area should be integrated and when the image object area should not be integrated. The predetermined condition includes, for example, a condition related to the positional relationship between the image object areas, a condition related to the degree of similarity of features between the image object areas, and the like. A specific example of the predetermined condition will be described in detail later.

  Regarding integration of image object areas, in one aspect, when integrating at least two image object areas, the area integration unit 14 unifies the label numbers of these image object areas into a common label number, and performs labeling after the unification. The image is output as the processing result. For example, when integrating the two image object regions of the white character “D” in FIG. 4, the region integration unit 14 reassigns the label number 3 to the pixel with the label number 4 to generate the labeling image shown in FIG. 5. Output as processing result. In another aspect, when integrating at least two image object regions, the region integration unit 14 outputs information indicating that as a processing result. For example, when integrating the two image object regions of the white character “D” in FIG. 4, the region integration unit 14 indicates that the image object regions with the label numbers 3 and 4 are integrated into one image object region. Information is output as a processing result.

  The image object output unit 15 outputs an image object or information for generating an image object based on the processing results of the region extraction unit 13 and the region integration unit 14.

  In one aspect, the image object output unit 15 outputs a mask image and a cut image for each image object as information for generating an image object. Here, the mask image is a binary image that is partitioned by a circumscribed rectangle of the image object area and represents the shape of the image object. For example, the mask image is an image in which the pixel value “0” is assigned to the pixels in the background area and the pixel value “1” is assigned to the pixels in the image object area. The cut image is an image obtained by cutting the input image with a circumscribed rectangle of the image object area. For example, the image object output unit 15 generates the three mask images 21 to 23 shown in FIG. 6 from the labeling image of FIG. 5 or from the labeling image of FIG. 4 and the integrated information, and the input image of FIG. Three full-color cut-out images 31 to 33 shown in FIG. 7 are generated from the labeling image of FIG.

  In another aspect, the image object output unit 15 generates and outputs an image object based on the mask image and the cut image. Specifically, an AND of the mask image and the cut image is taken, that is, the cut image is masked with the mask image to generate an image object. For example, the image object output unit 15 masks the three cut images 31 to 33 in FIG. 7 with the three mask images 21 to 23 in FIG. 6, respectively, and three full color image objects 41 to 43 shown in FIG. 8. Is generated.

  If the integration processing by the region integration unit 14 is not performed, for example, the four mask images 51 to 54 shown in FIG. 9 are generated from the labeling image shown in FIG. That is, the outer periphery and the inside of the outline character “D” are extracted as separate image objects.

  FIG. 10 is a flowchart showing an example of the operation of the image processing apparatus 10 according to the present embodiment. Hereinafter, an example of the operation of the image processing apparatus 10 will be described with reference to FIG.

  The image processing apparatus 10 receives an input image (for example, 24-bit full color image data) (S1).

  Next, the image processing apparatus 10 specifies the background color of the input image (S2).

  Next, the image processing apparatus 10 extracts each area partitioned by the background color area from the input image as an image object area (S3). That is, the image processing apparatus 10 detects an image object area on the background.

  Next, the image processing apparatus 10 performs an image object region integration process on the extracted image object region (S4). In other words, the image processing apparatus 10 determines that the at least two images when the relationship between at least two of the extracted image object areas satisfies a predetermined condition to be handled as one image object area. The object area is integrated into one image object area.

  Then, the image processing apparatus 10 outputs an image object or information for generating an image object based on the processing results of steps S3 and S4 (S5).

  Hereinafter, first to third specific examples will be shown as specific examples of the process of the region integration unit 14, that is, the integration process of step S4.

[First specific example]
As described above, the outer peripheral area and the inner area of the outline character are extracted as separate image object areas by the area extracting unit 13. However, since the outer periphery and the inner area of the white character constitute one image object, it is desirable that the outer peripheral area and the inner area of the white character are extracted as one image object area. Here, in the outline character, the circumscribed rectangle of the inner region is included in the circumscribed rectangle of the outer peripheral region, and the inner region and the outer peripheral region are close to each other. Therefore, when a circumscribed rectangle of an image object area is included in a circumscribed rectangle of another image object area, and the two image object areas are close to each other, the two image object areas have white characters. There is a high probability of configuring.

  In view of the above circumstances and the like, in this example, the region integration unit 14 uses other circumscribed rectangles of an image object region (hereinafter referred to as “included region”) among the image object regions extracted by the region extraction unit 13. Image object area (hereinafter referred to as "inclusion area") is included in a circumscribed rectangle, and the inclusion area and the inclusion area are combined into one image object area when the inclusion area is close to the inclusion area by a predetermined amount or more. To integrate.

  Specifically, the region integration unit 14 specifies a circumscribed rectangle of each image object region extracted by the region extracting unit 13 and investigates an inclusion relationship between circumscribed rectangles of arbitrary two image object regions. When the inclusion relationship is established, it is determined whether or not the two image object regions are close to each other according to a predetermined determination criterion. When it is determined that the two image object regions are close to each other, the two image object regions are Integrate into the image object area.

  Hereinafter, an example of determining whether or not the inclusion region is close to the inclusion region will be described with reference to FIG. In FIG. 11, the circumscribed rectangle 52 of the inner region (included region) 51 of the outline character “D” is surrounded by the circumscribed rectangle 54 of the outer peripheral region (included region) 53.

  For each of the four sides of the circumscribed rectangle 52 of the inclusion region 51, the region integration unit 14 determines whether the side is close to the inclusion region 53, and determines that a predetermined number of sides are close to the inclusion region 53. In the case where the inclusion area 51 is close to the inclusion area 53.

  Here, for the predetermined number, for example, what kind of white character such as alphabet, number, kanji, etc. is considered, or whether contact with the image end E of the white character as shown in FIG. What is necessary is just to set suitably according to etc. For example, “4” is set when it is desired to integrate only when the possibility of being a white character is very high, and “3” is set when the contact of the white character with the image edge is considered. Then, “2” is set when considering white characters of Chinese characters, and “1” is set when integrating without omission as much as possible. Further, for example, in consideration of a white character such as the Chinese character “da” shown in FIG. 13, in one specific aspect, the region integration unit 14 is orthogonal to each other among the four sides of the circumscribed rectangle of the inclusion region. When it is determined that at least two sides are close to the inclusion area, it is determined that the inclusion area is close to the inclusion area. That is, when it is determined that at least one of the two sides parallel to the horizontal direction of the image is close to the inclusion region and at least one of the two sides parallel to the vertical direction of the image is close to the inclusion region, It is determined that both areas are close to each other.

  Whether or not each side of the circumscribed rectangle of the inclusion area is close to the inclusion area may be determined by an appropriate method, for example, as follows.

  For each pixel on the side of the circumscribed rectangle 52, the region integration unit 14 inspects whether the interval between the pixel and the inclusion region 53 is equal to or less than a predetermined interval in the direction perpendicular to the side, and the inspection result Based on the above, it is determined whether or not the side is close to the inclusion region 53.

  For example, the region integration unit 14 determines that the side is close to the inclusion region 53 when the number of pixels having the interval equal to or smaller than the predetermined interval among the pixels on the side is equal to or larger than the predetermined number. The predetermined number may vary according to the total number of pixels on the side (that is, the length of the side), and is, for example, a number obtained by multiplying the total number of pixels on the side by a predetermined ratio.

  Here, it is considered that the larger the size of the white character, the larger the size of the circumscribed rectangle 52 of the inclusion area 51 and the gap between the inclusion area 51 and the inclusion area 53. In consideration of such circumstances and the like, in one aspect, the region integration unit 14 determines the size of the circumscribed rectangle 52 of the included region 51 so that the predetermined interval increases as the circumscribed rectangle 52 of the included region 51 increases. The predetermined interval is changed according to the above.

  In a specific aspect, the region integration unit 14 determines a predetermined pixel from each pixel on the side of the circumscribed rectangle 52 in a direction away from the inclusion region 51 perpendicular to the side, as indicated by a dashed arrow in FIG. It scans for several minutes, it is test | inspected whether it reaches the inclusion area 53, and it is determined based on the said test result whether the said edge adjoins the inclusion area 53. FIG. For example, when a predetermined number (or a predetermined ratio) or more of the scanned number reaches the inclusion area 53, it is determined that the side is close to the inclusion area 53. The region integration unit 14 changes the number of pixels to be scanned according to the size of the circumscribed rectangle 52 of the inclusion region 51.

  By the way, when a table is included in the input image, the table frame and the characters in the table frame have an inclusive relationship with each other, but it may be desired to extract them as separate image objects. That is, there is a demand for integrating the outer periphery and the inside of the outline character, but not integrating the frame of the table and the character in the frame.

  In consideration of the above-described circumstances and the like, in one aspect, the region integration unit 14 faces the interval from each pixel on the side of the circumscribed rectangle 52 to the inclusion region 53 and each pixel on the side of the circumscribed rectangle 52. Based on the width of the inclusion area 53, it is determined whether or not the side is close to the inclusion area 53. Specifically, for each pixel on the side of the circumscribed rectangle 52, the region integration unit 14 determines that the interval between the pixel and the inclusion region 53 is equal to or less than a predetermined interval in a direction perpendicular to the side. It is inspected whether or not the width of the opposing inclusion area 53 is equal to or larger than a predetermined width, and it is determined whether or not the side is close to the inclusion area 53 based on the inspection result. For example, the region integration unit 14 includes a side when the number of pixels on the sides of the circumscribed rectangle 52 is equal to or smaller than the predetermined interval and the width is equal to or larger than the predetermined width. It is determined that the area 53 is close. Also in this aspect, the predetermined number may change according to the total number of pixels on the side (that is, the length of the side), and is obtained, for example, by multiplying the total number of pixels on the side by a predetermined ratio. Is a number. Also in this aspect, the region integration unit 14 determines the predetermined interval according to the size of the circumscribed rectangle 52 of the included region 51 so that the predetermined interval increases as the circumscribed rectangle 52 of the included region 51 increases. May be changed.

  It is considered that the larger the size of the table, the larger the size of the circumscribed rectangle of the inclusion area, which is the character area in the frame, and the width of the border of the table. In consideration of the circumstances and the like, in one aspect, the region integration unit 14 determines the size of the circumscribed rectangle 52 of the included region 51 so that the predetermined width increases as the circumscribed rectangle 52 of the included region 51 increases. The predetermined width is changed according to the above.

  In one specific aspect, the region integration unit 14, as shown in FIG. 14, from each pixel on the side of the circumscribed rectangle 52 by a predetermined number of pixels in a direction perpendicular to the side and away from the inclusion region 51. Scan to check if the inclusion area 53 is reached. Furthermore, after reaching the inclusion area 53 in each of the above scans, the area integration unit 14 performs additional scanning for a predetermined number of pixels as indicated by the solid line arrow in FIG. Whether or not the side is close to the inclusion region 53 is determined based on the inspection result. For example, when a predetermined number (or a predetermined ratio) or more of the number of scanned lines reaches the inclusion area 53 and the additional scanning does not exit the inclusion area 53, it is determined that the side is close to the inclusion area 53. The region integration unit 14 changes the number of pixels to be scanned first and the number of pixels to be additionally scanned according to the size of the circumscribed rectangle 52 of the inclusion region 51.

  FIG. 15 is a flowchart illustrating an example of the integration process according to the first specific example. Hereinafter, an example of the integration process according to the first specific example will be described with reference to FIG.

  The image processing apparatus 10 specifies a circumscribed rectangle of each extracted image object area (S11).

  Next, the image processing apparatus 10 selects two image object areas from the extracted image object areas, and determines whether there is an inclusion relationship between the circumscribed rectangles of the two image object areas (S12). ).

  If it is determined that there is an inclusion relationship (S12: YES), the image processing apparatus 10 determines whether the inclusion region is close to the inclusion region (S13). Specifically, the image processing apparatus 10 performs scanning in the horizontal direction and the vertical direction from each pixel on each side of the circumscribed rectangle of the inclusion region, and each side of the circumscribed rectangle of the inclusion region is close to the inclusion region. When it is determined that all four sides are close to the inclusion region, it is determined that the inclusion region is close to the inclusion region, and otherwise, it is determined not to be close.

  When it is determined that the inclusion area is close to the inclusion area (S13: YES), the image processing apparatus 10 integrates the inclusion area and the inclusion area into one image object area (S14), and the process proceeds to step S15. Proceed.

  On the other hand, if it is determined in step S12 that there is no inclusion relationship (S12: NO), and if it is determined in step S13 that the inclusion region is not close to the inclusion region (S13: NO), the image processing apparatus 10 The process proceeds to step S15 without performing the integration process.

  In step S15, the image processing apparatus 10 determines whether or not the processing in steps S12 to S14 has been completed for all combinations of image object regions. If not (S15: NO), the processing is performed. Returning to step S12, if it has been completed (S15: YES), the process is terminated.

  For example, in the example of FIG. 11, the circumscribed rectangle 52 of the inclusion area 51 is included in the circumscription rectangle 54 of the inclusion area 53, and the circumscribed rectangle 52 is close to the inclusion area 53 in four directions. By the processing, the inclusion area 51 and the inclusion area 53 are integrated.

  On the other hand, in the example of FIG. 16, circumscribed rectangles 61 to 63 of the three numerical image areas of the pie graph are included in the circumscribed rectangle 64 of the area of the pie chart image, but the circumscribed rectangles 61 to 63 are not in four directions. Since it is not close to the pie chart image area, the three numerical image areas are not integrated into the pie chart image area by the integration process.

[Second specific example]
FIG. 17 shows an example of symbols included in the input image. In FIG. 17, a white area represents a background area, and 35 gray areas on the white area represent colored areas. The 35 colored areas are extracted as separate image object areas by the area extracting unit 13. However, in FIG. 17, since 35 colored areas are gathered to constitute one image object, it is desirable that the 35 colored areas are extracted as one image object area. Here, in the case of the design as shown in FIG. 17, the circumscribed rectangle of each region overlaps with the circumscribed rectangle of the adjacent region, and the characteristics (colors, etc.) of the two regions are similar to each other. Therefore, when the circumscribed rectangles of the two image object areas overlap with each other and the characteristics of the two image object areas are similar to each other, the two image object areas may constitute one image object area. high.

  In view of the above circumstances and the like, in this example, the region integration unit 14 overlaps the circumscribed rectangles of the two image object regions of the image object regions extracted by the region extraction unit 13, and the two image objects. When the features of the regions are similar to each other by a predetermined degree or more, the two image object regions are integrated into one image object region.

  Specifically, the region integration unit 14 specifies a circumscribed rectangle of each image object region extracted by the region extracting unit 13, and investigates the overlapping relationship between the circumscribed rectangles of any two image object regions. When the overlapping relationship is established, the feature amounts of the two image object regions are obtained, the two feature amounts are compared, and the two image object regions are converted into one image object according to the comparison result. Integrate into the area. Specifically, the region integration unit 14 determines whether or not the two feature amounts are similar according to a predetermined determination criterion, and when it is determined that they are similar, the two image object regions are converted into one image. Integrate into the object area.

  In one aspect, the feature amount includes the color information of the image object related to the image object region, the size of the circumscribed rectangle of the image object region, the shape of the image object region, the ratio of the image object region in the circumscribed rectangle, the image object This is at least one of attribute information (such as a uniform color area or a photographic area) of the image object related to the area. However, features other than those described above may be used as feature amounts.

  FIG. 18 is a flowchart illustrating an example of the integration process according to the second specific example. Hereinafter, an example of the integration process according to the second specific example will be described with reference to FIG.

  The image processing apparatus 10 specifies a circumscribed rectangle of each extracted image object area (S21).

  Next, the image processing apparatus 10 selects two image object areas from the extracted image object areas, and determines whether or not there is an overlapping relationship between the circumscribed rectangles of the two image object areas (S22). ).

  If it is determined that there is an overlapping relationship (S22: YES), the image processing apparatus 10 determines whether or not the feature amounts of the two image object areas are similar to each other (S23).

  When it is determined that the feature amounts are similar (S23: YES), the image processing apparatus 10 integrates the two image object areas into one image object area (S24), and the process proceeds to step S25.

  On the other hand, when it is determined in step S22 that there is no overlapping relationship (S22: NO), and when it is determined in step S23 that the feature amount is not similar (S23: NO), the image processing apparatus 10 performs the integration process. The process proceeds to step S25 without performing.

  In step S25, the image processing apparatus 10 determines whether or not the processing in steps S22 to S24 has been completed for all combinations of image object regions. If not (S25: NO), the processing is performed. Returning to step S22, if the process is completed (S25: YES), the process is terminated.

  By the above processing, for example, 35 image object areas of the design of FIG. 17 are integrated into one image object area.

[Third example]
FIG. 19 shows an example of symbols included in the input image. In FIG. 19, a white area represents a background area, and five gray areas on the white area represent colored areas. The five colored regions are extracted as separate image object regions by the region extracting unit 13. However, in FIG. 19, since five colored areas are collected to form one image object, it is desirable that the five colored areas are extracted as one image object area. Here, in the case of the design as shown in FIG. 19, the circumscribed rectangle of each region is close to the circumscribed rectangle of the adjacent region, and the features (colors, etc.) of the two regions are similar to each other. Therefore, when the circumscribed rectangles of the two image object areas are close to each other and the characteristics of the two image object areas are similar to each other, the two image object areas may constitute one image object area. high.

  In view of the above circumstances and the like, in this example, the region integration unit 14 is such that the circumscribed rectangles of the two image object regions of the image object regions extracted by the region extraction unit 13 are close to each other by a predetermined degree or more. When the features of two image object areas are similar to each other by a predetermined degree or more, the two image object areas are integrated into one image object area.

  In one aspect, the region integration unit 14 identifies a circumscribed rectangle of each image object region extracted by the region extraction unit 13, and determines the degree of proximity of the circumscribed rectangles of any two image object regions and the two image objects. Each feature amount of the region is obtained, and the two image object regions are integrated into one image object region according to the proximity degree and the feature amount. For example, the region integration unit 14 obtains a similarity degree between two feature amounts, and when the proximity degree X of the circumscribed rectangle is equal to or greater than a predetermined value X1 and the similarity degree Y of the feature amount is equal to or greater than a predetermined value Y1, Merge two image object areas. In this case, the predetermined value Y1 may be changed according to the proximity degree X so that the predetermined value Y1 decreases as the proximity degree X increases. Further, the predetermined value X1 may be changed according to the similarity degree Y so that the predetermined value X1 decreases as the similarity degree Y increases. Further, the region integration unit 14 may integrate two image object regions when the product X · Y of the proximity degree X and the similarity degree Y is equal to or greater than a predetermined value K. The proximity degree X is calculated based on, for example, the distance between the circumscribed rectangles in the horizontal and vertical directions and the overlapping width (or degree of overlap) of the circumscribed rectangles in the horizontal and vertical directions.

  In another aspect, the region integration unit 14 specifies a circumscribed rectangle of each image object region extracted by the region extracting unit 13, and establishes a proximity relationship in which a circumscribed rectangle of any two image object regions is close within a predetermined distance. investigate. When the proximity relationship is established, the feature amounts of the two image object regions are obtained, the two feature amounts are compared, and the two image object regions are converted into one image object according to the comparison result. Integrate into the area. Specifically, the region integration unit 14 determines whether or not the two feature amounts are similar according to a predetermined determination criterion, and when it is determined that they are similar, the two image object regions are converted into one image. Integrate into the object area. In this aspect, when the proximity relationship is established, the region integration unit 14 compares the two feature amounts, obtains the proximity degree of the circumscribed rectangle of the two image object regions, and compares the feature amount comparison result and the proximity degree. Accordingly, the two image object areas may be integrated into one image object area.

  In one aspect, the feature amount includes the color information of the image object related to the image object area, the size of the circumscribed rectangle of the image object area, the shape of the image object area, the ratio of the image object area in the circumscribed rectangle, the image object This is at least one of attribute information (such as a uniform color area or a photographic area) of the image object related to the area. However, features other than those described above may be used as feature amounts.

  FIG. 20 is a flowchart illustrating an example of the integration process according to the third specific example. Hereinafter, an example of the integration process according to the third specific example will be described with reference to FIG.

  The image processing apparatus 10 specifies a circumscribed rectangle of each extracted image object area (S31).

  Next, the image processing apparatus 10 selects two image object areas from the extracted image object areas, and determines whether or not there is a proximity relationship between the circumscribed rectangles of the two image object areas (S32). ). For example, the image processing apparatus 10 calculates a distance between circumscribed rectangles and determines whether or not the distance is a predetermined value or less.

  Then, when it is determined that there is a proximity relationship (S32: YES), the image processing apparatus 10 obtains the feature amounts of the two image object regions (S33). Further, the image processing apparatus 10 obtains the degree of proximity of the circumscribed rectangle between the two image object areas (S34).

  Next, the image processing apparatus 10 determines whether or not the two image object areas should be integrated based on the feature amount and the proximity degree (S35). For example, the image processing apparatus 10 calculates the similarity degree of the feature amount, and determines whether or not the product of the similarity degree and the proximity degree is a predetermined value or less.

  If it is determined that they should be integrated (S35: YES), the image processing apparatus 10 integrates the two image object areas into one image object area (S36), and advances the process to step S37.

  On the other hand, if it is determined in step S32 that there is no proximity relationship (S32: NO) and if it is determined in step S35 that it should not be integrated (S35: NO), the image processing apparatus 10 performs the integration process. Without performing this, the process proceeds to step S37.

  In step S37, the image processing apparatus 10 determines whether or not the processing in steps S32 to S36 has been completed for all combinations of image object regions. If not (S37: NO), the processing is performed. Returning to step S32, if the processing is completed (S37: YES), the processing is ended.

  By the above processing, for example, five image object areas of the design of FIG. 19 are integrated into one image object area.

  In addition, this invention is not limited to the said embodiment, It can change variously within the range which does not deviate from the summary of this invention.

It is a block diagram which shows an example of a structure of the image processing apparatus which concerns on embodiment. It is a figure which shows an example of an input image. It is a figure which shows an example of the binary image obtained by binarizing the image of FIG. 2 with a background color. It is a figure which shows an example of the labeling image obtained by labeling the image of FIG. FIG. 5 is a diagram illustrating an example of a labeling image obtained by performing integration processing on the image of FIG. 4. It is a figure which shows an example of a mask image. It is a figure which shows an example of a cropped image. It is a figure which shows an example of an image object. It is a figure which shows an example of the mask image in case integration processing is not implemented. 6 is a flowchart illustrating an example of an operation of the image processing apparatus according to the embodiment. It is a figure for demonstrating the example of determination whether a to-be-included area adjoins to an inclusion area. It is a figure which shows an example of a contact with the image edge part of a white character. It is a figure which shows an example of the outline character of the Chinese character "da". It is a figure for demonstrating the scan to an inclusion area | region, and an additional scan. It is a flowchart which shows an example of the integration process which concerns on a 1st specific example. It is a figure for demonstrating the integration process which concerns on a 1st specific example. It is a figure which shows an example of the pattern contained in an input image for demonstrating a 2nd specific example. It is a flowchart which shows an example of the integration process which concerns on a 2nd specific example. It is a figure which shows an example of the pattern contained in an input image for demonstrating a 3rd specific example. It is a flowchart which shows an example of the integration process which concerns on a 3rd specific example.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 Image processing apparatus, 11 Image reception part, 12 Background color specific part, 13 Area extraction part, 14 Area integration part, 15 Image object output part

Claims (8)

Image receiving means for receiving images;
Background color specifying means for specifying the background color of the image;
An area extracting means for extracting each area partitioned by the background color area from the image as an image component area;
When the relationship between at least two image component regions of the extracted image component regions satisfies a predetermined condition to be treated as one image component region, the at least two image component regions are set to 1 Region integration means for integrating into one image component region;
I have a,
The region integration means is configured such that a circumscribed rectangle of an image component region (hereinafter referred to as “included region”) among the extracted image component regions is defined as another image component region (hereinafter referred to as “include region”). The inclusion area and the inclusion area are integrated into a single image component area, and the inclusion area is included in a single image component area. For each of the four sides of the circumscribed rectangle of the inclusion area, it is determined whether the side is close to the inclusion area, and when it is determined that a predetermined number of sides are close to the inclusion area, the inclusion area is An image processing apparatus that determines to be close to the inclusion area . The image processing apparatus according to claim 1 ,
The region integration means, for each pixel on the side, inspects whether the interval between the pixel and the inclusion region is equal to or less than a predetermined interval in a direction perpendicular to the side, and based on the inspection result, An image processing apparatus that determines whether or not the side is close to the inclusion area. The image processing apparatus according to claim 2 ,
The image is characterized in that the region integration unit determines that the side is close to the inclusion region when the number of pixels having the interval equal to or smaller than the predetermined interval among the pixels on the side is equal to or larger than the predetermined number. Processing equipment. The image processing apparatus according to claim 1 ,
For each pixel on the side, the region integration unit is configured such that, in a direction perpendicular to the side, the interval between the pixel and the inclusion region is equal to or less than a predetermined interval, and the width of the inclusion region facing the pixel is An image processing apparatus that inspects whether or not a width is equal to or greater than a predetermined width and determines whether or not the side is close to the inclusion region based on the inspection result. The image processing apparatus according to claim 4 ,
The region integration unit may determine that the side is close to the inclusion region when the number of pixels of the pixels on the side is equal to or smaller than the predetermined interval and the width is equal to or larger than the predetermined number. An image processing apparatus characterized by determining. An image processing apparatus according to any one of claims 2 to 5 ,
The area integration means changes the predetermined interval according to the size of the circumscribed rectangle of the inclusion region so that the predetermined interval increases as the circumscribed rectangle of the inclusion region increases. Processing equipment. The image processing apparatus according to any one of claims 4 to 6 ,
The area integration unit changes the predetermined width according to the size of the circumscribed rectangle of the inclusion area so that the predetermined width increases as the circumscribed rectangle of the inclusion area increases. Processing equipment. On the computer,
The procedure for accepting images,
Identifying the background color of the image;
A procedure for extracting each area partitioned by the background color area from the image as an image component area;
When the relationship between at least two image component regions of the extracted image component regions satisfies a predetermined condition to be treated as one image component region, the at least two image component regions are set to 1 The steps to merge into one image component area,
An image processing program for causing the execution,
In the integration procedure, a circumscribed rectangle of an image component region (hereinafter referred to as “included region”) among the extracted image component regions is referred to as another image component region (hereinafter referred to as “include region”). The inclusion area and the inclusion area are integrated into a single image component area, and the inclusion area is included in a single image component area. For each of the four sides of the circumscribed rectangle of the inclusion area, it is determined whether the side is close to the inclusion area, and when it is determined that a predetermined number of sides are close to the inclusion area, the inclusion area is An image processing program that determines to be close to the inclusion area.

Images