JP6145983B2 - Image processing apparatus and computer program - Google Patents

Description

  The present invention relates to image processing for image data representing an image including a character region representing a character and a non-character region different from the character region.

  A technique for compressing target image data representing a target image including characters at a high compression rate is known (for example, Patent Document 1). This technique separates color character image data representing color characters, monochrome character image data representing monochrome characters, and non-character image data not containing characters. This technique improves the compression rate of the entire target image data by compressing each separated image by a compression method suitable for compression of each image.

  For example, monochrome character image data is compressed in MMR (Modified Modified Read) format, color character image data is compressed in GIF (Graphic Interchange Format) format, and non-character image data is compressed in JPEG (Joint Photographic Experts Group) format. It is compressed with.

JP 2005-1224066 A

  However, in the above technique, it cannot be said that the compression of the non-character image data is sufficiently devised, and there is a possibility that the target image data cannot be effectively compressed. For example, there is a possibility that the target image data in which the ratio of the character image data is smaller than the ratio of the non-character image data cannot be reduced effectively because the data size cannot be reduced sufficiently.

  A main advantage of the present invention is to provide a new technique for effectively compressing image data representing an image including a character region representing a character and a non-character region different from the character region.

  SUMMARY An advantage of some aspects of the invention is to solve at least a part of the problems described above, and the invention can be implemented as the following application examples.

Application Example 1 An acquisition unit that acquires target image data that represents a target image and includes a plurality of component data corresponding to a plurality of color components;
In the target image, a specifying unit that specifies a character region representing a character, a gray region representing a gray image different from the character, and a color region representing a color image different from the character,
A reduction processing unit for reducing the data size of the target image data,
A first processing unit that generates compressed character image data by executing a first process using partial image data corresponding to the character region;
A second process different from the first process is executed using the partial image data corresponding to the gray area, thereby generating compressed gray image data composed of one type of component value. A second processing unit, wherein the second processing is performed by performing processing for obtaining gray component data composed of one kind of component value and processing for compressing the gray component data; Generating the gray image data, the second processing unit;
By executing a third process different from the first process and the second process using partial image data corresponding to the color area, a compressed color composed of a plurality of types of component values A third processing unit for generating image data, wherein the third process performs a process of acquiring color component data including a plurality of component data and a process of compressing the color component data. Generating the compressed color image data, the third processing unit;
The reduction processing unit,
A generating unit that generates compressed target image data representing the target image using the compressed character image data, the compressed gray image data, and the compressed color image data;
An image processing apparatus comprising:

  According to the above configuration, the compressed character image data, the compressed gray image data, and the compressed color image data are generated using different processes. The compressed gray image data is generated by a second process including a process of acquiring gray component image data composed of one type of component value and a process of compressing the gray component image data. The compressed color image data is generated by a third process including a process of acquiring color image data including a plurality of component data and a process of compressing the color image data. As a result, processing suitable for a character, a gray image different from the character, and a color image different from the character are executed, so that the target image data is effectively compressed to generate compressed target image data. Can do.

  The present invention can be realized in various forms, for example, in the form of a method for realizing the function of the device, a computer program for realizing the function of the device, a recording medium on which the computer program is recorded, and the like. Can be realized.

It is a block diagram which shows the structure of the computer 200 as one Example of this invention. It is a flowchart of an image process. It is the schematic which shows the whole flow of an image area | region identification process. The calculation formula of the edge strength in the present embodiment is shown. It is a flowchart of an integration process. It is the schematic which shows integration of two area | regions. The calculation formula of the gradation difference TD under the third condition F3 is shown. It is the schematic which shows integration of four area | regions L34-L37. It is a figure which shows an example of the judgment table 292. It is explanatory drawing of the distribution width W and the number C of colors. It is a flowchart of a gray determination process. It is a flowchart of creation of a histogram. It is a figure which shows an example of the outer edge part of a non-character object area | region. It is a figure explaining judgment of a gray area and a color area. It is a flowchart of the reduction process of 1st Example. It is a figure explaining the image data produced | generated by the reduction process of 1st Example. It is a flowchart of the reduction process of 2nd Example. It is a figure explaining the image data produced | generated by the reduction process of 2nd Example.

A. First embodiment:
A-1: Configuration of Computer 200 Next, an embodiment of the present invention will be described based on examples. FIG. 1 is a block diagram showing a configuration of a computer 200 as an embodiment of the present invention. The computer 200 is, for example, a personal computer, a CPU 210, a volatile storage device 240 including a DRAM, a non-volatile storage device 290 including a flash memory and a hard disk drive, an operation unit 270 such as a touch panel and a keyboard, And a communication unit 280 which is an interface for communicating with an external device.

  The computer 200 is communicably connected to an external device (here, the scanner 300 and the multifunction device 400) via the communication unit 280. The scanner 300 is an image reading device that acquires scan data by optically reading an object (for example, a paper document). The multifunction device 400 includes an image reading unit (not shown) that acquires scan data by optically reading an object.

  The volatile storage device 240 is provided with a buffer area 241 for temporarily storing various intermediate data generated when the CPU 210 performs processing. The nonvolatile storage device 290 stores a driver program 291 and a determination table 292 used in image processing to be described later. The driver program 291 is provided in a form stored in, for example, a CD-ROM or DVD-ROM. Alternatively, the driver program 291 is provided in a form downloaded from a server connected to the computer 200 via a network.

  The CPU 210 functions as the scanner driver 100 by executing the driver program 291. The scanner driver 100 has an image processing function for executing a later-described image process on target image data (for example, scan data) to generate a compressed PDF file. The scanner driver 100 includes an acquisition unit 110, an identification unit 120, a reduction processing unit 130, and a generation unit 150 as functional units for realizing the image processing function. The acquisition unit 110 acquires target image data. The specifying unit 120 analyzes the target image data and executes region specification. The specifying unit 120 includes a determination unit 125 that performs a gray area determination process. The reduction processing unit 130 executes a reduction process for reducing the data size of the target image data. The reduction processing unit 130 includes a first processing unit 131 that performs processing related to a character region representing a character in the target image, and a second processing unit that performs processing related to a gray region representing a gray image different from the character included in the target image. 132, and a third processing unit 133 that performs processing related to a color region representing a color image different from the character included in the target image. The generation unit 150 generates a compressed PDF file using the image data processed by the reduction processing unit 130. Processing executed by each of these functional units 110 to 150 will be described later.

A-2: Image Processing FIG. 2 is a flowchart of image processing. In step S100, the acquisition unit 110 (FIG. 1) acquires scan data as target image data via the communication unit 280, and stores the acquired scan data in the buffer area 241 (FIG. 1). Specifically, the acquisition unit 110 controls the image reading unit of the scanner 300 or the multifunction device 400 to acquire scan data. The scan data represents, for example, a reading result of a paper document (also called a manuscript). The scan data is bitmap data that represents the color of each of the plurality of pixels as an RGB value.

  FIG. 3 is a schematic diagram showing an overall flow of the image region specifying process (the processes in steps S150 to S450 in FIG. 2). The target image SI in FIG. 3A is an example of an image represented by target image data (for example, scan data). In the target image SI, a plurality of pixels (not shown) are arranged in a matrix along a first direction D1 and a second direction D2 orthogonal to the first direction D1. Pixel data of one pixel (that is, RGB value) is, for example, the gradation values (hereinafter also referred to as component values) of three color components of red (R), green (G), and blue (B). Is included. In this embodiment, the number of gradations of each component value is 256 gradations. That is, it can be said that the target image data includes three component data respectively corresponding to the three RGB components. The component data is image data composed of one component value among the three component values.

  In the example of FIG. 3A, the target image SI includes a background Bg, three non-character objects Ob1 to Ob3, and four character objects Ob4 to Ob7. The non-character object is, for example, a photo object or a drawing object. The photographic object is an object representing a photograph obtained by photographing with a digital camera. The drawing object is an object representing drawing, such as an illustration, a table, a graph, a diagram, vector graphics, or a pattern. Of the three non-character objects of the present embodiment, two non-character objects Ob1 and Ob2 are gray (achromatic) objects, and one non-character object Ob3 is a color (chromatic) object. . Of the four character objects Ob4 to Ob7 in this embodiment, three character objects Ob4 to Ob6 (characters E, F, and G) are the same color characters, and one character object Ob7. It is assumed that (letter H) is a character of a different color from the other three character objects Ob4 to Ob6.

  Here, the gray object is an object that looks gray when the observer observes at a normal observation distance. For example, a case where the document is a document printed using three types of printing materials of CMY will be described as an example. In the scanned image obtained by scanning the original, each pixel representing an object that appears achromatic when viewed at a normal observation distance has a chromatic color, but the object that appears achromatic is It is included in the gray object in the embodiment. The color object is an object that looks chromatic when the observer observes at a normal observation distance.

  Here, in the target image SI, a partial image representing a character object is also called a character image, a partial image representing a photographic object is also called a photographic image, and a partial image representing a drawing object is also called a drawing image. Photo images and drawn images are collectively referred to as non-character images. A partial image representing a gray object is also called a gray image, and a partial image representing a color object is also called a color image.

  In step S150 of FIG. 2, the specifying unit 120 (FIG. 1) generates edge image data using the target image SI (scan data) and stores it in the buffer area 241. FIG. 3B is a schematic diagram of the edge image EI represented by the edge image data.

  The edge image EI represents the edge intensity at each pixel position in the target image SI. The edge strength represents the magnitude (for example, differentiation) of the gradation value with respect to the change in position in the image, that is, the magnitude of the difference in gradation value between a plurality of adjacent pixels. FIG. 4 shows a formula for calculating the edge strength in this embodiment. In the present embodiment, the specifying unit 120 calculates the edge strength Se for each of the three color components of RGB using a so-called Sobel operator.

  The gradation value P (x, y) in FIG. 4 represents the gradation value at a specific pixel position (x, y) in the target image SI. The position x indicates the pixel position in the first direction D1, and the position y indicates the pixel position in the second direction D2. As shown in the drawing, the edge intensity Se (x, y) at the pixel position (x, y) in the target image SI is obtained by calculating nine pixels in three rows and three columns adjacent to each other with the pixel position (x, y) as the center. Is used to calculate. The first and second terms of the calculation formula of FIG. 4 are absolute values of the sum of values obtained by multiplying the gradation values of the pixels at nine positions by the corresponding coefficients, respectively. The first term is the differentiation of the gradation value in the first direction D1 (that is, the differentiation in the horizontal direction), and the second term is the differentiation of the gradation value in the second direction D2 (that is, the differentiation in the vertical direction). is there.

  The edge image EI in FIG. 3B is obtained by averaging the edge strength of the R component, the edge strength of the G component, and the edge strength of the B component at each pixel position (hereinafter referred to as reference edge strength). ). Dotted lines Eg1 to Eg7 in FIG. 3B represent the positions of pixels (also referred to as edge pixels) having a relatively high reference edge strength. It can be seen that the edge image EI in FIG. 3B includes edge pixels Eg1 to Eg7 corresponding to the objects Ob1 to Ob7 of the target image SI, respectively.

  After generating the edge image data, in the subsequent step S200, the specifying unit 120 sets a block BL including a plurality of pixels on the edge image EI. The broken lines in FIG. 3B indicate the blocks BL arranged in a matrix on the edge image EI. One block BL is, for example, a block composed of pixels PX of BLn rows × BLn columns (BLn is an integer of 2 or more). As the value of BLn, for example, a value within the range of 10 to 50 can be adopted. Since the edge image EI and the target image SI have the same size (the same number of vertical and horizontal pixels), it can be said that the block BL is set on the target image SI.

  When the block BL is set, in the subsequent step S250, the specifying unit 120 specifies a solid area and a non-solid area in units of the block BL. A solid region is a region where the edge strength of the region is less than a predetermined reference, and a non-solid region is a region where the edge strength of the region is greater than or equal to a predetermined reference. Specifically, the specifying unit 120 calculates an average edge strength (ERave, EGave, EBave) for each block BL. The average edge strength (ERave, EGave, EBave) is calculated for each of the three RGB color components. The identifying unit 120 compares the average edge strength of the processing target block BL with a predetermined reference, and classifies the processing target block BL as either a solid block or a non-solid block. A solid block is a block BL whose average edge strength is smaller than a predetermined reference. A non-solid block is a block BL whose average edge strength is greater than or equal to a predetermined reference. In the present embodiment, the specifying unit 120 compares the average edge strength (ERave, EGave, EBave) with reference values (ETr, ETg, ETb) determined for each color component. As a result, when ERave <ETr, EGave <ETg, and EBave <ETb are satisfied, the specifying unit 120 classifies the processing target block BL as a solid block. When at least one of ERave ≧ ETr, EGave ≧ ETg, and EBave ≧ ETb is satisfied, the specifying unit 120 classifies the processing target block BL as a non-solid block.

  In the edge image EI of FIG. 3B, the non-solid block is hatched and the solid block is not hatched. After all the blocks BL are classified into a solid block and a non-solid block, the specifying unit 120 defines an area corresponding to one or more non-solid blocks adjacent to each other (continuous) as one non-solid area. As specified. Further, the specifying unit 120 specifies a region corresponding to one or more solid blocks adjacent to each other as one solid region. Thus, since one or more continuous non-solid blocks are incorporated into one non-solid area, the non-solid area is usually surrounded by a solid area. In the example of FIG. 3B, three non-solid regions L11 to L13 corresponding to the three non-character objects Ob1 to Ob3 of the target image SI (FIG. 3A) are specified. Further, one non-solid region L14 corresponding to the two character objects Ob4 and Ob5 of the target image SI and one non-solid region L15 corresponding to the two character objects Ob6 and Ob7 are specified. Yes. Furthermore, one solid area L10 corresponding to the background Bg of the target image SI is specified. Specifying a solid region and a non-solid region in the edge image EI has the same meaning as specifying a solid region and a non-solid region in the target image SI.

  In step S300, the specifying unit 120 uses a reference value for binarizing each non-solid region in the target image SI (hereinafter also referred to as a binarization reference value) as a non-solid region in the target image SI. Is determined for each of the non-solid regions L11 to L15 using pixel values (in other words, color values) in the solid region surrounding the periphery of the region. In this embodiment, the binarization reference value is determined for each RGB component. Specifically, the average value (Rr, Gr, Br) of the RGB component values for all the pixels in the solid area surrounding the non-solid area is adopted as the binarization reference value. In the example of FIG. 3B, since all the non-solid regions L11 to L15 are surrounded by one solid region L10 corresponding to the background Bg, the binarization criterion for all the non-solid regions L11 to L15. The value is the same value, that is, the average value of the component values in the solid region L10.

When the binarization reference values (Rr, Gr, Br) are determined, in the next step S350, the specifying unit 120 generates binary image data for each of the non-solid regions L11 to L15, and the buffer region 241. To store. In the present embodiment, the specifying unit 120 performs binarization processing using six threshold values R1, R2, G1, G2, B1, and B2 calculated using the binarization reference values (Rr, Gr, and Br). Execute.
R component lower limit threshold R1 = Rr−dV, R component upper limit threshold R2 = Rr + dV
G component lower limit threshold G1 = Gr−dV, G component upper limit threshold G2 = Gr + dV
B component lower limit threshold B1 = Br−dV, B component upper limit threshold B2 = Br + dV
Here, the value dV is a predetermined value. These values R1, R2, G1, G2, B1, and B2 are binarization reference values (Rr, Gr, Br), that is, colors that are relatively close to the average color of the solid area surrounding the binarization target solid area. That is, a color range relatively close to the background color.

  The specifying unit 120 uses these six threshold values R1, R2, G1, G2, B1, and B2 to set each pixel in the non-solid region in the target image SI to an object pixel and a non-object for each pixel. By classifying it into pixels, binary image data of a non-solid region is generated. For example, in the generated binary image data, the pixel value “1” indicates an object pixel, and the pixel value “0” indicates a non-object pixel.

Specifically, when the gradation values (Ri, Gi, Bi) of the three color components (RGB) of the pixel Pxi in the non-solid region satisfy all the following three conditions, the specifying unit 120: The pixel Pxi is classified as a non-object pixel, and when any of the following three conditions is not satisfied, the pixel Pxi is classified as an object pixel.
(First condition) R1 <Ri <R2
(Second condition) G1 <Gi <G2
(Third condition) B1 <Bi <B2

  In this way, an object is configured by classifying pixels that are relatively close to the background color calculated using the color of the pixels in the solid area as non-object pixels and classifying other pixels as object pixels. It is possible to generate binary image data that accurately identifies object pixels.

  FIG. 3C shows a binary image BI represented by the generated binary image data. Actually, separate binary image data is generated for each of the non-solid regions L11 to L15 described above, but in FIG. 3C, it is represented by one binary image BI.

  After the binary image data is generated, in the subsequent step S400, the specifying unit 120 uses the binary image data to specify the object region and the non-object region, and assigns an identifier to the specified region. Execute the attached labeling. As a result of labeling, for example, label data in which each area is associated with an identifier is generated and stored in the buffer area 241.

  Specifically, the specifying unit 120 converts one region including one or more continuous object pixels (that is, a pixel whose gradation value after binarization is “1”) to one Identifies as an object area. In addition, the specifying unit 120 converts one area formed of one or more continuous non-object pixels (that is, a pixel whose gradation value after binarization is “zero”) into one non-object. Specify as an area.

  In the example of FIG. 3C, seven object regions L21 to L27 respectively corresponding to the six objects Ob1 to Ob7 (FIG. 3A) of the target image SI and 1 corresponding to the background Bg of the target image SI. Pieces of non-object regions L20 are identified. The identification unit 120 assigns an identifier for identifying the area to the identified area. Since each pixel constituting the binary image BI corresponds to each pixel constituting the target image SI, it is shown in FIG. 3D that the regions L20 to L27 are specified in the binary image BI. As described above, in the target image SI, similarly, the regions L30 to L37 are specified. In the following description, the symbols L30 to L37 shown in FIG. 3D are basically used as symbols representing the respective regions (that is, the object region and the non-object region). Further, when simply calling an image corresponding to an object region, it refers to a corresponding partial image of the target image SI, and a pixel value of a pixel in the object region is a pixel value of a corresponding pixel of the target image SI, that is, The pixel value (for example, RGB value) corresponding to the target image data is indicated.

  Subsequent to the labeling, in step S450, the specifying unit 120 executes an integration process for integrating a plurality of object areas satisfying the integration condition among the plurality of labeled object areas. This integration processing is processing for specifying a plurality of character areas separated as different object areas as one object area.

  FIG. 5 is a flowchart of the integration process. In step S4500, the specifying unit 120 selects a background region from a plurality of regions specified by the specifying unit 120 (for example, the regions L30 to L37 in FIG. 3D). The background region is a solid region corresponding to the edge portion of the target image SI (FIG. 3A). In the example of FIG. 3D, the non-object region L30 is selected as the background region. This background region L30 is excluded from integration targets.

  After the background area is selected, the specifying unit 120 (FIG. 1) selects one unprocessed area as the process target area N in step S <b> 4505. Next, in step S4510, the specifying unit 120 determines whether or not the number of pixels in the processing target area N is equal to or less than a predetermined pixel number reference. The pixel number standard is determined in advance. For example, when the processing target area N represents a character to be integrated with another area, a value slightly larger than the maximum value that the number of pixels of the processing target area N can take can be adopted as the pixel number standard. The pixel number reference is set in advance to a value slightly larger than the maximum value that the number of pixels when the processing target area N represents one character. When the number of pixels in the processing target area N exceeds the pixel number reference (step S4510: NO), the specifying unit 120 returns to step S4505. As a result, the selected processing target area N is excluded from integration targets. . In this case, since the current processing target area N is larger than a typical character, there is a high possibility that it represents an object other than a character. By appropriately setting the pixel number reference, in the example of FIG. 3C, the three object regions L31 to L33 representing the non-character object are excluded from the integration targets, and the four object regions representing the character object are displayed. The object areas L34 to L37 are targeted for integration. In the present embodiment, the number of pixels in the processing target area N is the number of pixels included in the minimum rectangle circumscribing the processing target area N in the target image SI.

  FIG. 6 is a schematic diagram showing the integration of two regions. In the figure, a processing target area Ln representing the character “E” is shown. A rectangle LnR in the drawing is a minimum rectangle circumscribing the processing target region Ln. The number of pixels included in the rectangle LnR is the number of pixels in the processing target area Ln. Here, the “minimum rectangle circumscribing the region” is the following rectangle. That is, the rectangle is composed of two sides parallel to the first direction D1 and two sides parallel to the second direction D2. The upper side of the rectangle is in contact with the upper end of the region, the lower side of the rectangle is in contact with the lower end of the region, the left side of the rectangle is in contact with the left end of the region, and the right side of the rectangle is in contact with the right end of the region. Here, the upper side and the upper end are sides and ends on the opposite side of the second direction D2, the lower side and the lower end are sides and ends on the second direction D2, and the left side and the left end are in the first direction D1. The sides and ends on the opposite direction side, and the right side and the right end are sides and ends on the first direction D1 side. The specifying unit 120 may calculate the number of pixels in the processing target area N by counting only the pixels in the processing target area N. That is, the specifying unit 120 may calculate the number of pixels without counting pixels that are not included in the processing target area N among the plurality of pixels in the circumscribed rectangle.

  In S4510 of FIG. 5, when the number of pixels in the processing target area N is equal to or smaller than a predetermined pixel number reference (Step S4510: YES), in Step S4515, the specifying unit 120 initializes a list of candidate areas M for integration. To do. The specifying unit 120 generates a list of areas that have not been selected as the processing target area N in step S4505. For example, in the example of FIG. 3D, when the region L31 is selected as the processing target region N in step S4505 executed for the first time, the remaining six regions L32 to L37 are listed. When the region L32 is selected as the processing target region N in the next step S4505, the remaining five regions L33 to L37 are listed. Note that areas that have been integrated with other areas are excluded from the list.

  Next, in step S4520, the specifying unit 120 selects one unselected area as a candidate area M from the generated list. The specifying unit 120 determines whether or not to integrate the candidate area M into the process target area N in the following three steps S4525, S4530, and S4535. In steps S4525, S4530, and S4535, the following conditions are determined.

(Step S4525: First Condition F1) Number of Pixels in Candidate Region M ≦ Pixel Number Reference (Step S4530: Second Condition F2) First Distance Dis1 ≦ Distance Reference and Second Distance Dis2 ≦ Distance Reference (Step S4535: First 3 condition F3) gradation difference TD ≦ gradation difference reference

  When the candidate area M satisfies all these conditions F1, F2, and F3 (step S4525: YES, S4530: YES, and S4535: YES), the specifying unit 120 selects candidates in step S4540 of FIG. The region M is integrated into the processing target region N.

  The first condition F1 in step S4525 is the same as the condition in step S4510. When the candidate area M does not satisfy the first condition F1 (step S4525: NO), it is highly likely that the candidate area M represents an object of a different type from characters. In this case, the specifying unit 120 does not integrate the candidate area M into the process target area N by skipping step S4540.

  The second condition F2 in step S4530 is a condition that is satisfied when the candidate area M is relatively close to the process target area N. FIG. 6 shows an outline of the first distance Dis1 and the second distance Dis2 under the second condition F2. In the figure, a processing target area Ln and a candidate area Lm are shown. The target rectangle LnR is the smallest rectangle circumscribing the processing target region Ln, and the candidate rectangle LmR is the smallest rectangle circumscribing the candidate region Lm.

  As shown in FIG. 6A, the first distance Dis1 is the shortest distance along the first direction D1 between the target rectangle LnR and the candidate rectangle LmR, and is represented by, for example, the number of pixels. As shown in FIG. 6B, the range of the position of the target rectangle LnR in the first direction D1 (left end PnL to right end PnR) is the range of the position of the candidate rectangle LmR in the first direction D1 (left end PmL to right end PmR). The first distance Dis1 is zero when it overlaps at least a part of.

  As shown in FIG. 6B, the second distance Dis2 is the shortest distance along the second direction D2 between the target rectangle LnR and the candidate rectangle LmR, and is represented by, for example, the number of pixels. As shown in FIG. 6A, the range of the position of the target rectangle LnR in the second direction D2 (that is, the range of the upper end PnT to the lower end PnB) is the range of the position of the candidate rectangle LmR in the second direction D2 (that is, The second distance Dis2 is zero when it overlaps at least part of the upper end PmT to the lower end PmB.

  The distance reference of the second condition F2 is determined in advance. For example, as the distance reference, a value slightly exceeding the maximum value that the distance between two characters to be integrated can take can be adopted. When the candidate area M satisfies the second condition F2, it is highly likely that the candidate area M and the processing target area N represent characters included in the same character string. When the candidate area M does not satisfy the second condition F2 (step S4530: NO), it is highly likely that the candidate area M represents an object that is not related to the processing target area N. In this case, the specifying unit 120 does not integrate the candidate area M into the process target area N by skipping step S4540.

  The third condition F3 in step S4535 is satisfied when the color of the candidate area M is relatively close to the process target area N. FIG. 7 shows a formula for calculating the gradation difference TD under the third condition F3. In the present embodiment, the gradation difference TD is calculated by comparing the average color of the processing target area N (ie, Rav_n, Gav_n, Bav_n) and the average color of the candidate area M (ie, Rav_m, Gav_m, Bav_m) in the RGB color space. Is the square of the Euclidean distance between. The gradation difference reference for the third condition F3 is determined in advance. For example, as the gradation difference reference, the upper limit value of the color difference between two colors recognized by a normal observer can be adopted as substantially the same color. For example, when two colors are used as the colors of two characters, a normal observer who sees the two characters recognizes that the colors of the two characters are the same. In this case, it can be determined that the two colors are substantially the same color. Here, the condition of the third condition F3 is that, as will be described later, when a character image is compressed, a plurality of characters expressed in substantially different colors must be processed separately. It is. When the candidate area M does not satisfy the third condition F3 (step S4530: NO), the specifying unit 120 does not integrate the candidate area M into the process target area N by skipping step S4540.

  After integrating the candidate area M into the process target area N in step S4540 in FIG. 5 or after determining NO in any one of steps S4525, S4530, and S4535, in step S4545, the specifying unit 120 Then, it is determined whether or not the processing of all candidate areas M in the list has been completed. When the unprocessed candidate area M remains (step S4545: NO), the specifying unit 120 returns to step S4520 and executes the processes of steps S4520 to S4540 on the unprocessed candidate area M. When the process for all candidate areas M in the list is completed (step S4545: YES), the specifying unit 120 shifts the process to step S4550.

  In step S4550, the specifying unit 120 determines whether or not the processing target area N has been expanded after step S4515 is executed last, that is, the total number of candidate areas M integrated into the processing target area N is one or more. Or not. When the processing target area N is expanded (step S4550: YES), the specifying unit 120 executes the processes of steps S4515 to S4545 again using the expanded processing target area N. Therefore, the specifying unit 120 can integrate three or more regions.

  FIG. 8 is a schematic diagram showing the integration of the four regions L34 to L37. Here, the integration process proceeds in the order of FIG. 8 (A) to FIG. 8 (C).

  In FIG. 8A, the region L34 representing the character “E” is the processing target region N (FIG. 5: S4505). Since the region L35 representing the letter “F” arranged next to the region L34 satisfies the above conditions F1 to F3, the specifying unit 120 (FIG. 1) integrates the region L35 into the region L34 (step S4540). Further, the region L36 representing the character “G” and the region L37 representing the character “H” do not satisfy the second condition F2 because the distance from the region L34 is long. Therefore, the two regions L36 and L37 are not integrated into the region L34.

  As described above, when the region L35 is integrated into the region L34, in step S4550 of FIG. 5, the specifying unit 120 determines that the processing target region N (for example, the region L34) has been expanded (S4550: YES). ). In subsequent step S4515, the specifying unit 120 generates a list for the expanded region L34B (FIG. 8B) including the region L35. The generated list includes a region L36 and a region L37.

  In FIG. 8B, the expanded region L34B is the processing target region N. Since the region L36 arranged next to the region L34B satisfies the above conditions F1 to F3, the specifying unit 120 integrates the region L36 into the region L34B (step S4540). For the determination of the conditions F1 to F3, a minimum rectangle circumscribing the expanded region L34B (that is, a region representing the characters “E” and “F”) is used. The region L37 is not integrated with the region L34B because the distance from the region L34B is long.

  As described above, when the region L36 is integrated into the expanded region L34B, the specifying unit 120 (FIG. 1) expands the processing target region N (in this case, the region L34B) in step S4550 of FIG. (S4550: YES). In subsequent step S4515, the specifying unit 120 generates a list for the expanded region L34C (FIG. 8C) including the region L36. The generated list includes an area L37.

  In FIG. 8C, the expanded area L34C is the processing target area. The region L37 disposed next to the region L34C satisfies the condition F1 because the distance to the region L34C is relatively short, but does not satisfy the condition F3 because the color is substantially different from the region L34C. Therefore, the specifying unit 120 does not integrate the region L37 into the region L34C. Therefore, after the integration process, two character object areas L34C, which are an integration of the three character object areas L34 to L36 shown in FIGS. 3D and 8A, and a character object area L37. Is specified (FIG. 3D, FIG. 8C).

  If the processing target area N has not been expanded in step S4550 in FIG. 5 (step S4550: NO), in step S4555, the identifying unit 120 determines whether or not the processing for all the areas has been completed. To do. If an unprocessed area remains (step S4555: NO), the identifying unit 120 returns to step S4505. When the processing for all the areas is completed (step S4555: YES), the specifying unit 120 updates the area identifier in step S4560. Specifically, the specifying unit 120 updates the label data generated in S400 of FIG. 2 and stored in the buffer area 241. When the update of the label data ends, the specifying unit 120 ends the integration process.

  After the integration process, in step S500 of FIG. 2, the determination unit 125 (FIG. 1) determines whether the type of image, that is, the type of object in the area is “character” for each of the plurality of areas after the integration process. Object attribute determination processing is performed to determine whether or not.

  FIG. 9 is a diagram illustrating an example of the determination table 292. The determination unit 125 refers to the determination table 292 and executes object attribute determination processing. The determination unit 125 identifies the type according to the color distribution width W, the number of colors C, and the pixel density S. The determination unit 125 generates a histogram as shown in FIG. 10 for each identified object region, and calculates the distribution width W, the number of colors C, and the pixel density S. The generated histogram, distribution width W, number of colors C, and pixel density S are stored in the buffer area 241 and are deleted after the object determination processing, for example.

  FIG. 10 is an explanatory diagram of the distribution width W and the number C of colors. In the figure, a histogram of luminance is shown. This luminance histogram is a luminance histogram calculated from pixel values in a processing target object region (hereinafter referred to as a target region). In this embodiment, the luminance of each pixel is calculated from the gradation value of each pixel (gradation values of three color components of red R, green G, and blue B). As the calculation formula, for example, a calculation formula for calculating the Y component (luminance component) of the YCbCr color space from the RGB gradation values is used.

  The color number C is the number of luminance values whose frequency value (that is, the number of pixels) is equal to or greater than a predetermined threshold Th among 256 luminance values from 0 to 255. The histogram of FIG. 10 includes three peaks P1, P2, and P3 that exceed the threshold Th. In the example of FIG. 10, the number of colors C exceeds the width C1 of the portion exceeding the threshold Th of the first peak P1, the width C2 of the portion exceeding the threshold Th of the second peak P2, and the threshold Th of the third peak P3. This corresponds to the sum of the width C3 of the portion. In general, since characters are often expressed with a small number of colors, the number of colors C is relatively small when the target area represents a character object. When the target area represents a non-character object, the number of colors C is relatively large. For example, since a photographic object represents various colors of a photographed subject, the number of colors C is relatively large when the target region represents a photographic object.

  The distribution width W is a difference (width) between the lowest value and the highest value of the luminance values whose frequency value (that is, the number of pixels) is equal to or greater than a predetermined threshold Th. For the same reason as the description of the number of colors C, the distribution width W is relatively small when the target area represents a character object, and the distribution width W is relatively large when the target area represents a non-character object. .

  The pixel density S is the number of object pixels (the number of pixels per unit area) with respect to the total number of pixels in the minimum rectangle circumscribing the target region. In general, characters are written on the background with thin lines having a different color from the background. When the target area represents a character object, the pixel density S is relatively small. When the target area represents a non-character object, the pixel density S is relatively large. For example, since the photo object is likely to occupy most of the circumscribed minimum rectangle, the pixel density S is relatively large when the target region represents the photo object.

  The determination table 292 in FIG. 9 is created in consideration of the above description. Specifically, the determination unit 125 determines that the target region is a character object region representing a character object when a predetermined determination condition is satisfied. When the predetermined determination condition is not satisfied, the determination unit 125 determines that the target region is a non-character object region representing a non-character object.

As is understood from the determination table 292, the predetermined determination condition is that one of the following two conditions is satisfied.
1) The distribution width W is greater than or equal to the distribution width threshold value Wth, the color number C is less than the color number threshold value Cth, and the pixel density S is less than the pixel density threshold value Sth. 2) The distribution width W is less than the distribution width threshold value Wth. In addition, the pixel density S is less than the pixel density threshold Sth.

  In the example of FIG. 3D, the object areas L31, L32, and L33 are determined to be areas representing non-character objects (hereinafter also referred to as non-character object areas), and the object areas L34C and L37 are character objects. (Hereinafter also referred to as a character object region).

  Subsequent to the object attribute determination process, in step S550, the determination unit 125 executes a gray determination process.

FIG. 11 is a flowchart of the gray determination process.
In step S551, the determination unit 125 selects an area determined to be a non-character object area in step S500. In the example of FIG. 3D, three non-character object areas L31 to L33 are sequentially selected.

  In step S552, a smoothing process for smoothing the image in the selected non-character object region is executed. Specifically, the smoothing filter FL (FIG. 11) is applied to the partial image data corresponding to the selected non-character object region in the target image data. The smoothing filter FL is applied to each of the three component data (that is, R component data, G component data, and B component data) included in the partial image data to be processed. That is, the determination unit 125 arranges the smoothing filter FL so that the center position FC of the smoothing filter FL overlaps the target pixel in the image represented by the processing target component data. The determination unit 125 calculates an average value of a plurality of component values in the smoothing filter FL (for example, 5 pixels in the vertical direction × 5 pixels in the horizontal direction) centered on the target pixel. The determination unit 125 changes the component value of the target pixel to the calculated average value. The determination unit 125 sets all the pixels in each component data as the target pixel and executes the same processing. The partial image data after the smoothing process is stored in the buffer area 241.

  The significance of the smoothing process will be described. As described above, in an image represented by scan data obtained by reading a document, each pixel of an image that appears achromatic when observed at a normal observation distance may have a chromatic color. . For example, the document is a document printed using three types of printing materials of CMY. Even when individual pixels have a chromatic color, images that appear gray when viewed at a normal viewing distance (ie, regions) are identified by a relatively narrow region (eg, a normal viewing distance) The average color of the individual pixels within the smallest possible area) is considered to be achromatic. For this reason, by performing the smoothing process, when the individual pixels in the region that appears to be an achromatic color have a chromatic color, the color of these pixels can be brought close to the achromatic color. Therefore, it is possible to improve the accuracy of determining whether or not it is a gray region described later.

  In step S553 following the smoothing process, the determination unit 125 creates a histogram representing the distribution of the * a value and the * b value in the CIELAB color space (hereinafter also simply referred to as LaB color space), which is a device-independent color space. And stored in the buffer area 241. This histogram is, for example, a histogram obtained by counting the number of pixels having the * a value and the number of pixels having the * b value for each * a value and * b value.

FIG. 12 is a flowchart for creating a histogram.
In step S5531, the determination unit 125 selects a pixel in the non-character object region to be processed as a processing target pixel in the target image SI. In step S5532, the determination unit 125 calculates the edge intensity EP of the pixel in the binary image BI (FIG. 3C) corresponding to the processing target pixel. The edge strength EP is calculated using, for example, a calculation formula shown in FIG.

  When the edge strength EP is calculated, in the subsequent step S5533, the determination unit 125 determines whether or not the edge strength EP is equal to or less than the edge reference value Eth. When the edge intensity EP is equal to or less than the edge reference value Eth (step S5533: YES), the determination unit 125 updates the histogram using the pixel value of the processing target pixel. That is, the determination unit 125 converts the pixel value (RGB value) of the processing target pixel into a Lab value, and updates the above-described histogram according to the Lab value.

  When the edge strength EP is larger than the edge reference value Eth (step S5533: NO), the determination unit 125 skips step S5534. That is, when the edge intensity EP of the processing target pixel is larger than the reference value Et, the processing target pixel is not used for creating a histogram.

  FIG. 13 is a diagram illustrating an example of the outer edge portion of the non-character object region. As shown in FIG. 3C, in the binary image BI, the pixel values inside the object area are all “1” (specifically, it is represented by black in FIG. 3C). Therefore, pixels whose edge intensity EP is larger than the reference value Et appear in the outer edge portion of the object region (portion in contact with the non-object region), and do not appear in the region other than the outer edge portion of the object region. Therefore, as shown in FIG. 13, the pixels of the outer edge portions OE1 to OE3 (specifically, hatched regions in FIG. 13) of the non-character object regions L31 to L33 are excluded from the creation of the histogram. That is, the pixels of the outer edge portions OE1 to OE3 of the non-character object regions L31 to L33 are not used for determining whether or not they are gray regions, which are executed in the processes of steps S554 to S556 described later.

  In subsequent step S5535, determination unit 125 determines whether or not all pixels in the non-character object region to be processed have been selected. When all the pixels are selected (step S5535: YES), the determination unit 125 ends the creation of the histogram. If there is an unselected pixel (step S5535: NO), the determination unit 125 returns to step S5531, newly selects an unselected pixel, and repeats the processing of steps S5532 to S5535 described above.

  When the histogram is created, in step S554 in FIG. 11, the determination unit 125 determines that the ratio PN (that is, the chromatic color ratio) occupied by chromatic pixels to all the pixels in the non-character object region to be processed is gray. It is determined whether it is less than the reference value Nth.

FIG. 14 is a diagram for explaining determination of a gray area and a color area.
In the present embodiment, the determination unit 125 determines colors whose distance R to the achromatic color axis (that is, the axis of * a = * b = 0) in the Lab color space is less than a predetermined reference distance Rth (FIG. 14). A color whose distance R from the achromatic color axis (that is, the axis of * a = * b = 0) in the Lab color space is equal to or greater than a predetermined reference distance Rth is a chromatic color. Judge. Here, the distance R is the Euclidean distance in the Lab color space (specifically, it is represented by the formula R ² = (* a) ² + (* b) ² ). The reference distance Rth is a value that can separate the color recognized by the observer as an achromatic color and the color recognized as a chromatic color, and is determined empirically. The determination unit 125 refers to the histogram, calculates the chromatic color ratio PN, and determines whether it is less than the gray determination reference value Nth.

  When the chromatic color ratio PN is less than the gray determination reference value Nth (step S554: YES), the determination unit 125 determines that the non-character object region to be processed is a gray region (step S555). When the chromatic color ratio PN is equal to or greater than the gray determination reference value Nth (step S554: NO), the determination unit 125 determines that the non-character object region to be processed is a color region (step S556). FIG. 14A shows an example of the color distribution of the image in the non-character object area determined to be a gray area, and FIG. 14B shows the inside of the non-character object area determined to be a color area. 2 shows an example of the color distribution of the image. In the example of FIG. 14A, it can be seen that the color of most pixels is distributed in a range (also referred to as an achromatic color range) within the reference distance Rth from the achromatic color axis. On the other hand, in the example of FIG. 14B, it can be seen that a relatively high proportion of pixel colors are distributed in a range (also referred to as a chromatic color range) separated from the achromatic color axis by a reference distance Rth or more. As an example, the gray determination reference value Nth is preferably 0.3 or less, and more preferably 0.15 or less. When determining whether the non-character object region to be processed is a gray region or a color region, the determination unit 125 deletes the partial image data and the histogram data after the smoothing process from the buffer region 241. .

  In subsequent step S557, determination unit 125 determines whether or not all non-character object regions have been selected. If all the non-character object regions are selected (step S557: YES), the determination unit 125 ends the gray determination process. If there is an unselected non-character object area (step S557: NO), the determination unit 125 returns to step S551, newly selects an unselected non-character object area, and the above-described steps S552 to S557. Repeat the process.

  Here, the reason why the pixels of the outer edge portions OE1 to OE3 of the non-character object regions L31 to L33 are not used for determining whether or not the image is a gray image will be described. These outer edge portions OE1 to OE3 are located at the boundary between the object region and the non-object region (background region). These boundary portions may have a color different from the original color of the object area. A general scanner receives light from an original by the image sensor while moving the position of the image sensor relative to the original in the sub-scanning direction, and generates scan data. At this time, the pixel data of the boundary portion can be generated based on both the light from the object area of the document and the light from the background area. Accordingly, when the target image data is scan data as in the present embodiment, the boundary portion may be an intermediate color between the original color of the object area and the color of the background area. In this embodiment, the pixels of the outer edge portions OE1 to OE3 are not used for determining whether or not they are gray regions, thereby accurately determining whether the original color of the object region is an achromatic color or a chromatic color. can do.

  In the example of FIG. 3D, the two non-character object regions L31 and L32 are determined to be gray regions by the gray determination process of FIG. 11, and one non-character object region L33 is a color region. It is judged that there is.

  When the gray determination process is completed, in step S600 of FIG. 2, the reduction processing unit 130 executes a reduction process for reducing the data size of the target image data.

FIG. 15 is a flowchart of the reduction process of the first embodiment.
When the reduction process is started, first, in step S605, one object area (character object area or non-character object area) to be processed is selected. In the example of FIG. 3D, five object areas L31 to L33, L34C, and L37 are sequentially selected one by one.

  When the processing target object area is selected, in the next step S610, the reduction processing unit 130 determines whether or not the processing target object area is a character object area. In the case where the object area to be processed is a character object area (step S610: YES), the reduction processing section 130 causes the first processing section 131 of the reduction processing section 130 to include a character object area within the processing target character object area. A character color value TC representing the color of the character image is determined (step S620). For example, the first processing unit 131 calculates an average value (Rave1, Gave1, Bave1) of RGB component values for all pixels in the character object area to be processed as the character color value TC. The first processing unit 131 may calculate other values as the character color value TC. For example, only the pixels in the character object region excluding the outer edge portion (boundary portion with the background) of the character object region. May be used to calculate the character color value TC.

  When the character color value TC is determined, in the subsequent step S625, the third processing unit 133 of the reduction processing unit 130 uses the pixel of the pixel in the processing target character object region in the target image data stored in the buffer region 241. The value is changed to the background color value BC. That is, the third processing unit 133 deletes the character image in the character object area to be processed from the target image SI. As the background color value BC, for example, an average value (Rave2, Gave2, Bave2) of RGB component values for all pixels in the non-object area (background area) surrounding the character object area to be processed is adopted. . In the example of FIG. 3D, the background color value BC of the character object areas L34C and L37 is calculated using the pixel values of the pixels in the non-object area L30. Other values may be adopted as the background color value BC. For example, the third processing unit 133 compares the background color value BC with the character object area among a plurality of pixels in the non-object area surrounding the character object area to be processed. The background color value BC may be calculated using only a plurality of close pixels.

  In subsequent step S630, the first processing unit 131 obtains binary image data corresponding to the character object region to be processed, and compresses the binary image data by the FLATE compression method. In this step, compressed character image data representing characters in the character object area to be processed is generated. The generated compressed character image data is stored in the buffer area 241 until a compressed PDF file to be described later is generated (FIG. 2: S700). The binary image data corresponding to the character object area represents a binary image corresponding to the smallest rectangle circumscribing the character object. For example, the binary image represented by the binary image data corresponding to the character object region L34C in FIG. 3D is the binary image BI4 in FIG. Further, the binary image represented by the binary image data corresponding to the character object region L37 in FIG. 3D is the binary image BI7 in FIG.

  The FLATE compression method is a reversible compression method used for creating a ZIP file and is suitable for compression of an image having a relatively small number of gradations. If the FLATE compression method is used, binary image data can be compressed at a high compression rate and without reducing the resolution. Since it is considered that the readability is given priority over the color reproducibility, maintaining the resolution has priority over maintaining the gradation. For this reason, in this embodiment, the compressed character image data is generated by compressing the binary image data by the FLATE compression method. When the compressed character image data is generated by the FLATE compression method, the reduction processing unit 130 shifts the process to step S650.

  In step S610 of FIG. 6 described above, if the object area to be processed is not a character object area (step S610: NO), the reduction processing unit 130 indicates that the object area to be processed is a gray area. Whether or not (step S615). As described above, the gray area is an area determined as a gray area in the non-character object area by the gray determination process (FIG. 2: step S550). If the object area to be processed is not a gray area (step S615: NO), that is, if the object area to be processed is a color area, the reduction processing unit 130 shifts the process to step S650.

  When the object region to be processed is a gray region (step S615: YES), the second processing unit 132 of the reduction processing unit 130 acquires one component data representing an image in the gray region to be processed. (Step S635). In the present embodiment, the second processing unit 132 includes three component data (that is, R component data, G component data, and B component) included in the partial image data corresponding to the gray region to be processed among the target image data. Data) is selected. Since an achromatic color (that is, gray) is represented by RGB values having the same three component values, the three component data included in the partial image data corresponding to the gray region are considered to be substantially equal to each other. Therefore, in this step, any component data among the three component data may be selected.

  In subsequent step S640, the third processing unit 133 changes the pixel value of the pixel in the gray area to be processed to the background color value BC in the target image data stored in the buffer area 241. That is, the third processing unit 133 deletes the image in the gray area to be processed from the target image SI. The background color value BC is stored in the non-object region (that is, the background region) surrounding the gray region to be processed in the same manner as the background color value BC used when erasing characters in the character object region in step S625 described above. It is calculated using the pixel value of each pixel.

  In the next step S645, the second processing unit 132 arranges one component data acquired in step S635 on the gray component data. Specifically, the gray component data is one component data having a region having the same size as the target image SI, and in the initial state, all pixel values are values representing white (that is, the maximum luminance value). For example, “255” is set. First, the gray component data in the initial state is prepared in the buffer area 241. In this step, the second processing unit 132 acquires partial data corresponding to the gray area to be processed from the gray component data in step S635. It is replaced with one component data.

  In the next step S647, the second processing unit 132 acquires binary image data corresponding to the gray region to be processed, and arranges it on the gray mask data. Specifically, the gray mask data is binary image data having an area having the same size as that of the target image SI, that is, an area having the same size as that of the gray component data. 0 ”. First, the gray component data in the initial state is prepared in the buffer area 241. In this step, the second processing unit 132 corresponds to the gray area of the partial data corresponding to the gray area to be processed among the gray mask data. Replace with binary image data. The binary image data corresponding to the gray area represents a binary image corresponding to the smallest rectangle circumscribing the gray area (that is, the non-character object). For example, the binary image represented by the binary image data corresponding to the gray region (that is, the non-character object region) L31 in FIG. 3D is the binary image BI1 in FIG. Also, the binary image represented by the binary image data corresponding to the gray region (that is, the non-character object region) L32 in FIG. 3D is the binary image BI2 in FIG. When the binary image data corresponding to the gray area to be processed is acquired and placed on the gray mask data, the reduction processing unit 130 shifts the process to step S650.

  In step S650, the reduction processing unit 130 determines whether all object areas have been selected. If all the object areas have been selected (step S650: YES), the reduction processing unit 130 proceeds with the process to step S655. If there is an unselected object area (step S650: NO), the reduction processing unit 130 returns to step S605, newly selects an unselected object area, and performs the above-described processing of steps S610 to S650. repeat.

FIG. 16 is a diagram for explaining image data generated by the reduction process of the first embodiment.
When each of the above-described processes is completed for the target image SI and the process proceeds to step S655, one color component data (FIG. 16A) and two compressed character image data (FIG. 16B) )), One gray component data (FIG. 16C), and one gray mask data (FIG. 16D) are generated and stored in the buffer area 241 respectively.

  The color component data is generated by the processing by the third processing unit 133 of the reduction processing unit 130 (FIG. 15: steps S625 and S640). As understood from the processing of steps S625 and S640, the color component data includes the pixel value of each pixel in the character object area and the pixel value of each pixel in the gray area among the plurality of pixel values in the target image data. Is the image data obtained by substituting the background color value BC. The color component data includes partial image data corresponding to the color area in the target image data. Accordingly, as shown in FIG. 16A, the color component image CI represented by the color component data includes the background Bg corresponding to the non-object region L30 (FIG. 3D) of the target image SI and the color region L33. A partial image representing the color object Ob3 corresponding to (FIG. 3D) is included. In addition, the color component data includes three component data (that is, R component data, G component data, and B component data) corresponding to the three color components, like the target image data.

  The compressed character image data is generated by the processing by the first processing unit 131 of the reduction processing unit 130 (ie, step S630 in FIG. 15). As understood from the processing of step S630, the compressed character image data is binary image data obtained by binarizing the partial image data corresponding to the character object area in the target image data, and compressed by the FLAT compression method. It is data. In the example of FIG. 16B, a binary image TI1 representing three character objects Ob4 to Ob6 corresponding to the character object region L34C (FIG. 3D) of the target image SI, and a region L37 (FIG. 3D )), Two compressed character image data respectively representing a binary image TI2 representing one character object Ob7 are generated.

  The gray component data is generated by the processing (FIG. 15: steps S635 and S645) by the second processing unit 132 of the reduction processing unit 130. As understood from the processing in steps S635 and S645, the gray component data includes partial component data each representing one or more gray regions included in the target image SI. Therefore, as shown in FIG. 16C, the gray component image GI represented by the gray component data includes two gray objects Ob1 and Ob2 corresponding to the two gray regions L31 and L32 in the target image SI, respectively. , Two gray partial images PG1 and PG2 are included. Further, unlike the target image data, the gray component data is one component data composed of one component value.

  The gray mask data is generated by the processing (FIG. 15: step S647) by the second processing unit 132 of the reduction processing unit 130. As understood from the processing in step S647, the gray mask image MI represented by the gray mask data corresponds to the two gray partial images PG1 and PG2 in the corresponding gray component image GI (FIG. 16C), respectively. Two mask partial images PM1 and PM2 are included. The two mask partial images PM1 and PM2 indicate the positions of the object pixels in the corresponding two gray partial images PG1 and PG2. That is, the gray mask data is binary data composed of a value “1” representing the position of the object pixel in the gray component image GI and a value “0” representing the position of the non-object pixel. In other words, the gray mask image MI includes pixels to be displayed (that is, pixels having a pixel value “1”) and pixels not to be displayed (when the gray component image GI is displayed over the color component image CI). That is, a binary image defining a pixel having a pixel value “0”).

  In step S655, first, the second processing unit 132 of the reduction processing unit 130 compresses the gray component data using the JPEG compression method. The JPEG compression method is an irreversible compression method used for compression of image data generated by taking a picture with a digital camera, and is an image having a relatively high gradation, such as a photograph, that is, the number of colors C Therefore, it is suitable for compression of images with a gradual change in gradation. On the other hand, the JPEG compression method greatly deteriorates an edge whose gradation changes rapidly, and thus is not suitable for compression of an image in which reproducibility of an edge is important from the viewpoint of readability and appearance like characters. The gray component data compressed in this step is also called compressed gray image data. The generated compressed gray image data is stored in the buffer area 241 and the gray component data before compression is deleted.

  When the compressed gray image data is generated, in the next step S660, the second processing unit 132 generates the compressed gray mask data by compressing the gray mask data by the FLATE compression method. As described above, the FLATE compression method is suitable for compression of an image having a relatively small number of gradations such as binary data. The generated compressed gray mask data is stored in the buffer area 241, and the gray mask data before compression is deleted.

  When the compressed gray mask data is generated, in the next step S665, the third processing unit 133 of the reduction processing unit 130 compresses the color component data by the JPEG compression method. As described above, since the color component data includes three component data, the three component data are respectively compressed. The color component data compressed in this step is also called compressed color image data. The generated compressed color image data is stored in the buffer area 241, and the color component data before compression is deleted. When the compressed color image data is generated, the reduction processing unit 130 ends the reduction processing.

  When the reduction process is completed, in step S700 of FIG. 2, the generation unit 150 causes the compressed character image data, the compressed gray image data, the compressed color image data, and the compressed gray mask data. And generate a compressed PDF file.

  Specifically, the generation unit 150 stores the compressed color image data in the PDF file as image data to be displayed as the lowest layer.

  Further, the generation unit 150 stores the compressed character image data in the PDF file as image data to be displayed as a layer higher than the compressed color image data. The compressed character image data is stored in the PDF file in association with the character color value TC and the coordinate value CD. The character color value TC is an RGB value representing the color of the character, and is the value calculated in step S620 in FIG. The coordinate value CD is information representing a position where the binary image TI represented by the compressed character image data should be arranged with respect to the color component image CI represented by the compressed color image data. The coordinate value CD is represented by, for example, the coordinate value (X, Y) of the pixel at the upper left corner of the smallest rectangle that circumscribes the binary image TI. In the example of FIG. 16B, character image data representing a binary image TI1 representing three character objects Ob4 to Ob6 includes a character color value TC1 (R1, G1, B1) and a coordinate value CD1 (X1, Y1) are associated with each other. In addition, a character color value TC2 (R2, G2, B2) and a coordinate value CD2 (X2, Y2) are associated with the binary image TI2 representing one character object Ob7.

  Further, the generation unit 150 stores the compressed gray image data in the PDF file as image data to be displayed as a layer higher than the compressed color image data. The compressed gray image data is stored in the PDF file in association with the compressed gray mask data.

When the compressed PDF file is generated, for example, the scanner driver 100 stores the generated PDF file in, for example, the nonvolatile storage device 290 and the compressed character image data stored in the buffer area 241 and the compressed PDF file. After erasing the already-processed gray image data, the compressed color image data, and the compressed gray mask data, the image processing is terminated.
A PDF file can store a plurality of different types of image data in a single file. When an image is displayed using the file, a plurality of stored image data is superimposed on a single file. Standards have been established so that images can be reproduced. In step S700, the generation unit 150 stores each compressed image data (FIG. 16) in a PDF file in accordance with the PDF standard. Therefore, the compressed PDF file created in this embodiment is a PDF file viewing software. When used and displayed, the target image SI (FIG. 3A) can be reproduced.

  According to the first embodiment described above, the first processing unit 131 generates compressed character image data using the partial image data corresponding to the character region in the target image data. In addition, the second processing unit 132 generates compressed gray image data composed of one type of component value using partial image data corresponding to the gray region in the target image data. The second processing unit 132 acquires gray component data composed of one type of component value (FIG. 15: steps S635 and S645), processing for compressing gray component data (FIG. 15: step S655), To generate compressed gray image data. The third processing unit 133 generates compressed color image data composed of a plurality of types of component values using partial image data corresponding to the color area in the target image data. The third processing unit 133 performs a process of acquiring color component data including a plurality of component data (FIG. 15: steps S625 and S640) and a process of compressing the color component data (FIG. 15: step S665). As a result, compressed color image data is generated. The processes performed by the first processing unit 131, the second processing unit 132, and the third processing unit 133 are different from each other. As a result, processing suitable for a character, a gray image different from the character, and a color image different from the character are executed, so that the target image data is effectively compressed to generate compressed target image data. Can do.

  Specifically, for a character region in which edge reproducibility is considered to be more important than gradation, binary image data is compressed by the FLATE compression method to generate compressed character image data. As a result, an image representing a character region can be stored in a manner that maintains a high compression ratio and does not impair the readability of characters.

  Further, among non-character areas that can include photographic areas where gradation is more important than edge reproducibility, for gray areas, gray component data composed of one component is converted into JPEG compression. Compressed gray image data is generated. As a result, it is possible to increase the compression rate by using gray component data composed of one component, and to include gray photographs by using the JPEG compression method suitable for multi-tone compression. An image representing the gray area can be stored in a manner that does not impair the image quality (gradation, etc.) of the obtained gray area.

  Further, among the non-character areas that can include a photographic area where gradation is considered to be more important than edge reproducibility, the color area is composed of a plurality of types of color components (for example, three types of RGB). The color component data is compressed by the JPEG compression method to generate compressed color image data. As a result, an image representing the color area can be stored in a manner that does not impair the image quality (hue and gradation) of the color area that may include a color photograph.

  As a result, the target image data representing the target image SI including the character area, the gray area, and the color area is converted and saved in a format that can improve the compression rate while maintaining high image quality as a whole. Can do.

  Further, the second processing unit 132 selects one component data out of three component data (that is, R component data, G component data, and B component data) included in the partial image data corresponding to the gray area. Thus, gray component data is generated (FIG. 15: step S635). As a result, gray component data can be easily obtained.

  Further, the compressed gray image data is image data representing the gray component image GI shown in FIG. 16C, and one piece representing an area corresponding to the entire target image SI as can be seen from FIG. Image data. Further, the compressed color image data is image data representing the color component image CI shown in FIG. 16A, and as shown in FIG. 16A, 1 represents a region corresponding to the entire target image SI. Pieces of image data. As a result, for example, even when a relatively large number of gray regions exist, it is possible to suppress an excessive increase in the number of compressed image data.

  Furthermore, when the ratio PN of pixels representing a chromatic color among the plurality of pixels included in the object region to be determined is equal to or greater than the gray determination reference value Nth, the determination unit 125 determines the object region to be determined as If it is determined that the area is a color area and the ratio PN of pixels representing a chromatic color is less than the gray determination reference value Nth, the determination target area is determined to be a gray area (FIG. 11). Therefore, the determination unit 125 can appropriately determine whether the region is a gray region or a color region.

  Further, the determination unit 125 identifies a character region and a non-character region including a gray region and a color region according to the number of colors C (in other words, the number of color types) included in the object region to be processed. (FIGS. 9 and 10). As a result, the specification unit 120 can specify the character region and the non-character region with high accuracy by using the difference in gradation of the region represented by the number of colors C.

  Further, in the above-described embodiment, the specifying unit 120 performs the labeling and the integration process (FIG. 5) to separate and specify the character area for each character color. Then, binary data compressed for each character area is created, and a character color value TC is associated with each binary data (FIG. 16). Therefore, it is possible to create a compressed PDF file that is compressed at a high compression rate while reproducing the character color.

B. Second Embodiment FIG. 17 is a flowchart of a reduction process according to the second embodiment. In FIG. 17, the same steps as those of the reduction process (FIG. 15) of the first embodiment are denoted by the same reference numerals as those of FIG. 15, and steps different from those of the reduction process of the first embodiment are denoted by “ A ”is attached.

  In the reduction process of the second embodiment, steps S645A and S647A shown in FIG. 17 are executed instead of steps S645 and S647 (FIG. 15) in the first embodiment, and steps S655 and S660 in the first embodiment are not executed. .

  In step S645A of FIG. 17, the second processing unit 132 acquires the number of gradations of one component data (for example, representing the image of the gray area to be processed that is acquired in step S635 and stored in the buffer area 241). Gray component data representing a gray region image is generated. In the present embodiment, the second processing unit 132 converts one component data representing a gray area image into binary data (reducing the number of gradations to two gradations).

  In the subsequent step S647A, the second processing unit 132 compresses the two-gradation gray component data generated in step S645A by using the FLATE compression method suitable for compressing an image having a relatively small number of gradations. Of gray component data is generated. In this step, compressed gray image data representing the gray region image to be processed is generated. The generated compressed gray image data is stored in the buffer area 241 and the binarized component data before compression is deleted.

  As can be understood from the above description, in this embodiment, compressed gray image data is generated for each gray region. Accordingly, when there are a plurality of gray areas, a plurality of compressed gray image data are generated and stored in the buffer area 241.

  In the reduction process of the second embodiment, compressed gray image data is generated for each gray area, and therefore the process of step S655 of the first embodiment (a process of compressing one gray component data) does not exist. . In the reduction process of the second embodiment, the compressed gray image data is binary data. Therefore, when superimposed on the color component image CI, the pixel to be displayed (that is, the pixel having the pixel value “1”). ) And non-displayed pixels (that is, pixels having a pixel value “0”) are clear. For this reason, gray mask data (FIG. 16D) is not generated in the reduction process of the second embodiment. Therefore, in the reduction process of the second embodiment, the process of step S660 of the first embodiment does not exist.

FIG. 18 is a diagram for explaining image data generated by the reduction process of the second embodiment.
In the reduction process of the second embodiment, when the target image data representing the target image SI of FIG. 3A is used, as shown in FIG. 18, one compressed color component data (FIG. 18A )), Two compressed character image data (FIG. 18B), and two compressed gray component data (FIG. 18C) are generated in the buffer area 241. Each is stored.

  One compressed color component data (FIG. 18A) and two compressed character image data (FIG. 18B) are data having the same name in the first embodiment (FIG. 16A). ) And FIG. 16 (B)).

  The compressed gray component data is data representing a binary image having a size corresponding to a minimum rectangle circumscribing the gray region, which is generated for each gray region, that is, a non-character object region determined to be gray. Accordingly, in the example of FIG. 18C, two compressed gray image data respectively representing two binary images GI1 and GI2 representing two gray objects Ob1 and Ob2 are generated. In step S700 of FIG. 2, the compressed gray image data representing the binary image GI1 is stored in the PDF file in association with the coordinate value CD3 (X3, Y3), and the compressed gray image representing the binary image GI2 is stored. The image data is stored in the PDF file in association with the coordinate value CD4 (X4, Y4).

  According to the second embodiment described above, as in the first embodiment, the target image data representing the target image SI including the character area, the gray area, and the color area is maintained with high image quality as a whole. However, it can be converted and saved in a format that can improve the compression rate.

  Furthermore, according to the second embodiment, the second processing unit 132 executes a process for reducing the number of gradations of the component values included in the gray component data (FIG. 17: step S645A). As a result, the amount of gray component data can be further reduced.

  Further, the FLATE compression method suitable for relatively few gradations is adopted as the compression method for compressing the gray component data in accordance with the reduction of the number of gradations of the gray component data to 2 gradations (FIG. 17). : Step S647A). As a result, the compression rate can be further improved according to the number of gray levels of the gray component data.

  The compressed color image data is one piece of image data representing an area (that is, the color component image CI (FIG. 18A)) corresponding to the entire target image SI. The compressed gray image data represents one or more gray component images (two gray component images GI1 and GI2 in the example of FIG. 18C) that represent an area corresponding to a part of the target image SI. One or more partial image data. One or more compressed gray image data is associated with position information (coordinate values CD1, CD2 in the example of FIG. 18) indicating the position in the target image SI. As a result, for example, when the gray area is relatively small, it is only necessary to hold gray image data of a partial area, which is efficient.

C. Variations:
(1) In the first and second embodiments, the FLATE compression method is used to generate compressed character image data. However, the present invention is not limited to this. For example, another compression method suitable for compression of binary image data may be employed. For example, an MMR (Modified Modified Read) method (also referred to as CCITT-G4 method) that is a lossless compression method may be employed. good.

(2) In the first embodiment, the JPEG compression method is employed for both the generation of compressed gray image data and the generation of compressed color image data. However, the present invention is not limited to this. For example, the GIF (Graphic Interchange Format) format or the LZW compression used for compressing the TIFF (Tagged Image File Format) format image file may be employed.

(3) In step S645A (FIG. 17) of the second embodiment, the second processing unit 132 determines the number of gradations of one component data (eg, 255 gradations (8 Bit)) is reduced to 2 gradations (1 bit), but is not limited thereto. Instead, the second processing unit 132 sets the number of gradations of one component data representing the gray area image to 4 gradations (2 bits), 8 gradations (3 bits), and 16 gradations (4 Bit), 32 gradations (5 bits), 64 gradations (6 bits), or 128 gradations (7 bits).

(4) The scanner driver 100 may execute image processing in which part of the image processing in the first embodiment and part of the image processing in the second embodiment are combined. Specifically, the scanner driver 100 applies the one component data representing the image in the gray area acquired in step S635 of the reduction process (FIG. 15) of the first embodiment to the second embodiment. You may perform the process which reduces the number of gradations of step S645A of a reduction process (FIG. 17). In this case, one piece of gray component data with a reduced number of gradations and one piece of gray component data representing an area corresponding to the entire target image SI is generated. Conversely, the scanner driver 100 may omit the process of reducing the number of gradations in step S645A in the reduction process (FIG. 7) of the second embodiment. In this case, the number of gradations is not reduced, for example, one or more pieces of gray component data composed of component values of 256 gradations, each representing an area corresponding to a part of the target image SI The above gray component data is generated.

(5) In the first embodiment, in step S635 (FIG. 15), the second processing unit 132 includes, among the target image data, three component data included in the partial image data corresponding to the gray area to be processed. One component data representing a gray region is acquired by selecting one component data from among (ie, R component data, G component data, B component data). Instead, the second processing unit 132 converts one component data representing an image in the gray area to be processed into three component data (that is, R data) included in the partial image data corresponding to the gray area. (Component data, G component data, B component data) may be used. Specifically, the second processing unit 132 may calculate the luminance value Y of each of the plurality of pixels in the gray area using the following formula.
Y = R × 0.29871 + G × 0.58661 + B × 0.11448
R, G, and B in the above formula are pixel values of corresponding pixels of the three component data, that is, the R component value, the G component value, and the B component value. The second processing unit 132 may use one component data having the calculated luminance value as one component value of each pixel as one component data representing an image in the gray area.

(6) FIG. 2: The calculation formula of the edge strength Se in step S150 and the calculation formula of the edge strength EP in step S5532 in FIG. 12 are not limited to the calculation method using the Sobel operator in FIG. Any method can be adopted. For example, various edge detection operators such as a Prewitt operator or a Roberts Cross operator can be used. The edge strength is not limited to the RGB color components, and may be calculated using gradation values of other color components (for example, luminance).

(7) In the above-described embodiment, the determination unit 125 determines whether the color of each pixel is a chromatic color using the Lab color space. That is, the determination unit 125 converts the pixel value of the pixel (in other words, the color value) into a color value of the Lab color space, and the distance R between the color value and the achromatic color axis of the Lab color space is Whether or not the color of the pixel is a chromatic color is determined based on whether or not the distance is less than the reference distance Rth. However, the determination unit 125 is not limited to this, and in general, the determination unit 125 is relatively in the achromatic color axis in an arbitrary color space having an achromatic color axis, for example, an RGB color space, a Lab color space, an HSV color space, and a YCrCb color space. A pixel representing a close color may be determined to be a pixel having an achromatic color, and a pixel representing a color relatively far from the achromatic color axis may be determined to be a pixel having a chromatic color.

(8) The image processing function by the scanner driver 100 of the computer 200 may be realized by an image processing apparatus including an image reading unit that generates image data representing an object by optically reading the object (for example, MFP 400, scanner 300, or digital camera (not shown). In this case, the image processing apparatus may perform image processing (for example, the processing in FIG. 2) using the image data generated by its own image reading unit.

  In general, an image processing apparatus that realizes image processing (for example, the process of FIG. 2) is not limited to the computer 200, and may be various apparatuses. For example, a computer inside an image-related device such as a printer, a digital camera, or a scanner, a general-purpose personal computer, a server connected to a network, or the like can be employed. A plurality of computers that can communicate with each other via a network may share a part of the functions required for image processing and provide the image processing functions as a whole. In this case, the entirety of the plurality of computers corresponds to the image processing apparatus in the claims.

(9) In each of the above embodiments, a part of the configuration realized by hardware may be replaced with software, and conversely, part or all of the configuration realized by software is replaced with hardware. You may do it.

  As mentioned above, although this invention was demonstrated based on the Example and the modification, Embodiment mentioned above is for making an understanding of this invention easy, and does not limit this invention. The present invention can be changed and improved without departing from the spirit and scope of the claims, and equivalents thereof are included in the present invention.

    DESCRIPTION OF SYMBOLS 100 ... Scanner driver, 110 ... Acquisition part, 120 ... Identification part, 125 ... Judgment part, 130 ... Reduction process part, 131 ... 1st process part, 132 ... 1st 2 processing units, 133 ... third processing unit, 150 ... generation unit, 200 ... computer, 210 ... CPU, 240 ... volatile storage device, 241 ... buffer area, 270. ..Operation unit, 280 ... communication unit, 290 ... nonvolatile storage device, 291 ... driver program, 292 ... judgment table, 300 ... scanner, 400 ... multifunction device

Claims (9)

Target image data representing a target image including a character, a gray image different from the character, and a color image different from the character, and acquiring target image data including a plurality of component data corresponding to a plurality of color components An acquisition unit to
In the target image, and a character region representing the character, and gray regions represent different said gray image and text, and a specifying unit configured to specify a color region representing a different said color image as a character,
A reduction processing unit for reducing the data size of the target image data,
A first processing unit that generates compressed character image data by executing a first process using partial image data corresponding to the character region;
A second process different from the first process is executed using the partial image data corresponding to the gray area, thereby generating compressed gray image data composed of one type of component value. A second processing unit, wherein the second processing is performed by performing processing for obtaining gray component data composed of one kind of component value and processing for compressing the gray component data; look including generating a gray image data, gray image data of said compressed is one of image data representing a region corresponding to the whole of the target image, and the second processing unit,
By executing a third process different from the first process and the second process using partial image data corresponding to the color area, a compressed color composed of a plurality of types of component values A third processing unit for generating image data, wherein the third process performs a process of acquiring color component data including a plurality of component data and a process of compressing the color component data. look including generating a compressed color image data, color image data of said compressed is one of image data representing a region corresponding to the whole of the target image, and the third processing unit,
The reduction processing unit,
A generating unit that generates compressed target image data representing the target image using the compressed character image data, the compressed gray image data, and the compressed color image data;
An image processing apparatus comprising:   The image processing apparatus according to claim 1, further comprising:
  An area corresponding to the entire target image represented by the compressed gray image data and an area corresponding to the entire target image represented by the compressed color image data are displayed in an overlapping manner. A fourth processing unit for generating mask data defining pixels to be displayed and pixels not to be displayed for one region,
  The generation unit generates the compressed target image data using the compressed character image data, the compressed gray image data, the compressed color image data, and the mask data. An image processing apparatus. The image processing apparatus according to claim 1 or 2 ,
The image processing apparatus, wherein the gray component data includes one component data selected from the plurality of component data included in the partial image data corresponding to the gray region. It claims 1 An image processing apparatus according to claim 3,
The compression method used for compression of the gray component data by the second processing unit and the compression method used for compression of each of the plurality of component data included in the color component data by the third processing unit are the same An image processing apparatus. An image processing apparatus according to any one of claims 1 to 4 ,
It said second process includes a process for reducing the gradation number of component value included in the gray component data, the image processing apparatus. An image processing apparatus according to any one of claims 1 to 5 ,
The specifying unit includes a determination unit that determines whether an area in the target image is the gray area or the color area,
The determination unit determines that the determination target area is the color area when the ratio of pixels representing a chromatic color included in the determination target area is equal to or greater than a reference value.
The image processing apparatus, wherein the determination unit determines that the determination target area is the gray area when a ratio of pixels representing a chromatic color included in the determination target area is less than a reference value. The image processing apparatus according to claim 6 ,
The determination unit is the gray region by using a plurality of pixels excluding a plurality of outer edge pixels located at an outer edge portion of the determination target region from a plurality of pixels included in the determination target region. Or an image processing apparatus for determining whether the color area is present. An image processing apparatus according to any one of claims 1 to 7 ,
The specifying unit specifies the character region and the region including the gray region and the color region according to the number of types of colors included in the processing target region. Target image data representing a target image including a character, a gray image different from the character, and a color image different from the character, and acquiring target image data including a plurality of component data corresponding to a plurality of color components With the acquisition function to
In the target image, and a character region representing the character, and gray regions represent different said gray image and text, and the specific function of specifying a color region representing a different said color image as a character,
A reduction processing function for reducing the data size of the target image data,
A first processing function for generating compressed character image data by executing a first process using partial image data corresponding to the character region;
A second process different from the first process is executed using the partial image data corresponding to the gray area, thereby generating compressed gray image data composed of one type of component value. A second processing function, wherein the second processing is performed by performing processing for obtaining gray component data composed of one type of component value and processing for compressing the gray component data. look including generating a gray image data, gray image data of the compressed, the a single image data representing a region corresponding to the entire target image, and the second processing function,
By executing a third process different from the first process and the second process using partial image data corresponding to the color area, a compressed color composed of a plurality of types of component values A third processing function for generating image data, wherein the third processing performs processing for acquiring color component data including a plurality of component data and processing for compressing the color component data; look including generating a compressed color image data, color image data of said compressed is one of image data representing a region corresponding to the whole of the target image, and the third processing function,
The reduction processing function,
A generation function for generating compressed target image data representing the target image using the compressed character image data, the compressed gray image data, and the compressed color image data;
A computer program that causes a computer to realize

Images