JP2008092447A - Image processing apparatus, image output device, and image processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing apparatus, image output device, and image processing method, which appropriately corrects a segmentation defect included in a result of segmentation into a character/drawing region and a non-character/drawing region. <P>SOLUTION: An image processing apparatus 11 comprises: an image region segmenting unit 18 for segmenting image data of an image read by an image reader 10 into a character/drawing region for characters or drawings and a non-character/drawing region that is not the character/drawing region; a color flatness calculating unit 19 for deriving, for each pixel, a flatness indicating whether a change of image data is in a flat region not more than a predetermined value, based on the image data of the read image in units of pixels; and an image region correcting unit 20 for correcting a segmentation defect included in the segmentation result on the basis of the calculated flatness. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an image processing device, an image output device, and an image processing method.

  In image output devices such as copiers, fax machines, and printers, the read image is separated into a character area, a photographic area, and a halftone dot area (hereinafter referred to as image area separation). It is widely performed to perform processing and output a high-quality image. In this way, when image processing is performed for each area, an image quality defect occurs when an image area separation error occurs.

Therefore, an image processing apparatus including an image area separation unit, a variable determination unit, and a gradation reproduction control unit, wherein the image area separation unit converts a character area, a photograph area, and a halftone dot of an input image from input image data. Based on the determination result, the variable determination unit determines an image area separation variable that indicates the characteristics of the input image stepwise or continuously, and reproduces the gradation based on the image area separation variable. An image processing apparatus has been proposed in which the control unit changes gradation reproduction characteristics for input image data stepwise or continuously so that no image quality defect occurs even if an image area separation error occurs ( For example, see Patent Document 1.)
JP-A-8-65514

  However, in the conventional image processing apparatus, only the gradation reproduction characteristics are changed stepwise or continuously by the image area separation variables (character degree, halftone dot degree, and photo degree). Although it is possible to blur the separation boundary between the photograph and the halftone dot area, it is an area that should be originally determined as a photograph area as shown in FIG. 29 (B) and should be processed erroneously as shown in FIG. 29 (A). When processed, there may be a gap in image quality between the separation error area and the area where the image area is correctly separated in the vicinity of the separation error area, and unnaturalness may remain.

  The present invention has been made in order to solve the above-described problems, and is an image processing apparatus capable of appropriately correcting a separation failure included in a separation result separated into a character line drawing area and a non-character line drawing area, An object of the present invention is to provide an image output apparatus and an image processing method.

  In order to achieve the above object, an image processing apparatus according to a first aspect of the present invention provides an image data read by an image reading means, wherein a character or line drawing character line drawing area and a non-character line drawing area that is not a character line drawing area. Based on the image data of the read image, the flatness indicating whether each pixel is a flat region whose change in image data is a predetermined value or less is derived for each pixel based on the image data of the read image. Flatness deriving means for correcting, and correcting means for correcting a separation failure included in the separation result of the separating means based on the derived flatness.

  According to a second aspect of the invention, in the first aspect of the invention, the flatness deriving means calculates the flatness of each pixel from the average coordinates of the image data of each pixel and each pixel neighboring area in the uniform color space. It is derived based on distribution.

  According to a third aspect of the present invention, in the first or second aspect of the invention, the flatness deriving unit sequentially sets each pixel of the read image as a target pixel, and sets the target pixel and the target attention. Calculating the degree of variation of image data in the pixel vicinity region, and the deviation between the image data of the target pixel and the average value of the target pixel and the image data of the target pixel vicinity region; and The flatness of each pixel is derived by deriving the flatness of the target pixel based on the deviation.

  According to a fourth aspect of the invention, in the third aspect of the invention, the flatness derivation means calculates a standard deviation of data indicating hue and saturation included in the image data as the degree of variation, and the deviation The difference between the average of the hue data included in the image data of the pixel of interest and the region near the pixel of interest and the hue data included in the image data of the pixel of interest is calculated.

  According to a fifth aspect of the present invention, in the invention according to any one of the first to fourth aspects, the image data of the read image is subjected to image processing according to the separation result corrected by the correction means. Image processing means is further provided.

  According to a sixth aspect of the invention, in the fifth aspect of the invention, the image processing means is derived by the separation result corrected by the correction means and the flatness deriving means into the image data of the read image. Image processing is performed according to flatness.

  According to a seventh aspect of the present invention, in the invention according to any one of the first to sixth aspects, the separation unit sequentially sets each pixel of the read image as a target pixel, and the read image The image data of the rectangular area including the set target pixel is extracted from the image data of the image, and the image data of the read image is converted into the character line drawing area and the non-character line drawing area by performing predetermined processing on the extracted image data. Further, separation result arrangement data in which the separation result for each pixel is arranged corresponding to the position of each pixel of the read image is generated as the separation result, and the correction means is derived Using the flatness and a predetermined rectangular template, a separation defect having a shape of a rectangle or a series of rectangles is detected from the generated separation result arrangement data, and the detected separation defect is corrected. And wherein the door.

  The invention according to claim 8 is the invention according to claim 7, wherein the separation unit sequentially sets each pixel of the read image as a target pixel, and sets the target pixel set from the image data of the read image. After extracting the image data of the rectangular area including the image, and performing the contraction process for contracting the black part on the extracted image data, the expansion process for expanding the black part is performed, and the contraction process and the expansion process are performed for all the target pixels. Separating the character / line drawing area from the image data of the read image based on the image data obtained by inverting the image data after completion and the image data before the shrinking process, and image data other than the separated character / line drawing area Is generated as a non-character / line drawing area to generate the separation result arrangement data, and the correction means includes the derived flatness, a predetermined rectangular model, and the convergence. Using the processed image data, the above-generated separation results from the arrangement data of the continuous shape is rectangular or square separating detect bad, and correcting the defective separation of the detected.

  The invention according to claim 9 is the invention according to any one of claims 1 to 8, wherein the separating unit further separates the non-character / line drawing area into a halftone dot area and a photograph area. To do.

  An image output device according to a tenth aspect of the present invention provides an image reading unit for reading an image of a document, and converts the image data of the read image into a character / line drawing area and a non-character / line drawing area that is not a character / line drawing area. Flatness for deriving, for each pixel, flatness indicating whether or not each pixel is a flat region whose change in image data is a predetermined value or less based on image data of the read image and separation means for separating A deriving unit, a correcting unit that corrects a separation failure included in the separation result of the separating unit based on the derived flatness, and a separation result corrected by the correcting unit in the image data of the read image And image output means for outputting an image based on the image data subjected to the image processing by the image processing means.

  An image processing method according to an eleventh aspect of the invention includes a separation step of separating image data of an image read by an image reading unit into a character / line drawing character / line drawing region and a non-character / line drawing region that is not a character / line drawing region; A flatness deriving step for deriving flatness for each pixel, which indicates whether each pixel is a flat region whose change in image data is a predetermined value or less based on the image data of the read image; And a correction step of correcting a separation failure included in the separation result of the separation step based on the flatness.

  As described above, according to the inventions according to claims 1, 10, and 11, the separation failure included in the separation result separated into the character line drawing area and the non-character line drawing area is appropriately corrected. The effect of being able to be obtained.

  According to the second aspect of the invention, the flatness can be easily derived.

  According to the third and fourth aspects of the invention, the flatness can be derived appropriately.

  According to the fifth aspect of the present invention, appropriate image processing can be performed according to the appropriately corrected separation result.

  According to the sixth aspect of the present invention, appropriate image processing can be performed in accordance with the appropriately corrected separation result and flatness.

  According to the seventh aspect of the present invention, a rectangular shape or a continuous shape of rectangles generated when image data of a rectangular area is extracted and subjected to predetermined processing to be separated into a character line drawing area and a non-character line drawing area. Can be corrected.

  According to the eighth aspect of the present invention, the rectangles or rectangles generated when the image data of the rectangular area is extracted and subjected to the shrinking process or the expanding process to be separated into the character line drawing area and the non-character line drawing area are connected. It is possible to correct the separation failure of the shape with higher accuracy.

  According to the ninth aspect of the present invention, even when separating into a character / line drawing area, a halftone dot area, and a photographic area, it is possible to correct the separation defect similarly.

  Hereinafter, an example of an embodiment of the present invention will be described in detail with reference to the drawings. In the present embodiment, the present invention is applied to a copying machine that reads and copies an image of a document.

  FIG. 1 is a block diagram showing the configuration of the copying machine according to the present embodiment.

  The copying machine includes an image reading device 10, an image processing device 11, and an image output device 12.

  The image reading device 10 reads an image of a document and outputs image data in the RGB color space of the read document image to the image processing device 11.

  The image processing apparatus 11 includes a color conversion unit 13, a scaling unit 14, an image area separation processing unit 15, a scaling unit 16, and an image processing unit 17.

  The color conversion unit 13 uses the L * a * b * color space, which is one of the uniform color spaces, for the image data in the RGB color space obtained by reading with the image reading device 10 using a predetermined look-up table (LUT). Convert to image data. Here, the uniform color space refers to a color space in which a perceptible color difference and a spatial distance difference substantially coincide.

  The reduction / enlargement unit 14 performs enlargement / reduction processing on the image data converted by the color conversion unit 13.

  As shown in FIG. 1, the image area separation processing unit 15 includes an image area separation unit 18, a color flatness calculation unit 19, and an image area correction unit 20.

  Only the L * value of the image data in the L * a * b * color space converted by the color conversion unit 13 is input to the image area separation unit 18. The image area separation unit 18 separates the image indicated by the image data of the L * value into a character area and a photographic area according to a procedure as disclosed in Japanese Patent Laid-Open No. Hei 2-294485, and corresponds to the position of each pixel. Thus, data in which the image area separation result is mapped (referred to as image area separation tag image data) is generated.

  In the present embodiment, an image area in which image data has a peak near white and black and has a sharp change in image data, such as characters and line drawings, is referred to as a character area, and image data such as a photograph. An area of a photographic image that is an image with a relatively slow change is referred to as a photographic area, and an area of a halftone image that is a pseudo gradation image by halftone dot formation that is seen in a printed material is referred to as a halftone area. However, the image processing unit 17 according to the present embodiment distinguishes image processing between character regions and other regions and does not perform image processing that distinguishes between halftone dot regions and photographic regions. In 18, it is assumed that image area separation tag image data in which an image area separation result in which all areas other than the character area are photograph areas is mapped is generated.

  The color flatness calculation unit 19 calculates the color flatness for each pixel from the image data in the L * a * b * color space converted by the color conversion unit 13, and the color flatness corresponding to the position of each pixel. Is mapped data (referred to as color flatness tag image data).

  Here, the color flatness is a value indicating whether or not each pixel is a flat region where the change in the image data is a predetermined value or less. Here, the color is high in color or low in color. Including white / black distinction, four values of 0 (white flat / non-flat), 1 (black flat), 2 (intermediate flat), and 3 (color flat) are taken.

  The image area correction unit 20 uses the color flatness indicated by the color flatness tag image data generated by the color flatness calculation unit 19 to display the image area separation result indicated by the image area separation tag image data generated by the image area separation unit 18. Correct using degrees.

  The enlargement / reduction unit 16 converts the image area separation tag image data and the color flatness tag image data generated and corrected by the image area separation processing unit 15 to the same magnification as the enlargement / reduction process performed by the reduction / enlargement unit 14. Enlarge / reduce with.

  The image processing unit 17 displays the image area separation result indicated by the image area separation tag image data output from the enlargement / reduction unit 16 and the color flatness tag image data for the image data output from the enlargement / reduction unit 14. Image processing is performed based on the color flatness and output data is output.

  The image output device 12 forms an image on recording paper based on the output data output from the image processing unit 17.

  FIG. 2 is a block diagram illustrating a configuration of the image processing unit 17.

  The image processing unit 17 includes a space correction unit 21, a color conversion unit 22, a gradation correction unit 23, and a screen generation unit 24.

  The space correction unit 21 corrects the spatial frequency of the image data in the L * a * b * color space output from the reduction / enlargement unit 14 using a filter corresponding to the image area separation result and the color flatness.

  The color conversion unit 22 prints the image data in the L * a * b * color space corrected by the space correction unit 21 using a look-up table (LUT) corresponding to the image area separation result and color flatness. Convert to image data in YMCK color space.

  The tone correction unit 23 corrects the tone characteristics of the image data in the YMCK color space converted by the color conversion unit 22 using a look-up table (LUT) corresponding to the image area separation result and color flatness.

  The screen generation unit 24 determines in advance the image data in the YMCK color space corrected by the gradation correction unit 23 according to the image area separation result and the color flatness so that the image output device 12 can form an image. Halftone processing (processing for converting multi-gradation data into low-gradation data) is performed using the halftone screen pattern thus generated, and it is generated as output data and output to the image output device 12.

  FIG. 3 is a block diagram showing the configuration of the image area separation unit 18 described above.

  The image area separation unit 18 includes a binarization processing unit 25, a halftone dot extraction unit 26, an OR processing unit 27, a contraction processing unit 28, an expansion processing unit 29, an inversion processing unit 30, an AND processing unit 31, and an image region separation tag. The generation unit 32 is included.

  The binarization processing unit 25 performs generally known binarization processing on the L * value image data output from the color conversion unit 13 and converts the image data into binary image data. Then, the converted binary image data is output as image data of a binarized area.

  The halftone dot extraction unit 26 performs a general pattern matching process on the L * value image data output from the color conversion unit 13 to extract a halftone dot region, and the binary image of the extracted halftone dot region. Output data. For example, it is possible to use an algorithm that extracts a halftone dot area if the ratio of matching a template pattern of a predetermined size is equal to or higher than a predetermined level.

  The OR processing unit 27 outputs image data obtained by ORing the image data output from the binarization processing unit 25 and the image data output from the halftone extraction unit 26.

  The contraction processing unit 28 performs contraction processing on the image data output from the OR processing unit 27. In the shrinking process, a window image is extracted while moving a reference window of a predetermined size at a pitch of one pixel, and the black portion is thinned by performing black-and-white conversion of the pixel of interest located at the center of the window image. The image data is converted so as to be contracted.

  For example, a reference window of 5 × 5 pixels as shown in FIG. 4B is used for an input image (a binary image of white “0” and black “1”) as shown in FIG. The case where the contraction process is performed will be described as an example.

  In the contraction process, as shown in FIG. 4C, the window image is extracted so that the center of the reference window becomes the target pixel. If all the pixels in the extracted window image are black, the target pixel is converted to black “1”, and if there is at least one white pixel in the extracted window image, the target pixel is set to white “0”. Convert to Since the window image extracted at the position in FIG. 4C includes one white pixel, the target pixel is converted to a white pixel as shown in FIG. After the monochrome conversion for the target pixel is completed, the next pixel is set as the target pixel, the position of the reference window is moved, and the same processing is performed. Note that the window image extracted by the reference window is extracted from an image before black-and-white conversion (after the contraction process) by the contraction process.

  When the reference window is sequentially moved in this way to perform monochrome conversion of the pixel of interest, a processing result as shown in FIG. 4E is obtained. As shown in the figure, the processing result is an image in which the black portion is contracted. Further, since the reference window is rectangular, a block-shaped region is formed in the processing result of the contraction process.

  The processing result of the contraction process is not only output to the expansion processing unit 29 but also stored in a buffer (not shown) and used for image area correction of the image area correction unit 20 described later.

  The expansion processing unit 29 performs expansion processing on the image data output from the contraction processing unit 28. In the expansion process, a window image is extracted while moving a reference window of a predetermined size at a pitch of one pixel, and the black portion is thickened by performing black-and-white conversion of the pixel of interest located at the center of the window image ( The image data is converted so as to be expanded.

  For example, an input image (a binary image of white “0” and black “1”) as shown in FIG. 5A is used as a reference window of 5 × 5 pixels as shown in FIG. The case where the expansion process is performed will be described as an example.

  In the expansion process, as shown in FIG. 5C, the window image is extracted so that the center of the reference window becomes the target pixel. If at least one black pixel is included in the extracted window image, the target pixel is converted to black “1”. If all of the extracted window images are white pixels, the target pixel is Convert to white “0”. Since the window image extracted at the position in FIG. 5C includes one or more black pixels, the target pixel is converted into a black pixel as shown in FIG. After the monochrome conversion for the target pixel is completed, the next pixel is set as the target pixel, the position of the reference window is moved, and the same processing is performed. Note that the window image extracted by the reference window is extracted from the input image before black-and-white conversion by the expansion process.

  When the reference window is sequentially moved in this way to perform monochrome conversion of the pixel of interest, a processing result as shown in FIG. 5E is obtained. As shown in the figure, the processing result is an image in which the black portion is expanded. Since the reference window is rectangular, a block-shaped region is formed in the processing result of the expansion process.

  The inversion processing unit 30 performs monochrome inversion processing on the image data output from the expansion processing unit 29.

  The AND processing unit 31 outputs image data obtained by ANDing the image data subjected to the black and white inversion processing by the inversion processing unit 30 and the image data output from the OR processing unit 27.

  FIG. 6 is a diagram illustrating an example of image data obtained by ANDing the image data input to the AND processing unit 31 and the input image data.

  The image data of the image before the contraction process as shown in FIG. 6A and the image data of the image obtained by performing the black and white inversion process on the image after the expansion process as shown in FIG. Then, image data as shown in FIG. 6D is obtained. In the contraction process and the expansion process, the process is performed using the reference window. Therefore, the processing result of the AND processing unit 31 is also an image in which a block-shaped region is formed.

  The image area separation tag generation unit 32 determines whether each pixel is a character area corresponding to the position of each pixel of the image indicated by the image data converted by the color conversion unit 13 from the image data output from the AND processing unit 31. Image area separation tag image data in which an image area separation result indicating a photographic area is mapped is generated. That is, the tag value “1” of the character area is mapped to the pixel position that has become black “1” in the AND processing unit 31, and the tag value “0” of the photographic area is mapped to the other pixel positions.

  FIG. 7 is a diagram showing an example of the processing result in each part of the image area separation unit 18. Here, the processing flow of the image area separation unit 18 will be described with reference to FIG.

  The image shown in FIG. 7A is an image obtained by binarizing the image data of the input image by the binarization processing unit 25. Here, a part of a continuous image such as an image of a halftone dot region other than characters or a natural image such as a photographic paper photograph is extracted as a binarized region.

  The image shown in FIG. 7B is an image represented by the image data of the halftone dot region extracted by performing a general pattern matching process from the image data of the input image by the halftone dot extracting unit 26.

  The image shown in FIG. 7 (3) is an image represented by the image data output from the OR processing unit 27, and the logical sum of the image data of FIG. 7 (1) and the image data of the image of FIG. 7 (2). Is taken.

  The image shown in FIG. 7 (4) is obtained by contracting the image data of the image in FIG. 7 (3) by the contraction processing unit 28. Thereby, an image in which the black portion is thinned (shrinked) is generated. Since the character / line drawing is represented by a thin area (small area), the character area of the small area is erased by the shrinking process, and a relatively large halftone dot area remains.

  The image shown in FIG. 7 (5) is obtained by expanding the image data of the image of FIG. 7 (4) by the expansion processing unit 29. As a result, an image in which the black portion is fattened (expanded) is generated. As a result, the halftone dot region remaining after the contraction process is expanded.

  The image shown in FIG. 7 (6) is an image obtained by reversing the image data of the image in FIG. 7 (5) by the reversal processing unit 30.

  The image shown in FIG. 7 (7) is an image of the image data output from the AND processing unit 31, and is obtained by ANDing the image data of the images of FIG. 7 (6) and FIG. 7 (3). It is. As is apparent from this image, the final image is an image in which only the character area is extracted as black pixels. As a result, the image area can be separated from the character area and the other photographic areas.

  The document read by the image reading device 10 is a document output by the image output device 12 or another printer, that is, a low line number halftone dot, a YMCK line screen generally used in a multi-value laser printer, A document on which an image is formed on a rotation screen. When such a document is copied, image area separation performed at the time of copying may cause an image area separation error of a rectangle or a shape in which rectangles are connected, that is, a block shape, and an image quality defect may occur.

  FIG. 8 is an explanatory diagram for explaining a block-shaped image area separation error that occurs when an original image printed based on image data screen-processed by the dot TO line screen method is copied. The dot TO line screen method is one of methods for reproducing gradation by reproducing a highlight portion with dots and growing a dot component to form a line screen.

  The original image to be copied shown in FIG. 8A is an image whose entire area should be separated as a photographic area. When the original image is read, as shown in FIG. 8B, the area where the adjacent dots are independent from each other as a result of the reading is determined not to be a character area (a photographic area). However, as shown in FIG. 8C, the area of the line screen in which the adjacent dots of the reading result are continuous is determined to be a character area rather than a halftone area. As the reading resolution of the image reading apparatus 10 becomes higher, an enlarged image is input to the image area separation unit, and thus such misjudgment increases.

  When a halftone dot region extraction leak (an image region separation error region that is erroneously determined as a character region) occurs, the region is subjected to the above-described shrinkage processing and expansion processing as shown in FIG. 8D. It becomes a block shape.

  In such a state where an image area separation error has occurred, image processing with emphasis on fineness is performed on an area determined as a character area, and image on which gradation reproduction is emphasized on an area determined as a photographic area. When processing is performed and an image is output, an image quality defect occurs in a block shape as shown in FIG.

  FIG. 9 is an explanatory diagram for explaining block-shaped white spots that occur when copying an original image represented by low-line-number halftone dots.

  The document image to be copied shown in FIG. 9A is a halftone dot image with a low number of lines. Therefore, there is a case where the halftone dot interval is wide and cannot be detected as a halftone dot region by the halftone detection processing. For example, when a halftone dot area is detected by an algorithm in which a newspaper level (about 70-line halftone dot) is detected as a halftone dot area and a 25-line character area is detected with respect to the detection window, a halftone line of about 30-60 lines is detected. The point area becomes an image area separation boundary area, and an image area separation error occurs in this boundary area.

  Accordingly, the halftone dot extraction unit 26 determines that each dot point is a character as shown in FIG. 9B for a halftone dot region having a wide dot interval and a relatively low line number. In the case where the halftone dot extraction process has been missed in this manner, the subsequent OR processing, contraction processing, and expansion processing are performed to separate the image area, and image processing and image output are performed. As shown in (C), block-shaped white spots may occur.

  Therefore, in this image processing apparatus 11, the image area correction unit 20 corrects the image area separation error generated by the image area separation unit 18 using the color flatness calculated by the color flatness calculation unit 19.

  FIG. 10 is a flowchart showing the flow of the main routine of the color flatness calculation process performed by the color flatness calculation unit 19. FIG. 11 is a flowchart showing the flow of a subroutine called in the main routine of the color flatness calculation process.

  In step 100 of the main routine shown in FIG. 10, as shown in FIG. 12, the image represented by the image data in the L * a * b * color space converted by the color conversion unit 13 is a plurality of small regions of M × N pixels. The image data is divided so that it is divided into (M and N are natural numbers, but a sufficiently wide area for noise components).

  In step 102, for each small region, the color centroid (aG, bG) on the a * b * coordinate, and the standard deviation of a * and the standard deviation of b * (here, the standard deviation of a * and the standard deviation of b *) Are collectively referred to as saturation standard deviation).

  The color centroid (aG, bG) of each small region is calculated using the following equations (1) and (2).

The color centroid (aG, bG) is obtained by calculating the average value of the a * value and b * value of each small region in this way.

  In addition, the standard deviation σa * of a * and the standard deviation σb * of b * of each small region are calculated using the following equations (3) and (4).

The standard deviation σa * of a * and the standard deviation σb * of b * indicate the degree of variation in each small area.

  The calculated color centroid (aG, bG) and saturation standard deviation (standard deviation of a * and standard deviation of b *) are stored for each small area in a memory (not shown). Keep it.

  In step 104, the target pixel is set from the pixels constituting the image of the input image data.

  In step 106, the coordinate value (l, c) of the target pixel on the L * C * plane is calculated. Here, the coordinate value c of the C * axis indicating the saturation on the L * C * plane indicates the distance from the center on the a * b * plane, and is obtained by the following equation (5).

Thus, the C * coordinate value c of each pixel is obtained from the a * value and b * value of each pixel. Note that the L * coordinate value l of each pixel is the L * value of the input image data.

  In step 108, a subroutine for color flatness determination processing (FIG. 11) is executed.

  In step 200 of FIG. 11, it is determined whether or not the coordinate position (l, c) on the L * C * plane is located in the color region (region with high saturation). Note that such color determination processing based on the coordinate position (l, c) on the L * C * plane is performed using threshold values (THL_AVE1, THL_AVE2, THC_AVE1, THC_AVE2) shown in FIG. In step 200, if the C * coordinate value c is equal to or greater than the threshold value THC_AVE2, it is determined that the L * coordinate value l is located in the color region regardless of the value.

  If it is determined in step 200 that the coordinate position (l, c) on the L * C * plane is not located in the color area, the process proceeds to step 208, where the coordinate position ( It is determined whether l, c) is located in the intermediate area. Here, if the C * coordinate value c is greater than or equal to the threshold value THC_AVE1 and less than the threshold value THC_AVE2, it is determined that the L * coordinate value l is located in the intermediate region regardless of the value. Even if the C * coordinate value c is less than the threshold value THC_AVE1, if the L * coordinate value 1 is not less than the threshold value THL_AVE1 and less than the threshold value THL_AVE2, it is determined that the object is located in the intermediate region.

  If it is determined in step 208 that the coordinate position (l, c) on the L * C * plane is not located in the intermediate region, the process proceeds to step 216, where the coordinate position ( It is determined whether or not l, c) is located in a black region (a region with low saturation and close to black or a black region). Here, if the C * coordinate value c is less than the threshold THC_AVE1 and the L * coordinate value l is less than THL_AVE1, it is determined that the pixel is located in the black region.

  If it is determined in step 216 that the coordinate position (l, c) on the L * C * plane is not located in the black area, the coordinate position (l, c) is a white area (an area with low saturation). Then, it is determined that the image is located in a region close to white or white), and the process proceeds to step 224.

  In step 224, “0” indicating a white flat / non-flat region is set as the color flatness of the target pixel. Note that the white flat region is a region close to white or white and has a small change in hue and saturation, and the non-flat region indicates a region in which a change in hue and saturation is large.

  On the other hand, if it is determined in step 200 that the coordinate position (l, c) on the L * C * plane is located in the color area, the process proceeds to step 202, where M × N including the target pixel. It is determined whether or not the saturation standard deviation (standard deviation σa * of a * and standard deviation σb * of b *) of the small area is equal to or smaller than a threshold THC_RI. If it is determined that the saturation standard deviation is equal to or smaller than the threshold value THC_RI, it is determined in step 204 whether or not the flat hue angle is equal to or smaller than the threshold value THUEI. Here, the flat hue angle is an angular deviation of the coordinate value (a *, b *) of the pixel of interest with respect to the color centroid (aG, bG) on the a * b * plane, as shown in FIG.

  If it is determined in step 204 that the flat hue angle is equal to or smaller than the threshold THUEI, in step 206, “3” indicating a color flat region is set as the color flatness of the pixel of interest. The color flat region is a region having high saturation and small changes in hue and saturation.

  If it is determined in step 208 that the coordinate position (l, c) on the L * C * plane is located in the intermediate region, the process proceeds to step 210 and M × N including the pixel of interest is included. It is determined whether or not the saturation standard deviation (standard deviation σa * of a * and standard deviation σb * of b *) of the small area is equal to or smaller than a threshold value THC_RM. If it is determined that the saturation standard deviation is equal to or smaller than the threshold value THC_RM, it is determined in step 212 whether or not the flat hue angle is equal to or smaller than the threshold value THUEM.

  If it is determined in step 212 that the flat hue angle is equal to or smaller than the threshold THUEM, “2” indicating an intermediate flat region is set as the color flatness of the pixel of interest in step 214. The intermediate flat area is an intermediate area between the white and black areas and the color area and has a small change in hue and saturation.

  If it is determined in step 216 that the coordinate position (l, c) on the L * C * plane is located in the black region, the process proceeds to step 218, and M × N including the target pixel. It is determined whether or not the saturation standard deviation (standard deviation σa * of a * and standard deviation σb * of b *) of the small area is equal to or smaller than a threshold value THC_RB. If it is determined that the saturation standard deviation is equal to or smaller than the threshold value THC_RB, it is determined in step 220 whether or not the flat hue angle is equal to or smaller than the threshold value THUEB.

  If it is determined in step 220 that the flat hue angle is equal to or smaller than the threshold THUEB, “1” indicating a black flat region is set as the color flatness of the pixel of interest in step 222. The black flat region is a region close to black or black and having a small change in hue and saturation.

  If a negative determination is made in steps 202, 204, 210, 212, 218, and 220, the process proceeds to step 224, and “0” indicating a white flat / non-flat area is set as the color flatness of the target pixel. To do.

  After obtaining the color flatness for the pixel of interest, in step 110 of FIG. 10, it is determined whether or not the color flatness determination processing has been completed for all the pixels represented by the input image data. If it is determined in step 110 that the color flatness determination process has not been completed for all pixels, the process proceeds to step 112, the next pixel is set as the target pixel, the process returns to step 106, and the above process is repeated.

  By performing the color flatness determination process for each pixel in this way, the color flatness of each pixel is set to one of determination results 0 to 3 as shown in FIG. Note that the portions determined as the color flat region, the intermediate flat region, and the black flat region are flat portions with small changes in saturation and hue, and are highly likely to be non-character regions. In addition, a portion determined to be a white flat / non-flat region is a white region that is not mistakenly recognized as a character or line drawing, or has a large change in saturation or hue, resulting in a character region. Is likely.

  If it is determined in step 110 that the color flatness determination process has been completed for all pixels, the process proceeds to step 114.

  In step 114, based on the color flatness obtained for each pixel, color flatness tag image data in which the obtained color flatness is mapped corresponding to the position of each pixel of the image indicated by the image data is generated. Then, the main routine ends.

  Note that THC_RI, THC_RM, and THC_RB used as threshold values for the saturation standard deviation are preferably different values, but may be the same value. Similarly, THUEI, THUEM, and THUEB used as the threshold values of the flat hue angle may be different values or the same value.

  Further, here, an example of obtaining the standard deviation of a * and b * for obtaining the flatness has been described. However, the present invention is not limited to the standard deviation, and is not particularly limited as long as it is a value indicating the degree of variation such as variance. It is not limited.

  As described above, the color flatness of each pixel indicated by the color flatness tag image data obtained by the color flatness calculation unit 19 is output to the image area correction unit 20.

  FIG. 16 is a block diagram illustrating a configuration of the image area correction unit 20.

  The image area correction unit 20 includes an image area separation error block pattern detection processing unit 34 and a block correction processing unit 35.

  The image area separation error block pattern detection processing unit 34 includes an image area separation result indicated by the image area separation tag image data, a color flatness of each pixel indicated by the color flatness tag image data, and a contraction processing unit of the image area separation unit 18. 28 contraction processing results are input. The image area separation error block pattern detection processing unit 34 detects a part where an image area separation error has occurred in a block shape based on the image area separation result, the color flatness, and the contraction processing result, and the detection result is a block correction processing unit. 35.

  The block correction processing unit 35 receives the detection result of the image region separation error block pattern detection processing unit 34 and the image region separation result of the image region separation unit 18. The block correction processing unit 35 corrects and outputs the image region separation result of the image region separation unit 18 based on the detection result of the image region separation error block pattern detection processing unit 34. More specifically, a photographic region erroneously determined as a character region by the image region separation unit 18 is corrected so as to become an original photographic region.

  FIG. 17 is a block diagram illustrating a configuration of the image area separation error block pattern detection processing unit 34.

  The image area separation error block pattern detection processing unit 34 includes a block detection unit 36, a horizontal line detection unit 37, a vertical line detection unit 38, a block edge detection unit 39, an OR processing unit 40, an ON pixel addition unit 41, and threshold processing. Part 42.

  FIG. 18 is a diagram illustrating the configuration and processing contents of the block detection unit 36.

  The block detection unit 36 includes a first template matching processing unit 43, a second template matching processing unit 44, a third template matching processing unit 45, a fourth template matching processing unit 46, a fifth template matching processing unit 47, and an OR processing unit 48. The determination unit 49 is provided.

  The first template matching processing unit 43, the second template matching processing unit 44, the third template matching processing unit 45, the fourth template matching processing unit 46, and the fifth template matching processing unit 47 are configured with predetermined templates (detection patterns). ) Is used to perform template matching processing.

  More specifically, the first template matching processing unit 43 determines that the image area separation results of the 1 × 1 pixel area centered on the target pixel are all “1” (that is, the image area separation result of the target pixel). "1"), "1" is output, otherwise "0" is output.

  The second template matching processing unit 44 outputs “1” when the image area separation results of the 3 × 3 pixel region centered on the target pixel are all “1”, and outputs “0” otherwise. To do.

  The third template matching processing unit 45 outputs “1” when the image region separation results of the 5 × 5 pixel region centered on the target pixel are all “1”, and outputs “0” otherwise. To do.

  The fourth template matching processing unit 46 outputs “1” when the image region separation results of the 7 × 7 pixel region centered on the target pixel are all “1”, and “0” otherwise. Output.

  The fifth template matching processing unit 47 outputs “1” when the image region separation results of the 9 × 9 pixel region centered on the target pixel are all “1”, and outputs “0” otherwise. To do.

  Note that the template matching processing described here is an example, and the first template matching processing unit 43, the second template matching processing unit 44, the third template matching processing unit 45, the fourth template matching processing unit 46, and the It is not limited to the 5 template matching processing unit 47 alone.

  Output results from the first template matching processing unit 43, the second template matching processing unit 44, the third template matching processing unit 45, the fourth template matching processing unit 46, and the fifth template matching processing unit 47 are output to the OR processing unit 48. Is output.

  The OR processing unit 48 outputs results from the first template matching processing unit 43, the second template matching processing unit 44, the third template matching processing unit 45, the fourth template matching processing unit 46, and the fifth template matching processing unit 47. Is output as a template matching process result. That is, “1” is output to the determination unit 49 when “1” is output from at least one of the plurality of template matching processing units, and “0” is output to the determination unit 49 otherwise.

  The determination unit 49 receives the template matching processing result from the OR processing unit 48, the color flatness of the target pixel, and the contraction processing result in the target pixel contraction processing unit 28.

  In the determination unit 49, the template matching processing result for the target pixel is “1”, the color flatness is “1” or more, and the contraction processing result is “1: white pixel” (that is, in the contraction processing). “1” is output in the case of white pixels), and “0” is output in other cases.

  When “1” is output from the determination unit 49, the pixel of interest is imaged as a part of a block-shaped character region even though the change in saturation and the change in hue are small and flat. There is a high possibility that the area is separated.

  Here, in addition to the template matching processing result and the color flatness, the block detection is performed with reference to the contraction processing result. This is because a pixel that does not become a white pixel even after the shrinkage process is highly likely not to be a character area. By using the shrinkage process result, the accuracy of detecting an image area separation error can be further improved. it can.

  The horizontal line detection part 37 performs horizontal line detection using the template (horizontal line detection pattern) shown in FIG. Specifically, when the contraction processing result of the target pixel is “1: white pixel” and the image area separation result of all 1 × 11 pixels including the target pixel is “1”, “1” is output. To do.

  The vertical line detection unit 38 performs vertical line detection using the template (vertical line detection pattern) shown in FIG. Specifically, when the contraction processing result of the target pixel is “1: white pixel” and the image area separation result of all 11 × 1 pixels including the target pixel is “1”, “1” is output. To do.

  By the detection processing of the horizontal line detection unit 37 and the vertical line detection unit 38, it is possible to detect the horizontal lines and vertical lines constituting the edge of the block. The horizontal line detection pattern and the vertical line detection pattern described here are merely examples, and the size is not limited to 1 × 11 pixels or 11 × 1 pixels. The horizontal line detection unit 37 and the vertical line detection unit 38 use a detection pattern having a size larger than the size of the reference window used in the expansion processing unit 29 of the image area separation unit 18.

  In this case as well, vertical line and horizontal line detection is performed with reference to the contraction processing result for the same reason as the block detection described above.

  FIG. 21 is a diagram illustrating the configuration and processing contents of the block edge detection unit 39.

  The block edge detection unit 39 includes a first block edge detection processing unit 50, a second block edge detection processing unit 51, a third block edge detection processing unit 52, a fourth block edge detection processing unit 53, an OR processing unit 54, and a determination. A portion 55 is provided.

  The first block edge detection processing unit 50, the second block edge detection processing unit 51, the third block edge detection processing unit 52, and the fourth block edge detection processing unit 53 are respectively predetermined templates as shown in FIG. Whether or not there are a plurality of pixels whose edges are detected by the horizontal line detection unit 37 or the vertical line detection unit 38 continuously in a part away from the target pixel by a predetermined pixel (here, 5 pixels) using (detection pattern). If it exists, 1 is output.

  The OR processing unit 54 calculates the logical sum of the output results from the first block edge detection processing unit 50, the second block edge detection processing unit 51, the third block edge detection processing unit 52, and the fourth block edge detection processing unit 53. The result is output as a block edge detection result. That is, when “1” is output from at least one of the plurality of block edge detection processing units, “1” is output as the block edge detection result, and “0” is output otherwise.

  The determination unit 55 receives the block edge detection result, the color flatness of the target pixel, and the image area separation result of the target pixel from the OR processing unit 54.

  In the determination unit 55, when the block edge detection result for the target pixel is “1”, the color flatness is “1” or more, and the image area separation result is “1”, “1”. In other cases, “0” is output.

  Note that when “1” is output from the determination unit 55, the change in saturation and hue is small and flat, and is an edge portion of a vertical line or horizontal line pattern that is not a character or line drawing. Nevertheless, there is a high possibility that the pixel of interest is image-region separated as a character region.

  The detection result output from the block detection unit 36 and the detection result output from the block edge detection unit 39 are output to the OR processing unit 40. The OR processing unit 40 calculates the logical sum of the input detection results and outputs it to the ON pixel addition unit 41.

  The ON pixel adding unit 41 buffers the input detection result, and adds the detection result for each pixel included in a rectangular area of m × m pixels centered on the target pixel. For example, when the addition target is a rectangular area of 5 × 5 pixels, the addition result takes a value of 0 to 25.

  The threshold value processing unit 42 compares the addition result from the ON pixel addition unit 41 with a predetermined threshold value. “1” is output as the separation error block pattern detection result, and “0” is output to the block correction processing unit 35 if the addition result is less than the threshold value. Thereby, an image area separation error block pattern detection result is output for each pixel. The image area separation error block pattern detection result “1” indicates that a block pattern in which an area that should originally be a photographic area has been separated as a character area (that is, an image area separation error has occurred) has been detected. The image area separation error block pattern detection result “0” indicates that the block pattern in which the image area separation error has occurred is not detected.

  In this way, by adding the detection results for the pixels in a predetermined range including the target pixel and comparing it with the threshold value, the image area separation error can be detected with high accuracy.

  The block correction processing unit 35 corrects the image area separation result from the image area separation error block pattern detection result.

  FIG. 22 is a flowchart showing the flow of correction processing executed by the block correction processing unit 35.

  In step 300, a target pixel is set.

  In step 302, the image area separation result of the pixel of interest is acquired from the image area separation tag image data generated by the image area separation unit 18, and the image area separation for the pixel of interest is performed from the image area separation error block pattern detection processing unit 34. Get error block pattern detection result.

  In step 304, it is determined whether the image area separation result of the target pixel is “0 (photo area)” or “1 (character area)”. If it is determined that the image area separation result is “1”, the process proceeds to step 306 to determine whether the image area separation error block pattern detection result is “0” or “1”. To do. If it is determined that the image area separation error block pattern detection result is “1”, the image area separation result is corrected from “1” to “0” in step 308, and the process proceeds to step 310. .

  On the other hand, if it is determined in step 304 that the image area separation result is “0”, and if it is determined in step 306 that the image area separation error block pattern detection result is “0”, step 308 The process proceeds to step 310 without performing the correction process.

  In step 310, it is determined whether or not all pixels have been processed. If it is determined that all the pixels have not been processed, the process proceeds to step 312, the next pixel is set as the target pixel, the process returns to step 302, and the above process is repeated.

  On the other hand, when it is determined in step 310 that all pixels have been processed, the correction process ends.

  Thus, in this embodiment, when a block pattern in which an image area separation error has occurred is detected for a pixel determined to be a character area, the image area separation result of the pixel is corrected to a photographic area.

  The corrected image area separation tag image data and the color flatness tag image data calculated by the color flatness calculation unit 19 are output to the enlargement / reduction unit 16 and reduced / enlarged at the same magnification as that of the enlargement / reduction unit 14. It is output to the processing unit 17.

  The image processing unit 17 performs image processing on the image data output from the enlargement / reduction unit 14. At this time, an image area separation result indicated by the image area separation tag image data output from the reduction / enlargement unit 16, and Image processing is performed according to each color flatness indicated by the color flatness tag image data.

  FIG. 23 is an explanatory diagram illustrating the spatial frequency correction processing performed by the spatial correction unit 21 of the image processing unit 17.

  The spatial correction unit 21 performs filter processing on input image data (see FIG. 23A) using a filter (see FIG. 23B) in which DF (digital filter) coefficients are arranged in a matrix. To generate output data (see FIG. 23C) in which the spatial frequency is corrected.

  In this spatial frequency correction, different filters are used according to the image area separation result and the color flatness. In the present embodiment, as shown in FIG. 23A, the characteristics of the DF coefficient are determined according to the image area separation result and the color flatness, and a filter used for spatial frequency correction is selected according to this. A table in which the filter selection surface to be used for spatial frequency correction as shown in FIG. 23A is stored according to the image area separation result and the color flatness is stored and referred to in advance in a memory (not shown) or the like. .

  Note that the color conversion unit 22, the gradation correction unit 23, and the screen generation unit 24 in the image processing unit 17 also use the LUT and screen used for processing according to the image area separation result and the color flatness, similarly to the space correction unit 21. Etc. can be used with adaptive switching.

  The output data generated by image processing by the image processing unit 17 is output to the image output device 12, and the image output device 12 can output a high-quality image based on the output data.

  As described above, in the present embodiment, since the image area separation result is corrected based on the color flatness or the like, appropriate image processing can be performed according to the corrected image area separation result, and the image Image quality defects due to area separation errors can be reduced.

  In particular, it is originally necessary to correct the image area separation error itself, unlike the conventional method of blurring stepwise or continuously at the boundary part of the photograph area erroneously determined as the character area. The gradation can be reproduced. In the conventional method, an attempt is made to avoid the image area separation error by setting the detection parameter to that of a photograph, but as a result, the detection level of characters having a trade-off relationship is lowered. However, if the image area separation correction is performed based on the color flatness as described above, the character detection level does not decrease.

  In addition, when the image area separation method using a reference window or the like is used as described above, the conventional method that blurs in a stepwise or continuously changing manner at the boundary between the character area and the photographic area is low. What about block-type image area separation errors that occur when copying using a line number halftone dot, a YMCK multi-line (line) screen that is common in multi-value laser printers, or a document with an image formed on a rotation screen? It cannot be solved and the image quality defect cannot be corrected. However, if the block pattern is detected and corrected as described above, such an image area separation error can be corrected.

  Further, as described above, in this embodiment, for example, image area separation tag image data and color flatness tag image are not used for performing image area separation correction processing using YMCK image data of, for example, 8-bit 256 gradations. All data is generated in 1-bit (0, 1) or 2-bit (0-3) units, and these are used to perform image area separation correction processing. This not only shortens the processing time but also requires processing. There is also an advantage that the memory capacity is small.

  In addition, erroneous detection may occur when block pattern detection is performed on 8-bit multi-gradation image image data, but an image area separation error block pattern based on image area separation tag image data or color flatness tag image data. Since the detection process is performed, the correction can be performed with high accuracy without erroneous detection.

  In addition, since the color flatness is used not only in the image area separation correction process but also in the image processing of the subsequent image processing unit, optimal image processing corresponding to the color flatness can be performed.

  The present invention is not limited to the above-described embodiments, and various design changes can be made within the scope of the invention described in the claims.

  For example, in the above embodiment, an example in which an input image is separated into a character area and a photographic area and image processing is performed according to each area has been described. However, the present invention is not limited to this. The image area may be separated into four areas, a halftone dot area and a photographic area, and image processing may be performed according to each area. Hereinafter, a modification example in which the image area is separated into four areas will be described.

  The basic configuration of this modification is the same as the configuration shown in FIG. However, the image area separation unit is configured as follows, and the reference numeral 18a is described as being distinguished from the image area separation unit 18 of the above embodiment.

  FIG. 24 is a diagram illustrating a configuration of the image area separation unit 18a according to the modification. In FIG. 24, the same or equivalent parts as in FIG. 3 are denoted by the same reference numerals, and the description thereof is simplified or omitted.

  The image area separation unit 18a includes a binarization processing unit 25, a halftone dot extraction unit 26, an OR processing unit 27, a contraction processing unit 28, an expansion processing unit 29, an inversion processing unit 30, an AND processing unit 31, and an image region separation tag. A generation unit 56 and a color black determination unit 57 are included.

  In the above embodiment, the processing result (0 or 1) output from the AND processing unit 31 has been described as an example in which 1 is a character region and 0 is a photo region. In this modification, 1 is a character region. , 0 will be described as a non-character area.

  As in the above embodiment, the halftone dot extracting unit 26 extracts a halftone dot region from the image data output from the color converting unit 13, and binary image data (0: non-halftone dot) of the extracted halftone dot region. Dot, 1: halftone dot). Furthermore, in this modification, the halftone dot extraction unit 26 outputs the extraction result of the halftone dot region to the image area separation tag generation unit 56 and the image area correction unit 20a (see FIG. 27).

  The color black determination unit 57 uses the a * value and b * value of the image data in the L * a * b * color space to divide the input image into a region with high saturation (color region) and a region with low saturation (color region). Black area). Specifically, the coordinate value on the C * axis is calculated using the above-described equation (5), and the calculated value is compared with a predetermined threshold value to determine whether the color region is a black region or not. . This determination result (0: black, 1: color) is output to the image area separation tag generator 56.

  Hereinafter, the determination result of the color / black determination unit 57 input to the image area separation tag generation unit 56 is referred to as D0, the extraction result of the halftone extraction unit 26 is referred to as D1, and the processing result of the AND processing unit 31 is referred to as D2. .

  FIG. 25 is a table showing the relationship between the input data to the image area separation tag generation unit 56 and the output data output from the image area separation tag generation unit 56.

  As described above, D0, D1, and D2 are input to the image area separation tag generation unit 56 for each pixel as input data. Since each input data D0, D1, D2 is 1-bit data of 0 or 1, the input to the image area separation tag generator 56 is 3 bits.

  The image area separation tag generator 56 converts the 3-bit input data into 2-bit output data (TD0, TD1) according to the relationship shown in the table of FIG. 25, and converts TD1 as shown in the table of FIG. Image area separation tag image data of 2 bits for each pixel having the upper bit (second bit) and TD0 as the lower bit (first bit) is generated.

  The output data TD1 becomes the input data D2 (character, non-character) from the AND processing unit 31. The output data TD0 is the input data D1 when TD1 indicates the non-character area (0), and the input data D0 when TD1 indicates the character area (1).

  The image area separation tag image data generated here is corrected by the subsequent image area correction processing. In this modification, the image area correction unit that performs this correction is configured as follows, and the code is 20a. The description will be made separately from the image area correction unit 20 of the above embodiment.

  FIG. 27 is a diagram illustrating a configuration of the image area correction unit 20a according to the modification. In FIG. 27, parts that are the same as or equivalent to those in FIG. 16 are given the same reference numerals, and descriptions thereof are simplified or omitted.

  The image area correction unit 20 includes a tag cutout processing unit 58, an image area separation error block pattern detection processing unit 34, and a block correction processing unit 59.

  The tag cutout processing unit 58 cuts out the upper one bit corresponding to the output data TD1 from the image region separation tag image data of 2 bits for each pixel and outputs it to the image region separation error block pattern detection processing unit 34. As described above, since the output data TD1 corresponds to the input data D2 (character, non-character) from the AND processing unit 31, the image area separation error block pattern detection processing unit 34 performs the detection described in the above embodiment. Detection processing similar to the processing can be performed.

  The block correction processing unit 59 corrects the 2-bit image area separation tag image data according to the image area separation error block pattern detection result.

  FIG. 28 is a flowchart showing the flow of correction processing executed by the block correction processing unit 59 in the modification.

  In step 400, a target pixel is set.

  In step 402, the image area separation result of the target pixel is obtained from the image area separation tag image data generated by the image area separation unit 18a, and the halftone dot extraction result (D1) of the halftone dot extraction unit 26 is obtained. Furthermore, the image area separation error block pattern detection result for the target pixel is acquired from the image area separation error block pattern detection processing unit 34.

  In step 404, it is determined whether the upper bit (TD1) of the image area separation result of the target pixel is “0” or “1”. If it is determined that the upper bit (TD1) of the image area separation result is “1”, the process proceeds to step 406, where the image area separation error block pattern detection result is “0” or “1”. Is determined. If it is determined that the image area separation error block pattern detection result is “1”, in step 408, the image area separation result is changed to a photographic area or a halftone area according to the halftone extraction result D1. To do. Specifically, the upper bit TD1 is corrected from “1” to “0”, and the lower bit TD0 is corrected to the same value as the halftone extraction result D1. After the correction process in step 408, the process proceeds to step 410.

  On the other hand, if it is determined in step 404 that the upper bit (TD1) of the image area separation result is “0”, and if it is determined in step 406 that the image area separation error block pattern detection result is “0”. In step S408, the process proceeds to step S410 without performing the correction process in step S408.

  In step 410, it is determined whether all pixels have been processed. If it is determined that all the pixels have not been processed, the process proceeds to step 412 to set the next pixel as a target pixel and return to step 402 to repeat the above process.

  On the other hand, when it is determined in step 410 that all pixels have been processed, the correction process ends.

  As described above, in this modified example, when a block pattern in which an image area separation error has occurred is detected for a pixel determined to be a black character area or a color character area, the image area separation result of the pixel is displayed as a photographic area or a network. Correct to point area.

  The corrected image area separation tag image data and the color flatness tag image data calculated by the color flatness calculation unit 19 are output to the enlargement / reduction unit 16 and reduced / enlarged at the same magnification as that of the enlargement / reduction unit 14. It is output to the processing unit 17.

  The image processing unit 17 performs image processing on the image data output from the enlargement / reduction unit 14. At this time, an image area separation result indicated by the image area separation tag image data output from the reduction / enlargement unit 16, and Image processing is performed in the same manner as in the above embodiment in accordance with each of the color flatness indicated by the color flatness tag image data. In this modification, the image area is divided into four types of areas, so that image processing suitable for each area can be performed.

  Although the modification has been described above, the same effect as that described in the above embodiment can be obtained in the modification.

  In the embodiment and the modification described above, the color flatness calculated by the color flatness calculating unit is “3: color flat region”, “2: intermediate flat region”, “1: black flat region”, “0: Although the four values of “white flat / non-flat region” are used, the present invention is not limited to this, and, for example, two colors, “1: flat region” with high flatness and “0: non-flat region” with low flatness, are not color-coded. Only the value may be used. This also corrects the image area separation error.

  Further, the image area separation processing in the image area separation units 18 and 18a is not limited to the method described in the above-described embodiment and modification examples. For example, a character area and a photograph area (halftone dot area) may be separated only by pattern matching using a rectangular template. As a simple pattern matching process, scoring is performed based on the degree of matching with the template pattern, and the result of scoring a window of a predetermined size is added to the target pixel, and then an average value is calculated. There is a method of separating into a halftone dot region and a non-halftone dot region (character region) by a comparison process comparing the values. Even with such a technique, since the template used for pattern matching is rectangular, a block-shaped image area separation error may occur in the separation boundary area. Since image area separation using a template is a widely performed technique, it is possible to correct such block shape image area separation errors with high accuracy by correcting using block pattern detection results as described above. it can.

  Further, when an image area separation method that does not perform contraction processing or expansion processing is adopted, the image area separation error can be corrected based on only the color flatness and the block pattern detection result without considering the contraction processing result. it can. It should be noted that the image area separation method described in the above embodiments and modifications can also correct the image area separation error without using the contraction processing result. As a result, although there is a slight decrease in accuracy, the image area separation process can be performed with higher accuracy than in the past.

  It is also possible to correct the image area separation error based only on the color flatness. For example, in an image area separation method in which a block-shaped image area separation error does not occur (for example, when separating into a character area and a non-character area only by edge extraction processing), the image area separation error is corrected only by color flatness. You may do it.

  Even with an image area separation method in which a block-shaped image area separation error occurs, it is possible to correct the image area separation error based only on the color flatness. In the image gamut separation methods described in the above embodiments and modifications, if an image gamut separation error is corrected only by color flatness, there is a possibility that characters in a color background such as catalog headline characters will be corrected to a photographic region. However, it is possible to prevent deterioration in image quality due to switching of image processing between characters and photographs (or halftone dots). Therefore, when the color flatness is “3 (color flat area)”, a high-quality output result can be obtained even if the portion of the image area separated into the character area is corrected to the photographic area (or halftone dot area).

  In the embodiment and the modification described above, an example has been described in which RGB image data is converted into image data in the L * a * b * color space, and image area separation processing, flatness calculation processing, and image processing are performed. The image data is not limited to the L * a * b * color space, and may be converted into image data of another color space as long as it is image data of a uniform color space.

1 is a block diagram showing a configuration of a copying machine according to an embodiment of the present invention. It is a block diagram which shows the structure of an image process part. It is a block diagram which shows the structure of an image area separation part. It is explanatory drawing explaining the contraction process of a contraction process part. It is explanatory drawing explaining the expansion process of an expansion process part. It is a figure which shows an example of the image data which took the logical product of the image data input into the AND process part of an image area separation part, and this input image data. It is the figure which showed an example of the processing result in each part of an image area separation part. FIG. 5 is an explanatory diagram for explaining a block-shaped image area separation error that occurs when a document image printed based on image data screen-processed by the dot TO line screen method is copied. FIG. 6 is an explanatory diagram for explaining block-shaped white spots that occur when copying a document image expressed by a low line number halftone dot. It is a flowchart which shows the flow of the main routine of the color flatness calculation process which a color flatness calculation part performs. It is a flowchart which shows the flow of the subroutine called by the main routine of a color flatness calculation process. It is a figure which shows a state when the image which the image data of L * a * b * color space represents is divided | segmented into the several small area | region of MxN pixel. It is a figure which shows the threshold value for determining each pixel to any of a color area | region, a white area | region, a black area | region, and an intermediate | middle area | region on a L * C * plane. It is explanatory drawing explaining a flat hue angle. It is a table showing the type of color flatness. It is a block diagram which shows the structure of an image area correction | amendment part. It is a block diagram which shows the structure of an image area separation error block pattern detection process part. It is a figure which shows the structure and process content of a block detection part. It is a figure which shows an example of the template used by the horizontal line detection of a horizontal line detection part. It is a figure which shows an example of the template used for a vertical line detection in a vertical line detection part. It is a figure which shows the structure and process content of a block edge detection part. It is a flowchart which shows the flow of the correction process which a block correction process part performs. It is explanatory drawing explaining the spatial frequency correction process performed in the spatial correction part of an image process part. It is a figure which shows the structure of the image area separation part in a modification. 10 is a table showing a relationship between input data to an image area separation tag generation unit and output data output from the image area separation tag generation unit 56 in a modified example. It is a table | surface which shows the value of each bit, and the image area separation content corresponding to this at the time of producing | generating image area separation tag image data of each pixel 2 bits. It is a figure which shows the structure of the image area correction | amendment part in a modification. It is a flowchart which shows the flow of the correction process which the block correction process part in a modification performs. It is explanatory drawing explaining that an image quality degrades by performing an image process in the state in which the image area separation error generate | occur | produced, and outputting an image.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Image reader 11 Image processing apparatus 12 Image output apparatus 15 Image area separation process part 17 Image process part 18, 18a Image area separation part 19 Color flatness calculation part 20, 20a Image area correction part 25 Binarization process part 26 Network Point extraction unit 28 Shrinkage processing unit 29 Expansion processing unit 30 Inversion processing unit 32 Image region separation tag generation unit 34 Image region separation error block pattern detection processing unit 35 Block correction processing unit 36 Block detection unit 37 Horizontal line detection unit 38 Vertical line detection unit 39 Block edge detection unit 56 Image area separation tag generation unit 57 Color black determination unit 59 Block correction processing unit

Claims (11)

Separating means for separating the image data of the image read by the image reading means into a character or line drawing character line drawing area and a non-character line drawing area that is not a character line drawing area;
Flatness deriving means for deriving flatness for each pixel, which indicates whether each pixel is a flat region where the change in the image data is a predetermined value or less based on the image data of the read image;
Correction means for correcting a separation failure included in the separation result of the separation means based on the derived flatness;
An image processing apparatus. The image processing apparatus according to claim 1, wherein the flatness deriving unit derives the flatness of each pixel based on a distribution from average coordinates of image data of each pixel and each pixel neighboring area in a uniform color space. The flatness deriving means sequentially sets each pixel of the read image as a target pixel, the degree of variation in image data of the set target pixel and the region near the target pixel, and the image data of the target pixel Calculating the deviation from the average value of the image data of the target pixel and the region near the target pixel, and deriving the flatness of the target pixel based on the calculated degree of variation and the deviation. The image processing apparatus according to claim 1, wherein flatness is derived for each. The flatness deriving unit calculates a standard deviation of data indicating hue and saturation included in the image data as the degree of variation, and the image data of the pixel of interest and the region near the pixel of interest as the deviation. The image processing apparatus according to claim 3, wherein the difference between the average of the hue data included and the hue data included in the image data of the target pixel is calculated. 5. The image processing apparatus according to claim 1, further comprising an image processing unit that performs image processing on the image data of the read image according to the separation result corrected by the correction unit. . The image according to claim 5, wherein the image processing means performs image processing on the image data of the read image according to the separation result corrected by the correction means and the flatness derived by the flatness derivation means. Processing equipment. The separation unit sequentially sets each pixel of the read image as a target pixel, extracts image data of a rectangular area including the set target pixel from the image data of the read image, and extracts the extracted image The image data of the read image is separated into a character line drawing area and a non-character line drawing area by performing predetermined processing on the data, and further, the separation result for each pixel is obtained at the position of each pixel of the read image. Generate separation result arrangement data arranged correspondingly as the separation result,
The correction means uses the derived flatness and a predetermined rectangular model,
The image processing apparatus according to any one of claims 1 to 6, wherein a separation defect having a shape of a rectangle or a series of rectangles is detected from the generated separation result arrangement data, and the detected separation defect is corrected. The separation unit sequentially sets each pixel of the read image as a target pixel, extracts image data of a rectangular area including the set target pixel from the image data of the read image, and extracts the extracted image Image data that has undergone contraction processing that contracts the black portion of the data, and then expands processing that expands the black portion, and image data and contraction processing obtained by inverting the image data after the contraction processing and expansion processing have been completed for all pixels of interest A character / line drawing area is separated from the image data of the read image based on the previous image data, and image data other than the separated character / line drawing area is separated as a non-character / line drawing area, and the separation result arrangement data is obtained. Generate and
The correction means uses the derived flatness, a predetermined rectangular template, and the image data after the shrinkage processing to separate a rectangle or a shape in which rectangles are connected from the generated separation result arrangement data. The image processing apparatus according to claim 7, wherein a defect is detected and the detected separation defect is corrected. The image processing apparatus according to claim 1, wherein the separating unit further separates the non-character / line drawing area into a halftone dot area and a photograph area. Image reading means for reading an image of a document;
Separating means for separating the image data of the read image into a character or line drawing character line drawing area and a non-character line drawing area that is not a character line drawing area;
Flatness deriving means for deriving flatness for each pixel, which indicates whether each pixel is a flat region where the change in the image data is a predetermined value or less based on the image data of the read image;
Correction means for correcting a separation failure included in the separation result of the separation means based on the derived flatness;
Image processing means for performing image processing on the image data of the read image according to the separation result corrected by the correction means;
Image output means for outputting an image based on image data subjected to image processing by the image processing means;
An image output device. A separation step of separating the image data of the image read by the image reading unit into a character or line drawing character line drawing area and a non-character line drawing area that is not a character line drawing area;
A flatness derivation step for deriving flatness for each pixel, which indicates whether each pixel is a flat region whose change in image data is a predetermined value or less, based on the image data of the read image;
A correction step of correcting a separation failure included in the separation result of the separation step based on the derived flatness;
An image processing method comprising: