JP2009194740A - Halftone dot background area detection method, image processing apparatus, image forming apparatus, computer program, and recording medium - Google Patents

Abstract

A halftone dot background detection method capable of accurately detecting a halftone dot background region from input image data, an image processing apparatus for performing halftone dot background detection processing using this method, and a halftone dot background region The present invention provides an image forming apparatus that forms an image based on the detection result, a computer program that detects a halftone dot background area, and a recording medium on which the computer program is recorded.
A character feature amount calculation unit 41 of a region separation processing unit 24 calculates an edge degree as a character feature amount from an input image, and a dot feature amount calculation unit 42 calculates a complexity degree as a dot feature amount. The photograph feature value calculation unit 43 calculates an entropy value as a feature value. The gray scale conversion processing unit 25 detects a halftone dot background region from the input image based on the calculated feature values. The halftone dot background region is detected as a region where the halftone dot feature amount exceeds a predetermined amount and the character feature amount and the photographic feature amount do not exceed the predetermined amount, that is, a halftone dot region that is not a character region and a photo region.
[Selection] Figure 3

Description

  The present invention relates to a halftone dot region detection method for detecting a dot region (halftone dot region) composed of halftone dots from an input image, and an image processing apparatus for detecting halftone dot regions using this method. The present invention relates to an image forming apparatus for forming an image based on a detection result of a halftone dot background area, a computer program for detecting a halftone dot background area, and a recording medium on which the computer program is recorded.

  2. Description of the Related Art Conventionally, an image forming apparatus that forms an image relating to a document read by a scanner or a document supplied from a PC (personal computer) or the like on a printing sheet by an electrophotographic method or an inkjet method has been widely used. The types of originals handled by the image forming apparatus include, for example, character originals, halftone originals, photo originals, and mixed originals. In order to reproduce the original image with higher image quality, depending on the original type It is necessary to perform optimum image processing and image forming processing.

  In addition, with the development of image processing technology, image forming apparatuses can reproduce color (multicolor) images with high image quality. In recent years, high-performance image formation such as full-color digital copiers and multi-function peripherals. The device has been commercialized. However, even in such an image forming apparatus, from the viewpoint of saving toner or ink or the like, there are not a few chances for image formation in black monochrome, so-called monochrome (hereinafter simply referred to as monochrome) output.

  A halftone original is an image formed by arranging a large number of small dots, and it is possible to perform pseudo gradation expression with a limited number of colors by appropriately adjusting the size of the dots and the interval between the dots. Therefore, this is a technique that is frequently used when forming an image expressed in multiple gradations. In addition, there are many opportunities to use a manuscript in which characters are drawn on an image expressed by halftone dots, that is, a manuscript composed of a halftone dot background and halftone dots. However, when a color original composed of a halftone dot background and halftone dots is output in monochrome as described above, the visibility or legibility of the halftone characters may be reduced.

  Patent Document 1 proposes an image reproduction apparatus capable of reproducing an image so as to improve the legibility of a character on a halftone dot or a character on a color ground when reproducing a color image in monochrome. Has been. This image reproduction device generates a monochrome signal from a color input signal, detects a color edge, a monochrome edge, or a halftone dot region from the input signal, and performs a gamma conversion process that positively contrasts around the character. Or, whitening processing is performed.

Japanese Patent Application Laid-Open No. H10-228688 proposes an image area discrimination device that can accurately discriminate a character area on a halftone dot. This image area discriminating apparatus includes a halftone dot discriminating unit that discriminates a halftone dot region from an input image, a background level discriminating unit that discriminates a background level, an average brightness calculating unit that calculates an average brightness of a predetermined region, and an edge. An edge discriminating unit for discriminating. The halftone dot discrimination unit outputs the discrimination result as a halftone dot area signal. The comparison unit compares the threshold calculated according to the background level determined by the background level determination unit and the average brightness calculated by the average lightness calculation unit, and the comparison unit outputs the comparison result as a character area signal. Further, the edge determination unit outputs the determination result as an edge region signal. The image area discriminating apparatus calculates a logical product of the halftone dot area signal, the character area signal, and the edge area signal, and outputs the calculation result as a halftone dot character edge area signal.
JP 2001-352453 A JP 2002-368993 A

  Many attempts have been made to improve the reduction in visibility that occurs when a color document is output in black and white, and as described above, it is important to improve the visibility and legibility of characters on a halftone dot background. . However, the image reproduction apparatus described in Patent Document 1 uses only the edge information of a character when detecting a character on a halftone dot background in order to add contrast around the character on the halftone dot background. It is carried out. For this reason, when the difference between the density of the halftone dots and the density of the characters is small, there is a possibility that the edge of the character cannot be sufficiently detected, and it is sufficient to improve the visibility and legibility of the characters on the halftone dot background May not be able to ensure accurate accuracy.

  The image area discriminating apparatus described in Patent Document 2 detects a character area on a halftone dot background by comparing the threshold value calculated from the background level (the density of the halftone dot background) with the average lightness (character density). It is the structure to do. For this reason, as in the above-described image reproducing apparatus of Patent Document 1, when the difference between the density of the halftone dots and the density of the characters is small, the character area cannot be detected with sufficient accuracy by comparing the densities. There is a possibility that sufficient accuracy for improving the visibility and legibility of characters on the halftone dot background cannot be secured.

  The present invention has been made in view of such circumstances, and an object of the present invention is to input an original even when the difference between the density of a halftone dot background and the density of characters on a halftone dot is small. Halftone dot region detection method capable of accurately detecting a halftone dot region from an image according to the present invention, an image processing apparatus for detecting a halftone dot region using this method, based on the detection result of the halftone dot region Another object of the present invention is to provide an image forming apparatus for forming an image, a computer program for detecting a halftone dot background area, and a recording medium on which the computer program is recorded.

  A halftone dot background region detection method according to the present invention is a halftone dot background region detection method for detecting a background region composed of halftone dots from an image, and calculates a character feature amount indicating a feature of a character image from the image. A halftone dot feature amount indicating a feature of a halftone dot image is calculated from the image, a photographic feature amount indicating a feature of a photographic image is calculated from the image, and it is determined whether or not the calculated character feature amount exceeds a predetermined amount. Determining whether the calculated halftone dot feature amount exceeds a predetermined amount, determining whether the calculated photograph feature amount exceeds a predetermined amount, and based on a combination of the determination results, Is detected.

  The halftone dot background region detection method according to the present invention is characterized in that, as the character feature amount, a gradient between a pixel value of each pixel and a pixel value of a peripheral pixel of the pixel is calculated from the image.

  The halftone dot background region detection method according to the present invention is characterized in that, as the halftone dot feature amount, a degree of change in pixel values of adjacent pixels in one or a plurality of directions is calculated from the image.

  The halftone dot background region detection method according to the present invention is characterized in that entropy with respect to the appearance frequency of a pixel value is calculated for each predetermined region of the image as the photographic feature amount.

  In the halftone dot background region detection method according to the present invention, it is determined that the halftone dot feature amount exceeds a predetermined amount, and it is determined that the character feature amount and the photographic feature amount do not exceed the predetermined amount, respectively. An area is detected as the halftone dot background area.

  An image processing apparatus according to the present invention includes a halftone dot background area detecting unit that detects a background area composed of halftone dots from an input image, and shows the characteristics of a character image from the image. Character feature amount calculating means for calculating a character feature amount, halftone dot feature amount calculating means for calculating a halftone dot feature amount indicating a feature of a halftone image from the image, and a photographic feature amount indicating a feature of a photographic image from the image Calculated by the photograph feature amount calculating means, the character feature amount determining means for determining whether or not the character feature amount calculated by the character feature amount calculating means exceeds a predetermined amount, and the halftone dot feature amount calculating means A halftone dot feature amount determination unit that determines whether or not a halftone dot feature amount exceeds a predetermined amount, and a photographic feature amount determination that determines whether or not the photographic feature amount calculated by the photographic feature amount calculation unit exceeds a predetermined amount Means. Land area detecting means, wherein the text feature quantity determining means, are to detect the halftone background region based on a combination of the determination result by the halftone characteristic amount determination unit and a photographic characteristic amount determination unit.

  An image forming apparatus according to the present invention includes the above-described image processing apparatus, an input unit that inputs an image to the image processing apparatus, and a halftone dot background area detection unit of the image processing apparatus. An image correcting unit that performs image correction on the halftone dot background area detected by the image correction unit, and an image forming unit that forms an image based on the image subjected to the image correction by the image correcting unit.

  According to another aspect of the present invention, there is provided a computer program for causing a computer to perform a process of detecting a background area composed of halftone dots from an input image. The computer shows characteristics of a character image from the image. Calculating a character feature amount; calculating a halftone dot feature amount indicating a feature of a halftone image from the image; calculating a photographic feature amount indicating a feature of a photographic image from the image; Determining whether the character feature amount exceeds a predetermined amount; determining whether the calculated halftone dot feature amount exceeds a predetermined amount; and whether the calculated photographic feature amount exceeds a predetermined amount And a step of detecting a halftone dot background region based on a combination of the determination results. .

  A recording medium according to the present invention is recorded with the computer program described above.

  In the present invention, a character feature amount indicating a character image feature, a halftone dot feature amount indicating a halftone image feature, and a photographic feature amount indicating a photographic image feature are calculated from an input image. By determining whether or not each calculated feature amount exceeds a predetermined threshold value, whether the image is a character region, a halftone dot region, or a photo region It can be judged whether there is. Therefore, the entire input image is determined, and the halftone dot background region can be detected from the image by the combination of the determination results based on each feature amount, and image processing such as density correction is performed on the halftone dot background region. It can be applied with high accuracy.

  At this time, as the character feature amount, a gradient (that is, an edge degree or an edge gradient) between the pixel value of each pixel of the input image and the pixel values of the surrounding pixels can be calculated. The edge degree can be calculated using, for example, a Sobel filter. By calculating the edge degree rather than detecting only the presence or absence of an edge, it is possible to extract a possible edge without missing even if the edge degree is small.

  In addition, as the halftone dot feature amount, the degree of change of adjacent pixel values can be calculated in one or more directions of an image in which a plurality of pixels are arranged vertically and horizontally. ) Can be calculated. The calculation direction preferably includes the main scanning direction of the image, and the directions other than the main scanning direction include, for example, a direction perpendicular to the main scanning direction and an angle of 45 ° with respect to the main scanning direction. There are directions. Rather than detecting only the presence or absence of halftone dots, by calculating the complexity of the halftone dots, it is possible to extract the possible halftone dots without omission even if the complexity is small.

  In addition, as a photograph feature amount, an appearance frequency (that is, a histogram) of pixel values can be obtained for each predetermined region of an input image, and entropy (average information amount) of the histogram can be calculated. When a histogram is created for the density of each pixel in the image, a histogram with a wide range of density changes is obtained in the photographic image. Therefore, when entropy is calculated from the histogram for a photographic image, the entropy value is larger than for other images. Thus, by calculating the entropy of the histogram, it is possible to accurately extract a photographic region.

  By determining whether or not the calculated character feature amount exceeds a predetermined amount, it is possible to extract a region that is highly likely to be included in the character region from the image, and whether or not the halftone dot feature amount exceeds a predetermined amount. An area that is likely to be included in the halftone dot area can be extracted by determining, and an area that is highly likely to be included in the photographic area by determining whether or not the photograph feature amount exceeds a predetermined amount Can be extracted. The halftone dot background area is an area that is included in the halftone dot area and not included in the character area and the photograph area. Therefore, it is possible to detect an area where it is determined that the halftone dot feature amount exceeds the predetermined amount and the character feature amount and the photographic feature amount do not exceed the predetermined amount as the halftone dot background region.

  In the case of the present invention, the character feature value, the halftone dot feature value, and the photographic feature value are calculated from the input image, and an area that is included in the halftone area and not included in the character area and the photographic area is halftone dot background. With the configuration of detecting as a region, a halftone dot background region can be detected from an image with high accuracy, so that image processing suitable for the halftone dot background region can be performed to form an image. Therefore, it is possible to accurately perform image processing such as density correction so as to widen the difference between the density of the halftone dot background area and the density of the character area, and the character on the halftone dot background when monochrome output is performed. Visibility and legibility can be improved with certainty.

  Hereinafter, the present invention will be specifically described with reference to the drawings showing embodiments thereof. FIG. 1 is a block diagram showing a configuration of an image forming apparatus according to the present invention. In the figure, reference numeral 1 denotes a CPU (Central Processing Unit) that performs control processing of various parts of the image forming apparatus and various arithmetic processing. The CPU 1 includes functions such as a ROM (Read Only Memory) 2, a RAM (Random Access Memory) 3, an operation unit 4, a display unit 5, a memory mounting unit 6, an image input device 11, an image output device 12, and an image processing device 20. The blocks are connected, and these functional blocks operate under the control of the CPU 1.

  Note that the image forming apparatus shown in this embodiment is a so-called digital color copying machine, but has a function of outputting an image in monochrome. FIG. 1 illustrates functional blocks related to monochrome output of the image forming apparatus. It is. When the image forming apparatus performs monochrome output, R (red), G (green), and B (blue) signals input from the image input apparatus 11 to the image processing apparatus 20 are output by the image processing apparatus 20 to K (black). ) And output to the image output device 12.

  The ROM 2 is configured by a non-volatile memory element such as a mask ROM, an EEPROM (Electrically Erasable Programmable ROM) or a flash memory, and stores a program executed by the CPU 1 and data necessary for the execution in advance. After the image forming apparatus is powered on, the CPU 1 reads out the program from the ROM 2 and executes it to start processing. The RAM 3 is composed of a large-capacity memory element capable of rewriting data by SRAM (Static RAM), DRAM (Dynamic RAM), flash memory, or the like, and is set by the user and data generated when the CPU 1 performs arithmetic processing. Various setting values are stored.

  The operation unit 4 includes a function key, a start key, a numeric keypad, a power key, and the like, and can accept an operation such as a process start command or a setting input from the user. The operation unit 4 gives the received user operation to the CPU 1. The display unit 5 includes a liquid crystal display, and displays an operation state of the image forming apparatus, a warning message to the user, various menus for assisting the user's operation, and the like on the liquid crystal display based on a display command from the CPU 1. Note that the operation unit 4 and the display unit 5 may be a touch panel configured integrally with each other.

  The memory mounting unit 6 is a so-called memory card slot, and a memory card 100 such as an SD memory card or a memory stick can be mounted detachably. When the memory card 100 is loaded in the memory loading unit 6, the CPU 1 can read data stored in the memory card 100 and write data to the memory card 100. When the program code 101 is stored in the memory card 100, the CPU 1 reads the program code 101 from the memory card 100 and stores it in the RAM 3, and executes the program code 101 to perform processing related to the program code 101. It can be carried out. If the ROM 2 is composed of a rewritable memory element, the program code 101 read from the memory card 100 may be stored in the ROM 2.

  The image input device 11 includes a light source that irradiates light on a document, and a scanner that includes a light receiving unit in which an image sensor such as a CCD (Charge Coupled Device) is arranged in parallel. Based on the control of the CPU 1, the image input device 11 irradiates light from the light source to the document and moves the light receiving unit in one direction to separate the reflected light from the document into three color components of RGB and receive the light. Read by the image processing unit and provided to the image processing apparatus 20 as RGB analog signals.

  The image processing apparatus 20 includes an A / D conversion unit 21, a shading correction unit 22, an input tone correction unit 23, a region separation processing unit 24, a gray scale conversion processing unit 25, a spatial filter processing unit 26, a scaling processing unit 27, An output tone correction unit 28 and a tone reproduction processing unit 29 are provided. The A / D converter 21 converts the RGB analog signal supplied from the image input device 11 into, for example, 8-bit digital signals, that is, RGB image data. The converted RGB image data is subjected to shading correction. Part 22 is given.

  The shading correction unit 22 performs a shading process on the RGB image data given from the A / D conversion unit 21 to remove various distortions generated in the illumination system, imaging system, imaging system, and the like of the image input device 11. Then, the RGB image data after the shading process is given to the input tone correction unit 23. The input tone correction unit 23 performs processing such as color balance adjustment and contrast adjustment on the RGB image data from which distortion has been removed by the shading correction unit 22, and the processed RGB image data is processed. This is given to the region separation processing unit 24.

  For each pixel in the RGB image data given from the input tone correction unit 23, the region separation processing unit 24 is a character feature amount indicating a character image feature, a dot feature amount indicating a halftone image feature, and a photograph. Photo feature quantities indicating the features of an image (continuous tone area such as a photographic paper photograph) are calculated, and each pixel belongs to any of a plurality of areas such as a character area, a halftone dot area, or a photo area based on the calculation result. Is determined, and each pixel is separated. The region separation processing unit 24 provides information (region separation result) indicating which region each pixel belongs to as a processing result to the spatial filter processing unit 26 and the gradation reproduction processing unit 29, and also performs input gradation correction. The RGB image data supplied from the unit 23 is supplied to the grayscale conversion processing unit 25 as it is. In addition, when separating the character area, the area separation processing unit 24 performs edge detection processing of the image data and performs separation based on the detection result of the edge. Edge detection of image data can be performed by performing image processing using a Sobel filter, for example. In addition, the region separation processing unit 24 provides the character scale, halftone dot feature amount, and photographic feature amount of each pixel calculated for region separation to the grayscale conversion processing unit 25.

  The grayscale conversion processing unit 25 uses the RGB color image data given from the region separation processing unit 24 based on the character feature value, halftone dot feature value, and photographic feature value given from the region separation processing unit 24. A process of converting to monochrome image data by K is performed. The gray scale conversion processing unit 25 gives the converted K image data to the spatial filter processing unit 26. Details of processing performed by the region separation processing unit 24 and the gray scale conversion processing unit 25 will be described later.

  The spatial filter processing unit 26 performs spatial filter processing using a digital filter on the K image data provided from the grayscale conversion processing unit 25 based on the region separation result provided from the region separation processing unit 24. As a result, the spatial frequency characteristics of the image are corrected, and it is possible to prevent the image output from the image output device 12 from being blurred or deteriorated in graininess. For example, in a region separated into character regions by the region separation processing unit 24, the spatial filter processing unit 26 performs sharp emphasis filter processing to emphasize high-frequency components, particularly for the purpose of improving character reproducibility. . Further, for example, the spatial filter processing unit 26 performs low pass filter processing on the region separated into halftone dot regions by the region separation processing unit 24 to remove the input halftone component. The K image data subjected to the spatial filter processing in this way is supplied from the spatial filter processing unit 26 to the scaling processing unit 27.

  In the image forming apparatus, the user can set enlargement / reduction of an image to be formed by operating the operation unit 4. The scaling processing unit 27 performs an enlargement calculation process or a reduction calculation process on the K image data that has been subjected to the spatial filter process by the spatial filter process unit 26 so as to obtain a predetermined magnification set by the user. The K image data that has been subjected to the scaling process of the scaling process unit 27 is subjected to various correction processes by the output tone correction unit 28 and is then supplied to the tone reproduction processing unit 29.

  The gradation reproduction processing unit 29 performs gradation reproduction processing on the K image data provided from the output gradation correction unit 28 in accordance with the region separation result provided from the region separation processing unit 24. The gradation reproduction process is a process of binarization or multi-value processing so that image data is classified into a plurality of pixels and intermediate gradations can be reproduced. For example, in a region separated into character regions by the region separation processing unit 24, the tone reproduction processing unit 29 reproduces high frequency components of the image output device 12 for the purpose of particularly improving the reproducibility of characters. Appropriate binarization processing or multilevel processing is performed. Further, for example, in an area separated into halftone dot areas by the area separation processing unit 24, the output tone correction unit 28 converts the density of the image data into a halftone dot area ratio that is a characteristic value of the image output device 12. A gradation correction process is performed, and an intermediate gradation process is performed so that the gradation reproduction processing unit 29 can separate the image data into pixels and reproduce each gradation. For example, the gradation reproduction processing unit 29 performs binarization processing or multi-value processing suitable for gradation reproducibility in the image output apparatus 12 on the region separated into the photographic regions by the region separation processing unit 24.

  The K image data processed by the gradation reproduction processing unit 29 is temporarily stored in a storage unit such as an image memory (not shown), and then read from the storage unit in accordance with the timing of image formation under the control of the CPU 1. And output to the image output device 12. The image output device 12 outputs given image data onto a recording paper such as paper, and is a printer such as an electrophotographic system or an inkjet system. For example, in the case of the electrophotographic system, the image output device 12 is a charger that charges the photosensitive drum to a predetermined potential, and emits a laser beam in accordance with the given image data to electrostatic latent image on the surface of the photosensitive drum. A laser writer for generating an image, a developer for supplying toner to the electrostatic latent image generated on the surface of the photosensitive drum to make it visible, and a transfer for transferring the toner image formed on the surface of the photosensitive drum onto paper Equipped with containers. Note that the image output device 12 may be a display device such as a display.

  In the image processing apparatus 20 shown in FIG. 1, the gray scale conversion processing unit 25 is arranged immediately after the region separation processing unit 24, but is not limited to this, and after the spatial filter processing unit 26. It may be arranged at another position, and may be arranged at any position as long as it is at least after the region separation processing unit 24 and before the gradation reproduction processing unit 29.

  Further, in the image processing apparatus 20 of FIG. 1, only functional blocks when the image forming apparatus performs monochrome output are illustrated, but the image forming apparatus also has a function of outputting an image in color. FIG. 2 is a block diagram illustrating a configuration of the image processing apparatus 20 when the image forming apparatus performs color output. When performing color output, the image processing apparatus 20 replaces the processing performed by the grayscale conversion processing unit 25 to the gradation reproduction processing unit 29 in FIG. 1 with a color correction unit 25a, a black generation and under color removal unit 25b, and a spatial filter. Processing is performed by the processing unit 26a, the scaling processing unit 27a, the output tone correction unit 28a, and the tone reproduction processing unit 29a.

  Even when color output is performed, the processes performed by the A / D conversion unit 21, the shading correction unit 22, and the input tone correction unit 23 of the image processing unit 20 are the same as those in the case of the monochrome output described above. The region separation processing unit 24 gives the region separation result to the black generation and under color removal unit 25b, the spatial filter processing unit 26a, and the gradation reproduction processing unit 29a, and the RGB image data given from the input gradation correction unit 23 is used as it is. The color correction unit 25a is supplied with the character feature value, halftone dot feature value, and photographic feature value of each pixel.

  The color correction unit 25 a converts the RGB image data given from the region separation processing unit 24 into image data in a color space of C (cyan), M (magenta), and Y (yellow), and at the image output device 12. In order to realize faithful color reproduction, color correction processing is performed on each color component of CMY. Specifically, the color correction processing is processing for removing color turbidity based on spectral characteristics of CMY toners or inks including unnecessary absorption components from the CMY color components. The color correction unit 25a gives the converted CMY image data to the black generation and under color removal unit 25b.

  The black generation and under color removal unit 25b calculates a color component of K (black) based on the region separation result given from the region separation processing unit 24 and each color component of the CMY image data given from the color correction unit 25a. A black generation process to be generated is performed, and a process of generating a new CMY component by subtracting the K component generated by the black generation process from each original CMY color component is performed. As a result, the CMY image data provided from the color correction unit 25a is converted into image data having four CMYK color components and is provided to the spatial filter processing unit 26a.

  The spatial filter processing unit 26a performs a spatial filter process using a digital filter on the CMYK image data provided from the black generation and under color removal unit 25b based on the result of the region separation provided from the region separation processing unit 24. As a result, the spatial frequency characteristics of the image are corrected, and it is possible to prevent the image output from the image output device 12 from being blurred or deteriorated in graininess. The CMYK image data subjected to the spatial filter processing by the spatial filter processing unit 26a is subjected to scaling processing by the scaling processing unit 27a, and various correction processing is performed by the output tone correction unit 28. Thereafter, it is given to the gradation reproduction processing unit 29a. The gradation reproduction processing unit 29a performs gradation reproduction processing on the CMYK image data provided from the output gradation correction unit 28a in accordance with the region separation result provided from the region separation processing unit 24.

  For example, a region separated into character regions by the region separation processing unit 24 is a high-frequency component enhancement amount in the sharp enhancement filter processing of the spatial filter processing unit 26a, particularly for the purpose of improving the reproducibility of black characters or color characters. Is set larger. Similarly, the gradation reproduction processing unit 29a performs a binarization process or a multi-value process suitable for reproducing the high frequency component of the image output device 12 for the purpose of improving the reproducibility of black characters or color characters. .

  For example, the region separated into halftone dot regions by the region separation processing unit 24 is subjected to low-pass filter processing for removing the input halftone dot component by the spatial filter processing unit 26a. Further, gradation correction processing for converting the density of CMYK image data into a halftone dot area ratio that is a characteristic value of the image output device 12 is performed by the output gradation correction unit 28a, and the gradation reproduction processing unit 29a performs this image data processing. Are divided into pixels and intermediate gradation processing is performed so that each gradation can be reproduced.

  For example, for the region separated into the photographic region by the region separation processing unit 24, the gradation reproduction processing unit 29a performs binarization processing or multi-value processing suitable for gradation reproduction in the image output apparatus 12. .

  As described above, the image forming apparatus can selectively perform both color output and monochrome output. The output method can be selected, for example, by switching the color mode or the monochrome mode when the user operates the operation unit 4. Alternatively, the image processing device 20 is provided with a color determination unit that automatically determines whether the image data input from the image input device 11 is a color image or a monochrome image, and the color mode is automatically set according to the determination result. Alternatively, the monochrome mode may be switched.

  In FIG. 1, only functional blocks related to monochrome output are illustrated in the image processing apparatus 20, and only functional blocks related to color output are illustrated in FIG. 2, but these functional blocks are actually image processing apparatuses. 20 is provided in parallel. At this time, the A / D conversion unit 21, the shading correction unit 22, the input tone correction unit 23, and the region separation processing unit 24 are function blocks common to the monochrome output and the color output, and the gray scale conversion processing unit 25 and the color correction unit 25a and a black generation and under color removal unit 25b are provided in parallel, and a spatial filter processing unit 26 (26a), a scaling processing unit 27 (27a), an output tone correction unit 28 (28a), and a tone reproduction processing unit 29 ( 29a) may be a common functional block.

  Next, details of processing performed by the region separation processing unit 24 and the gray scale conversion processing unit 25 will be described. FIG. 3 is a block diagram illustrating a configuration of the region separation processing unit 24. The region separation processing unit 24 includes a character feature amount calculation unit 41 that calculates a character feature amount from image data, a halftone feature amount calculation unit 42 that calculates a halftone feature amount, and a photographic feature amount calculation unit that calculates a photographic feature amount. 43. The RGB image data input from the input tone correction unit 23 to the region separation processing unit 24 is provided to the character feature amount calculation unit 41, the halftone feature amount calculation unit 42, and the photographic feature amount calculation unit 43, respectively. The data is directly output to the scale conversion processing unit 25.

The character feature amount calculation unit 41 of the region separation processing unit 24 performs a filtering process using an edge detection filter (Sobel filter) on the input image data, thereby obtaining a pixel value of each pixel of the image as a character feature amount. A gradient (edge degree) with the pixel values of the peripheral pixels is calculated. FIG. 4 is a schematic diagram for explaining a method of calculating the edge degree by the character feature amount calculation unit 41. The character feature amount calculation unit 41 calculates the edge degree for each pixel of the image data. The edge degree is calculated based on nine (3 × 3) pixels of the pixel of interest and the eight surrounding pixels. Do. FIG. 4A shows 3 × 3 image data including the pixel A ₅ as the target pixel and the surrounding pixels A _{1 to} A ₄ and pixels A _{6 to} A ₉ . Further, (b1) in FIG. 4 shows a vertical edge detection filter, and (b2) shows a horizontal edge detection filter. In FIG. 4 and the following description, the edge degree is detected with a size of 3 × 3. However, the present invention is not limited to this, and the edge degree is detected with a size suitable for the characteristics of the image forming apparatus. Can be done. The coefficient of the edge detection filter is an example and is not limited to this. For simplification of illustration, one numerical value is shown as the pixel value of each pixel. Actually, the pixel value of each pixel is composed of three values of RGB, and the following calculation is performed for each color. Shall be performed respectively.

  The character feature amount calculation unit 41 first extracts image data having a size of 3 × 3 as shown in FIG. 4A according to the target pixel of the input image data, and corresponds to each pixel value of the image data. The edge detection filter coefficients to be integrated are integrated, and the absolute value of the sum is calculated for each direction. In addition, the character feature amount calculation unit 41 adds the calculation results for each direction calculated using the two edge detection filters in FIGS. 4B1 and 4B2 to obtain the edge degree of the target pixel.

The calculation performed by the character feature amount calculation unit 41 is shown below.
α ₁ = | (−1) × A ₁ + 0 × A ₂ + 1 × A ₃
+ (− 1) × A ₄ + 0 × A ₅ + 1 × A ₆
+ (− 1) × A ₇ + 0 × A ₈ + 1 × A ₉ | (Formula 1-1)
α ₂ = | (−1) × A ₁ + (− 1) × A ₂ + (− 1) × A ₃
+ 0 × A ₄ + 0 × A ₅ + 0 × A ₆
+ 1 × A ₇ + 1 × A ₈ + 1 × A ₉ | (Formula 1-2)
(Edge degree) = α ₁ + α ₂ (Formula 1-3)

  The character feature amount calculation unit 41 provides the calculated edge degree to the gray scale conversion processing unit 25 as a character feature amount and also to the photograph feature amount calculation unit 43. Further, the character feature amount calculation unit 41 compares the edge degree calculated as the character feature amount with a predetermined threshold value, and determines whether or not the target pixel is a pixel included in the character region. The character feature amount calculation unit 41 determines that the target pixel is included in the character region when the edge degree exceeds the threshold value, and outputs a character region signal indicating the determination result as one of the region separation results. For example, the character region signal is a 1-bit digital signal, and when it is determined that the pixel of interest is included in the character region, the character region signal is set to “H” and the pixel of interest is not included in the character region. If it is determined, the character area signal is set to “L”.

  FIG. 5 is a flowchart showing the procedure of the character feature amount calculation process performed by the character feature amount calculation unit 41. First, the character feature amount calculation unit 41 uses the image data input from the input tone correction unit 23 as 3 × 3 image data of the target pixel to be processed and its surrounding pixels (FIG. 4A). Reference) is acquired (step S1). Next, the character feature amount calculation unit 41 performs a filtering process using a vertical edge detection filter (see FIG. 4B1) (step S2) and a horizontal edge detection filter (see FIG. 4B2). Is used (step S3), and the edge degree is calculated by adding the values obtained by the respective filter processes (step S4).

  Next, the character feature amount calculation unit 41 checks whether or not the calculated edge degree exceeds a predetermined threshold (step S5), and determines whether or not the target pixel is included in the character region. When the edge degree exceeds the threshold value (S5: YES), the character feature amount calculation unit 41 sets the character area signal to “H” and outputs it (step S6), and when the edge degree does not exceed the threshold value (step S6). (S5: NO), the character area signal is set to "L" and output (step S7). After outputting the character area signal in step S6 or S7, the character feature amount calculation unit 41 checks whether or not the calculation of the edge degree has been completed for all the pixels of the image data (step S8), and ends for all the pixels. If not (S8: NO), the process returns to step S1 to continue calculating the edge degree of the remaining pixels. If all the pixels are finished (S8: YES), the character feature amount calculation process is performed. finish.

The halftone dot feature amount calculation unit 42 of the region separation processing unit 24 performs processing for calculating the degree of change (complexity) of the pixel values of adjacent pixels as the halftone dot feature amount with respect to the input image data. FIG. 6 is a schematic diagram for explaining a calculation method of the complexity by the halftone dot feature amount calculation unit 42. The halftone dot feature amount calculation unit 42 calculates the complexity for each pixel of the image data. The complexity is calculated based on the pixel of interest, the eight surrounding pixels, and the surrounding pixels. (A) in the figure shows 3 × 3 image data including the pixel A ₅ as the target pixel and the surrounding pixels A _{1 to} A ₄ and pixels A _{6 to} A ₉ . The halftone dot feature amount calculation unit 42 is in the right direction (see FIG. 6 (c1)), the down direction (see FIG. 6 (c2)), the lower right direction (see FIG. 6 (c3)), and the lower left direction (see FIG. 6). 6 (see c4)), the sum of absolute values of differences is calculated from the pixel values of adjacent pixels, and the four sums calculated for the four directions including the main scanning direction (right direction) are calculated. The result of the addition is taken as the complexity. In FIG. 6 and the following description, the complexity is detected with a size of 3 × 3. However, the present invention is not limited to this, and the complexity is detected with a size suitable for the characteristics of the image forming apparatus. Can be done. Also, the detection direction of the complexity level is an example, and the present invention is not limited to this, and the complexity level may be detected in other directions. However, it is desirable that the detection direction includes the main scanning direction (right direction in this example) of the image. For simplification of illustration, one numerical value is shown as the pixel value of each pixel. Actually, the pixel value of each pixel is composed of three values of RGB, and the following calculation is performed for each color. Shall be performed respectively.

For example, in the case of FIG. 6 (c1), the halftone feature quantity calculating unit 42 calculates the sum beta ₁ carries out an operation following.
β ₁ = | A ₂ −A ₁ | + | A ₃ −A ₂ | + | B ₁ −A ₃ |
+ | A ₅ −A ₄ | + | A ₆ −A ₅ | + | B ₂ −A ₆ |
+ | A ₈ -A ₇ | + | A ₉ -A ₈ | + | B ₃ -A ₉ | (Formula 2-1)
Similarly, in the case of FIG. 6C2, the halftone dot feature amount calculation unit 42 calculates the sum β ₂ by performing the following calculation.
β ₂ = | A ₄ −A ₁ | + | A ₅ −A ₂ | + | A ₆ −A ₃ |
+ | A ₇ -A ₄ | + | A ₈ -A ₅ | + | A ₉ -A ₆ |
+ | C ₁ -A ₇ | + | C ₂ -A ₈ | + | C ₃ -A ₉ | (Formula 2-2)
Similarly, in the case of FIG. 6 (c3), the halftone dot feature value calculation unit 42 calculates the total β ₃ by performing the following calculation.
β ₃ = | A ₅ -A ₁ | + | A ₆ -A ₂ | + | D ₁ -A ₃ |
+ | A ₈ -A ₄ | + | A ₉ -A ₅ | + | D ₂ -A ₆ |
+ | D ₃ -A ₇ | + | D ₄ -A ₈ | + | D ₅ -A ₉ | (Formula 2-3)
Similarly, in the case of FIG. 6 (c4), the halftone dot feature amount calculation unit 42 calculates the total β ₄ by performing the following calculation.
β ₄ = | E ₁ −A ₁ | + | A ₄ −A ₂ | + | A ₅ −A ₃ |
+ | E ₂ −A ₄ | + | A ₇ −A ₅ | + | A ₈ −A ₆ |
+ | E ₃ −A ₇ | + | E ₄ −A ₈ | + | E ₅ −A ₉ | (Formula 2-4)
Therefore, the complexity that the halftone dot feature amount calculation unit 42 calculates as the halftone dot feature amount is as follows.
(Complexity) = β ₁ + β ₂ + β ₃ + β ₄ (Formula 2-5)
Note that there may be cases where the difference cannot be calculated because there are no adjacent pixels at the right end, left end, and lower end of the image. For such a part for which the difference cannot be calculated, the difference is set to “0”, or the difference is calculated by regarding the pixel value of the nonexistent pixel as “0”, and each obtained difference May be calculated.

  The halftone dot feature amount calculation unit 42 gives the calculated complexity to the grayscale conversion processing unit 25 as a halftone dot feature amount. Further, the halftone dot feature amount calculation unit 42 compares the complexity calculated as the halftone dot feature amount with a predetermined threshold value, and determines whether or not the target pixel is a pixel included in the character area. . The halftone dot feature amount calculation unit 42 determines that the target pixel is included in the halftone dot region when the complexity exceeds a threshold value, and outputs a halftone dot region signal indicating the determination result as one of the region separation results. For example, the halftone area signal is a 1-bit digital signal, and when it is determined that the target pixel is included in the halftone area, the halftone area signal is set to “H”, and the target pixel is set in the halftone area. If it is determined not to be included, the dot area signal is set to “L”.

  FIG. 7 is a flowchart showing a procedure of halftone dot feature amount calculation processing performed by the halftone dot feature amount calculation unit 42. First, the halftone dot feature value calculation unit 42 obtains 3 × 3 image data of the target pixel to be processed and its surrounding pixels from the image data input from the input tone correction unit 23 (FIG. 6A). And the image data relating to the surrounding pixels are acquired (step S21). Next, the halftone dot feature amount calculation unit 42 calculates the sum of the absolute values of the pixel value differences between the pixels of the 3 × 3 image data and the pixels adjacent in the right direction (see step S22, FIG. 6 (c1)). ), The sum of absolute values of pixel value differences with pixels adjacent in the downward direction is calculated (see step S23, FIG. 6 (c2)), and the absolute value of the pixel value difference with pixels adjacent in the diagonally lower right direction is calculated. The sum of the values is calculated (step S24, see FIG. 6 (c3)), and the sum of the absolute values of the difference between the pixel values of the pixels adjacent in the diagonally lower left direction is calculated (step S25, see FIG. 6 (c4)). ).

  Next, the halftone dot feature value calculation unit 42 adds the four calculated sums to calculate the complexity (step S26), and checks whether the complexity exceeds a predetermined threshold (step S27). Then, it is determined whether or not the target pixel is included in the halftone dot region. When the complexity level exceeds the threshold value (S27: YES), the halftone dot feature amount calculation unit 42 sets the halftone area signal to “H” and outputs it (step S28), and the complexity level does not exceed the threshold value. In the case (S27: NO), the dot area signal is set to “L” and output (step S29). After outputting the halftone dot region signal in step S28 or S29, the halftone dot feature value calculation unit 42 checks whether or not the complexity calculation has been completed for all the pixels of the image data (step S30). If not completed (S30: NO), the process returns to step S21 to continue calculating the complexity of the remaining pixels. If all pixels are completed (S30: YES), the halftone dot feature amount The calculation process ends.

  The photo feature amount calculation unit 43 of the region separation processing unit 24 is based on the RGB image data input from the input tone correction unit 23 and the edge degree input as the character feature amount from the character feature amount calculation unit 41. The entropy value with respect to the appearance frequency of the pixel value obtained for each predetermined area of the image is calculated as a photographic feature amount. 8 and 9 are schematic diagrams for explaining the entropy value calculation method by the photographic feature value calculation unit 43. FIG. 8A shows a part of the image data, and FIG. 8D shows the edge degree with respect to this. The photograph feature quantity calculation unit 43 first performs downsampling processing on the given image data and edge degree data to reduce the data quantity.

  The downsampling process is a process of calculating a representative value from data in a predetermined range such as 4 × 4 or 8 × 8, and setting the representative value as a value after downsampling. The data amount is reduced to 1/16 or 1 / It can be reduced to 64 or the like. The representative value calculation method is, for example, a method of selecting any one pixel value in a predetermined range as a representative value, a method of using an average value, total value, maximum value, minimum value, or the like of data in a predetermined range as a representative value, etc. Various methods can be used. In the present embodiment, the average value of data in a predetermined range is calculated as a representative value for the downsampling processing of image data, and the total value of data in the predetermined range is used as the representative value for the downsampling processing of edge degree data. It shall be calculated.

As an example, FIG. 8 shows an example in which image data and edge degree data are down-sampled in a predetermined range of 2 × 2. In this case, the average value a ₁ = (A ₁ + A ₂ + A ₃ + A ₄ ) / 4 is calculated for the four pixels A _{1 to} A ₄ of the original image data (see FIG. 8A), and after downsampling. Image data (see FIG. 8A '). Further, the total value b ₁ = B ₁ + B ₂ + B ₃ + B ₄ is calculated for the _four edge degrees B _{1 to} B _{4 in} the original edge degree data (see FIG. 8D), and the edge after downsampling is calculated. As data of the degree (see FIG. 8 (d ′)). The photograph feature quantity calculation unit 43 performs the same downsampling process on all data.

  After the end of the downsampling process, the photographic feature quantity calculation unit 43 performs a conditional local histogram generation process based on the calculated data. This conditional local histogram extracts, for example, only image data having a corresponding downsampled edge degree equal to or less than a predetermined threshold from a predetermined range of image data such as 11 × 11 for the downsampled image data. FIG. 5 is a graph showing the correspondence between pixel values and frequencies of the image data that has been obtained. By extracting only the image data whose edge degree is less than or equal to the threshold value, it is possible to exclude data that causes a sharp change in density at the edge, such as a character image, from the generation target of the histogram. The feature amount of the photographic image can be calculated with high accuracy by suppressing the deformation of the image.

  Next, the photographic feature amount calculation unit 43 performs entropy value calculation processing for the obtained conditional local histogram. The entropy value can be calculated by the following calculation formula. In the following calculation formula, “i” is the pixel value of the downsampled image data (0 to 255 (in the case of 8-bit data)), and “HIST (i)” is the frequency of the pixel value i. , “NUM” is the total number of pixels included in the histogram.

  The entropy value calculated by the photograph feature value calculation unit 43 according to the above (Equation 3) is a quantitative representation of the distribution shape of the histogram, and the value increases as the distribution of the histogram gently increases. The sharper the value, the smaller the value. In many cases, the distribution of pixel values of a photographic image is relatively wide. Therefore, the larger the entropy value calculated by the photographic feature amount calculation unit 43, the higher the possibility that the photographic image is a photographic image. Therefore, the photographic feature value calculation unit 43 provides the calculated entropy value to the gray scale conversion processing unit 25 as a photographic feature value.

As an example, FIG. 9 shows a conditional local histogram generated for a 3 × 3 region of image data a _{1 to} a ₉ and edge degree data b _{1 to} b ₉ after downsampling processing. FIG. 9A shows an example of numerical values applied to the image data a _{1 to} a ₉ after downsampling processing (see FIG. 9A ′), and the edge degree data b ₁ to b1 after downsampling processing FIG. 9 (d ″) is an example of numerical values applied to b ₉ (see FIG. 9 (d ′)).

The photo feature quantity calculation unit 43 excludes, for example, image data having an edge degree of “7” or more from the image data of the 3 × 3 region after downsampling processing, and only the image data having an edge degree of “6” or less Are extracted as histogram generation targets. In the illustrated example, the image data a ₁ is excluded (see the hatched pixels in FIG. 9A). The photographic feature amount calculation unit 43 calculates the frequency for each pixel value of the extracted eight image data. The histogram is generated by checking (see FIG. 9E).

Next, the photographic feature value calculation unit 43 calculates the entropy value of the histogram using the above-described (Equation 3) based on the generated histogram. When (Equation 3) is applied to the histogram of FIG.
(Entropy value) =-(1/8) log ₂ (1/8)-(3/8) log ₂ (3/8)-(1/8) log ₂ (1/8)-(2/8) log ₂ (2/8) − (1/8) log ₂ (1/8)
= 2.2
Incidentally, the entropy value = 2.2 is the entropy value calculated by focusing on a ₅ in the image data after the down-sampling shown in FIG. 9 (a '), photo feature quantity calculating unit 43 is a ₅ The entropy value of the four pixels of the original image data used for the calculation of is treated as 2.2. In addition, the photographic feature amount calculation unit 43 may average the entropy values for the calculated pixels in a predetermined area such as 3 × 3, and may use the averaged entropy value as the final photographic feature amount.

  The photographic feature amount calculation unit 43 gives the calculated entropy value to the gray scale conversion processing unit 25 as a photographic feature amount. The photo feature amount calculation unit 43 compares the entropy value calculated as the photo feature amount with a predetermined threshold value, and determines whether each pixel of the image data is a pixel included in the photo area. To do. When the entropy value exceeds the threshold value, the photo feature amount calculation unit 43 determines that the pixel of interest is included in the photo region, and outputs a photo region signal indicating the determination result as one of the region separation results. For example, the photographic area signal is a 1-bit digital signal. If it is determined that the target pixel is included in the photographic area, the photographic area signal is set to “H” and the target pixel is not included in the photographic area. If it is determined, the photo area signal is set to “L”.

  FIG. 10 is a flowchart showing the procedure of the photographic feature value calculation process performed by the photographic feature value calculation unit 43. First, the photograph feature value calculation unit 43 acquires image data input from the input tone correction unit 23 (step S41), and acquires edge degree data input from the character feature value calculation unit 41 (step S41). S42). Next, the photographic feature value calculation unit 43 performs a downsampling process on the acquired image data and edge degree data (step S43) to reduce the data amount.

  After the downsampling process is completed, the photographic feature amount calculation unit 43 extracts the image data with the edge degree equal to or less than the threshold value based on the downsampled image data and the edge degree data within a predetermined range. A local histogram is generated (step S44). The photograph feature quantity calculation unit 43 calculates the entropy value by the above calculation formula based on the generated local histogram (step S45). At this time, the photograph feature value calculation unit 43 may average the calculated entropy values in a predetermined region.

  Next, the photograph feature value calculation unit 43 checks whether or not the calculated entropy value exceeds a predetermined threshold (step S46), and determines whether or not the target pixel is included in the photograph area. When the entropy value exceeds the threshold value (S46: YES), the photo feature value calculation unit 43 sets the photo area signal to “H” and outputs it (step S47), and when the entropy value does not exceed the threshold value (step S47). (S46: NO), the photographic area signal is set to "L" and output (step S48). After outputting the photographic area signal in step S47 or S48, the photographic feature amount calculation unit 43 checks whether or not the entropy value calculation has been completed for all the pixels of the image data (step S49), and ends for all the pixels. If not (S49: NO), the process returns to step S41 to continue calculating entropy values for the remaining pixels. If all pixels have been completed (S49: YES), the photographic feature value calculation process is performed. finish.

Gray scale conversion in which RGB image data, edge degree calculated as a character feature amount, complexity calculated as a halftone dot feature amount, and entropy value calculated as a photo feature amount are given from the region separation processing unit 24 The processing unit 25 performs processing for detecting a halftone dot background region from the image data based on each feature amount, corrects the density of the pixels detected as the halftone dot background region, and converts the corrected color image from white. Conversion to a monochrome image (grayscale image) expressed in light and dark up to black is performed. That is, the gray scale conversion processing unit 25 operates as a character feature amount determination unit, a halftone dot feature amount determination unit, and a photographic feature amount determination unit that respectively determine whether each feature amount exceeds a predetermined value (threshold). It operates as a halftone dot region detection unit that detects a halftone dot region based on a combination of determination results for each feature amount, and further corrects the density of a color image according to the detected density of the halftone dot region. It operates as an image correction means for performing image correction such as. The grayscale conversion processing unit 25 compares the edge degree, complexity, and entropy value calculated as the feature amount for each pixel of the image data with predetermined threshold values, and the feature amount of each pixel is a predetermined value shown below. When the condition is satisfied (that is, when the determination result based on the threshold value of each feature amount is the following combination), this pixel is set as a pixel of the halftone dot background region.
(Edge degree <threshold A) and (complexity> threshold B) and (entropy value <threshold C) (Expression 4)

  Here, it is assumed that thresholds A, B, and C are appropriately determined in advance at the design stage of the image forming apparatus. The threshold A is a threshold for determining whether each pixel is a pixel in the character area, and the threshold B is a threshold for determining whether each pixel is a pixel in the halftone area. The threshold value C is a threshold value for determining whether each pixel is a pixel in the photographic area. Therefore, in other words, the above condition is a condition for detecting not a character area in image data but a halftone dot area and a non-photograph area as a halftone dot background area.

  The gray scale conversion processing unit 25 detects a pixel satisfying the above conditions as a halftone dot background region, corrects the density of the pixel detected as the halftone dot background region (details will be described later), and then converts the gray color from the color image. Convert to scale image. The conversion from a color image to a grayscale image includes, for example, a method of calculating a luminance value from RGB pixel values, a method of calculating a weighted average of RGB pixel values, or a method of using a G component among RGB pixel values. Various methods can be used.

  FIG. 11 is a flowchart illustrating a procedure of processing performed by the grayscale conversion processing unit 25. The gray scale conversion processing unit 25 first acquires the image data input from the region separation processing unit 24 (step S61), and acquires three feature amounts, that is, a character feature amount, a halftone dot feature amount, and a photographic feature amount. (Step S62). Next, the grayscale conversion processing unit 25 performs a halftone dot region detection process for detecting a halftone dot region from the image data using the acquired image data and feature amount (step S63).

  FIG. 12 is a flowchart illustrating a procedure of halftone dot background area detection processing performed by the grayscale conversion processing unit 25, and is processing performed in step S63 of FIG. Note that the three threshold values A, B, and C used in the halftone dot background area detection process shown in FIG. 12 are determined in advance at the design stage or manufacturing stage of the image forming apparatus based on the hardware characteristics of the image forming apparatus. An appropriate value is set. In the halftone dot background area detection process, the gray scale conversion processing unit 25 first checks whether or not the edge degree acquired as the character feature amount of the target pixel is smaller than a predetermined threshold A (step S81). If it is equal to or greater than the threshold A (S81: NO), it is determined that the pixel of interest is not a pixel in the halftone dot background region but a pixel in another region (step S85).

  When the edge degree is smaller than the threshold A (S81: YES), the gray scale conversion processing unit 25 checks whether or not the complexity acquired as the halftone dot feature amount of the target pixel is larger than a predetermined threshold B (Step S81). S82). When the complexity is equal to or less than the threshold value B (S82: NO), the grayscale conversion processing unit 25 determines that the target pixel is not a pixel in the halftone dot background region but a pixel in another region (step S85).

  When the complexity is larger than the threshold B (S82: YES), the grayscale conversion processing unit 25 checks whether or not the entropy value acquired as the photographic feature amount of the target pixel is smaller than the threshold C (step S83). When the entropy value is greater than or equal to the threshold value C (S83: NO), the grayscale conversion processing unit 25 determines that the pixel of interest is not a pixel in the halftone dot background region but a pixel in another region (step S85). When the entropy value is smaller than the threshold value C (S83: YES), the grayscale conversion processing unit 25 determines that the pixel of interest is a pixel in the halftone dot background region (step S84). The grayscale conversion processing unit 25 that has performed the determination in step S84 or S85 ends the halftone dot background area detection process, and returns to the process of the flowchart illustrated in FIG.

  After the halftone dot area detection processing is completed, the grayscale conversion processing unit 25 checks whether the target pixel is a pixel in the halftone dot area based on the detection result (step S64). When the target pixel is a pixel in the halftone dot background region (S64: YES), the gray scale conversion processing unit 25 performs density correction of the target pixel (step S65), and converts the color target pixel into gray scale (step S65). Step S66). If the target pixel is not a halftone dot background pixel (S64: NO), the grayscale conversion processing unit 25 converts the color target pixel to grayscale without performing density correction (step S66).

  After the conversion to the gray scale is completed, the gray scale conversion processing unit 25 checks whether or not the gray scale conversion processing has been completed for all the pixels of the input image data (step S67), and ends the processing for all the pixels. If not (S67: NO), the process returns to step S61 and the above process is repeated. When the process for all the pixels is completed (S67: YES), the grayscale conversion processing unit 25 ends the process.

  The grayscale image data obtained by the conversion processing of the grayscale conversion processing unit 25, that is, the K component image data, is supplied to the spatial filter processing unit 26. Thereafter, the K component image data is subjected to appropriate image processing by the spatial filter processing unit 26, scaling processing unit 27, output tone correction unit 28, and tone reproduction processing unit 29, and is output by the image output device 12. Is output.

Next, the density correction process for the pixels in the halftone dot background area performed by the gray scale conversion processing unit 25 will be described. The gray scale conversion processing unit 25 checks whether or not the difference between the density of the halftone background area and the density of the characters on the halftone background area is small, that is, whether or not the density difference is within a predetermined range. When the difference is within the predetermined range, the density of the pixel in the halftone dot background region is corrected according to the following two correction equations. In the following correction formula, X _in is the density value of the pixel before correction, and X _out is the density value of the pixel after correction. Further, the density value of the pixel is assumed to be in the range of 0 to 255, the value 0 being the highest density (dark), and the value 255 being the lowest density (light).
(Density correction formula 1)
X _out = X _in +50 (X _in ≦ 100)
X _out = 150 + (X _in −100) / 3 (100 <X _in ≦ 175)
X _out = X _in (X _in > 175)
(Density correction formula 2)
X _out = X _in (X _in ≦ 120)
X _out = 120 + (X _in −120) / 3 (120 <X _in ≦ 220)
X _out = 255− (255−X _in ) × 3 (X _in > 220)
The above density correction formula 1 is used when the density of the halftone dot background region is higher than a predetermined density. The density correction formula 2 is used when the density of the halftone dot background region is lower than a predetermined density. The above-mentioned predetermined density, which is a criterion for determining whether to use the density correction formula 1 or the density correction formula 2, is set to an appropriate value in advance at the design stage of the image forming apparatus. It may be configured to calculate for each image data from the average density of the entire input image data. The constants included in the density correction formula 1 and the density correction formula 2 are merely examples, and the present invention is not limited to this.

  13 and 14 are schematic diagrams for explaining the density correction formula 1. FIG. 13 is a graph showing the density correction formula 1, and FIG. 14 shows the calculation results using the density correction formula 1. It is shown as a table. If the density of the halftone dot background region is higher than the predetermined density and the correction for increasing the density of the halftone dot background region is performed, the visibility of characters on the halftone dot background region may be reduced. Therefore, the gray scale conversion processing unit 25 performs correction for reducing the density of pixels included in the halftone dot background region having a density higher than the predetermined density. By using the density correction formula 1, it is possible to greatly reduce the density of pixels having a high density of 0 to 100. In addition, with respect to pixels having medium densities of values 101 to 175, the higher the density, the larger the density can be reduced.

  15 and 16 are schematic diagrams for explaining the density correction formula 2. FIG. 15 is a graph showing the density correction formula 2, and FIG. 16 shows the calculation results using the density correction formula 2. It is shown as a table. When the density of the halftone dot background region is lower than the predetermined density, if the correction for reducing the density of the halftone dot background region is performed, the density may be reduced to the extent that the halftone dot background region cannot be visually recognized. Therefore, the gray scale conversion processing unit 25 performs correction for increasing the density of pixels included in the halftone dot background region having a density lower than the predetermined density. By using the density correction formula 2, it is possible to increase the density of the pixel having a relatively low density of 120 to 255 with the correction amount for the pixel of the density 220 as a peak.

The gray scale conversion processing unit 25 may perform the calculation using the density correction formula 1 or the density correction formula 2 every time density correction is performed. Only the values of _in and X _out are stored in association with each other as a table, and the corrected value is obtained by referring to the stored table at the time of density correction. As a result, the density correction processing by the gray scale conversion processing unit 25 can be speeded up.

  As described above, the gray scale conversion processing unit 25 determines the density of the halftone dot background area and the density on the halftone dot background area in order to determine the density correction area using the density correction formula 1 or the density correction formula 2. It is necessary to check whether the difference from the character density is within a predetermined range. Further, the gray scale conversion processing unit 25 determines whether to use either the density correction formula 1 or the density correction formula 2 so that the density of the halftone dot background region that is the target of density correction is higher than a predetermined density. It is necessary to check whether or not. Therefore, the gray scale conversion processing unit 25 calculates the average density of the halftone dot area in the vicinity of the character area in advance based on the edge degree given as the feature quantity of the character area and the complexity given as the feature quantity of the halftone dot area. The calculation and the comparison with the character density in advance are performed before (or in parallel with) the halftone dot background area detection process and the correction process. Hereinafter, the density correction process including the preceding process performed by the gray scale conversion processing unit 25 will be described.

  FIGS. 17 and 18 are schematic diagrams for explaining the density correction processing performed by the grayscale conversion processing unit 25. As an example, the case where density correction processing is performed on image data of 7 pixels × 5 pixels ( A) to (f) are shown in time series. First, the gray scale conversion processing unit 25 determines a pixel in the halftone dot region where the complexity given as the feature amount of the halftone dot region exceeds a predetermined threshold B (which may be the same as the threshold B shown in FIG. 12). Extract as In FIG. 17A, the extracted pixels in the halftone dot region are shown hatched. Further, the gray scale conversion processing unit 25 extracts a pixel whose edge degree given as a feature amount of the character area exceeds a predetermined threshold A (which may be the same as the threshold A shown in FIG. 12) as a character area pixel. To do. In FIG. 17B, the extracted pixels in the character area are shown with hatching.

Next, the gray scale conversion processing unit 25 performs labeling on the pixels extracted as the halftone dot area in the following procedure.
(1) The initial value of the label is “1”.
(2) In order from the upper left of the image data, pixels in an unprocessed halftone dot region are set as target pixels to be labeled.
(3) It is checked whether or not the upper left, upper, upper right, or left pixel is labeled with respect to the target pixel. If the label is attached, this label is set as the target pixel label. However, when there are a plurality of different labels, the minimum label is set as the label of the pixel of interest, and the plurality of labels are associated and stored as the same label.
(4) When the upper left, upper, upper right, or left pixel is not labeled with respect to the target pixel, an unadded label is attached to the target pixel. Thereafter, the label value is incremented by one.
(5) Repeat (2) to (4) for all pixels in the halftone dot region of the image data.

  A halftone dot region labeled according to the above procedure is shown in FIG. In the illustrated example, the halftone dot region is separated into two regions, a halftone dot region labeled 1 and a halftone dot region labeled 2 (and the label 3 regarded as the same label). Yes. That is, by performing labeling according to the above procedure, the grayscale conversion processing unit 25 can group pixels in continuous or close halftone dot regions.

  Next, the grayscale conversion processing unit 25 uses the same label (including the label considered to be the same) for the halftone dot region including the pixel determined to be the character region or the halftone dot region near the pixel determined to be the character region. ) The average value of the pixel densities is calculated. In the example of FIG. 17C, the grayscale conversion processing unit 25 calculates the average density of the eight pixels labeled 2 and 3. Thereafter, the gray scale conversion processing unit 25 acquires the density of the pixels in the character area and compares it with the calculated average density of the halftone dot area, and the difference between the density of the character and the halftone dot area is within a predetermined range. Check if there is any. When there are a plurality of pixels in the corresponding character area and a plurality of densities, the difference between the density of the character and the density of the halftone dot area is the smallest difference from the average density of the halftone dot area. What is necessary is just to check whether it is in a predetermined range. The calculation of the average density for the halftone dot region of the same label may be performed in parallel with the labeling process.

  In the illustrated example, it is assumed that the difference between the density of the character and the density of the halftone dot area is within a predetermined range, and FIG. 18D shows hatched pixels in the halftone dot area where the density difference from the character is small. . Further, the gray scale conversion processing unit 25 detects pixels in the halftone dot background region by the halftone dot region detection process shown in FIG. 12, and an example of the detection result is hatched in the pixels in the halftone dot region. This is shown in FIG. From these results, the gray scale conversion processing unit 25 is a halftone dot region (see FIG. 18D) with a small density difference from the character and a halftone dot background region (see FIG. 18E). The region is set as a region for density correction (see FIG. 18F).

  The gray scale conversion processing unit 25 calculates the average density of the obtained area, and when the average density is higher than the predetermined density, performs density correction using the above density correction formula 1 on the area, thereby obtaining the average density. If the density is lower than the predetermined density, density correction using density correction formula 2 is performed on this area (corresponding to steps S64 and S65 in the flowchart shown in FIG. 11). After the density correction, the gray scale conversion processing unit 25 converts the image data into a gray scale and outputs it.

  Through the above processing, the feature quantities of the character area, halftone dot area, and photograph area are calculated from the input image data, the halftone dot background area is detected based on these feature quantities, and the pixel density of the halftone dot background area Correction can be performed. In the following, these series of processes will be described with reference to actual examples. FIG. 19 is a schematic diagram illustrating an example of image data. (A) in the figure is image data obtained by reading the document by the image input device 11 and is input from the image input device 11 to the image processing device 20. The image data of this example is composed of a character 51 and a halftone dot photo 52 having no background, and a character 54 and a halftone dot photo 55 having a background 53 by a halftone dot. Also, (b) to (d) in the figure show the result of region separation by the region separation processing unit 24, and (b) is a character region output by the character feature value calculation unit 41 of the region separation processing unit 24. (C) corresponds to a halftone dot region signal output from the halftone dot feature amount calculation unit 42, and (d) corresponds to a photographic region signal output from the photographic feature amount calculation unit 43. However, the region separation results shown in (b) to (d) are ideal results and may actually include some errors.

  FIG. 20 is an enlarged view of the periphery of the area A in the image data shown in FIG. 19A, and the pixel value (density value) of each pixel in the 12 × 12 size image data is indicated by a numerical value from 0 to 255. It is. However, the actual image data has values for each of the three colors RGB, but in order to simplify the description, in the following, only the G component will be illustrated and described, and the R component will be described. The illustration and description of the components B and B are omitted. The R component and the B component may be processed in the same manner as the G component, except that the determination threshold value is different. In the following description, various types of processing will be described on the assumption that the 12 × 12 size data shown in FIG. 20 is one image data given from the image input device 11.

  First, the image processing apparatus 20 calculates a character feature amount from the input image data. FIG. 21 is a schematic diagram showing the result of calculating the edge degree as the character feature amount from the image data shown in FIG. The edge degree is calculated by the character feature amount calculating unit 41 of the region separation processing unit 24 using the two edge detection filters shown in FIGS. 4B1 and 4B2. The character feature amount calculation unit 41 performs the above-described calculation (Equation 1-1) using the target pixel and the surrounding pixels and the vertical edge detection filter in FIG. calculate. In addition, the character feature amount calculation unit 41 performs the calculation of (Equation 1-2) by using the target pixel and the surrounding pixels and the horizontal edge detection filter of FIG. calculate. After that, the character feature amount calculation unit 41 uses the value obtained by adding the vertical edge degree and the horizontal edge degree according to (Equation 1-3) as the final edge degree of the target pixel.

  Since the two edge detection filters are each a 3 × 3 size filter, calculation using the edge detection filter cannot be performed on pixels existing in the upper, lower, left, and right ends of the image data. Therefore, in this example, without using the edge detection filter for the upper, lower, left, and right end pixels, the edge degree of the adjacent pixel is copied to obtain the edge degree of the end pixel (in FIG. 21, the edge degree is expressed as the edge degree). The copied pixel is surrounded by a broken line). However, the method of determining the edge degree of the end pixel is not limited to this, and the calculation by the edge detection filter may be performed by regarding the value of the pixel that is insufficient with respect to the 3 × 3 size as 0.

  The character feature amount calculation unit 41 gives the calculated edge degree to the gray scale conversion processing unit 25 and the photograph feature amount calculation unit 43 as a character feature amount. Further, the character feature amount calculation unit 41 determines the character region by comparing the calculated edge degree with a predetermined threshold value. In this example, the threshold value is set to 500, and the pixel area is hatched to indicate a character area. The character feature amount calculation unit 41 outputs the comparison result as a character area signal.

  Further, the image processing apparatus 20 calculates a halftone dot feature amount from the input image data. FIG. 22 is a schematic diagram showing a result of calculating the complexity as the halftone dot feature amount from the image data shown in FIG. The halftone dot feature amount calculation unit 42 uses the target pixel, its surrounding pixels, and further its surrounding pixels, and calculates the absolute difference between the values of adjacent pixels in each direction shown in FIGS. 6 (c1) to (c4). The sum of the values is calculated based on (Expression 2-1) to (Expression 2-4), and the four sums calculated for each direction are added based on (Expression 2-5). Calculate the degree.

Note that in order to calculate the complexity, the pixel of interest and its surrounding pixels and further surrounding pixels are required, so the pixels existing near the top, bottom, left, and right edges of the image data are represented by (Equation 2-1) ~ (Equation 2-4) may not be calculated. Therefore, in this example, when there are no surrounding pixels (A _{1 to} A ₄ and A _{6 to} A ₉ ) with respect to the target pixel (A _{5 in} FIG. 6), the calculation result of the difference relating to the non-existing pixel Is set to 0, and when the surrounding pixels (B _{1 to} B ₃ , C _{1 to} C ₃ , D _{1 to} D ₅ , E _{1 to} E ₅ ) do not exist, the value of the non-existing pixel is set to 0. The complexity is calculated by calculating the difference and obtaining the sum of the absolute values of the differences. However, the method for determining the complexity of the pixels near the edge is not limited to this.

  The halftone dot feature amount calculation unit 42 gives the calculated complexity to the gray scale conversion processing unit 25 as a halftone dot feature amount. The halftone dot feature value calculation unit 42 determines the halftone dot region by comparing the calculated complexity with a predetermined threshold value. In this example, the threshold value is set to 2450, and in FIG. 22, a halftone dot region is shown by hatching pixels whose complexity exceeds 2450. The halftone dot feature value calculation unit 42 outputs the comparison result as a halftone dot region signal.

  The image processing apparatus 20 calculates an entropy value as a photographic feature amount from the input image data. FIG. 23 is a schematic diagram showing a calculation process for calculating a photographic feature amount from the image data shown in FIG. FIG. 24 is a schematic diagram showing a result of calculating entropy as a photographic feature amount from the image data shown in FIG. FIG. 25 is a schematic diagram showing a halftone dot region in the image data shown in FIG. The photograph feature quantity calculation unit 43 first down-samples the input image data and the edge quantity calculated by the character feature quantity calculation unit 41, respectively. In this example, the photographic feature amount calculation unit 43 calculates the average pixel value of four pixels for each 2 × 2 size and sets it as a representative value, thereby converting the image data of 12 × 12 size to 6 × 6. Downsampling to a size (see FIG. 23A). Further, the photographic feature amount calculation unit 43 calculates the total amount of the four edge amounts for each 2 × 2 size and sets the edge amount calculated for each pixel of the image data as a representative value, thereby obtaining a 12 × 12 size. Are down-sampled to 6 × 6 size (see FIG. 23B).

  Next, the photographic feature amount calculation unit 43 determines whether or not the down-sampled edge amount exceeds a predetermined threshold value. In this example, this threshold is set to 1000, and in FIG. 23B, the edge amount exceeding the threshold is hatched. Thereafter, the photographic feature amount calculation unit 43 excludes the data corresponding to the edge amount exceeding the threshold (shown by hatching in FIG. 23A) from the downsampled image data, and removes the remaining data. The entropy value is calculated based on this.

  The photographic feature quantity calculation unit 43 creates a histogram for each 3 × 3 region of the pixel of interest and surrounding pixels of the downsampled image data (illustration of the histogram is omitted). An entropy value based on Equation 3) is calculated. FIG. 24A shows the result of calculating the entropy value for each pixel of the image data after downsampling. Since the histogram is created for each 3 × 3 area, it is not possible to create a histogram for pixels existing at the top, bottom, left, and right edges of the downsampled image data, and entropy values can be calculated. Can not. Therefore, in this example, the creation of the histogram and the calculation of the entropy value are not performed for the upper, lower, left, and right end pixels, and the entropy value of the adjacent pixel is copied and used as the entropy value of the end pixel (in FIG. 24A). The pixel where the entropy value is copied is shown surrounded by a broken line). However, the method for determining the entropy value of the end pixel is not limited to this.

  In this example, the photographic feature value calculation unit 43 further averages the calculated entropy values in a 3 × 3 size region of the pixel of interest and surrounding pixels. The average value of the entropy shown in FIG. 24B is obtained by averaging the entropy values shown in FIG. Since the averaging is performed with a 3 × 3 size, the entropy values corresponding to the pixels existing at the upper, lower, left, and right edges of the image data cannot be averaged. Therefore, in this example, the average value of the entropy value of the adjacent pixel is copied as the average value of the entropy value of the end pixel without averaging for the upper, lower, left and right end pixels (FIG. 24B). (The pixel where the average value of the entropy value is copied is surrounded by a broken line). However, the determination method of the average value of the entropy value regarding the end pixel is not limited to this.

  The photographic feature amount calculation unit 43 gives the averaged entropy value to the gray scale conversion processing unit 25 as a photographic feature amount. In addition, the photograph feature amount calculation unit 43 determines a photograph region by comparing the calculated entropy value with a predetermined threshold value. In this example, the threshold value is set to 2.57, and FIG. 24B shows a photographic region by hatching pixels whose entropy value exceeds the threshold value. Further, the pixels of the image data before downsampling corresponding to this photographic area are shown in FIG. 25 with hatching. The photograph feature amount calculation unit 43 outputs the photograph area determination result as a photograph area signal.

  The image processing apparatus 20 that has calculated the character feature amount, the halftone dot feature amount, and the photographic feature amount detects the halftone dot background region based on these feature amounts. FIG. 26 is a schematic diagram showing a halftone dot background region in the image data shown in FIG. The grayscale conversion processing unit 25 given the character feature amount, the halftone dot feature amount, and the photographic feature amount respectively compares each feature amount with a predetermined threshold value to determine the condition shown in the above (Equation 4). Pixels that satisfy are detected as a halftone dot background region. In this example, threshold value A = 500, threshold value B = 2450, threshold value C = 2.57 in (Equation 4), and (edge degree <500) and (complexity> 2450) and (entropy value <2.57) Pixels in the halftone dot base region satisfying the above are hatched in FIG.

  As described above, the grayscale conversion processing unit 25 can detect the halftone dot background region from the given image data. However, for the pixels at the top, bottom, left, and right edges of the image data (shown surrounded by broken lines in FIG. 26), the amount of information is insufficient at the time of calculating the feature amount, and there is a risk of including erroneous determination or error. Therefore, the image forming apparatus may be configured not to use the detection results for the pixels at the upper, lower, left, and right edges of the image data when using the detection results of the halftone dot background region in the subsequent processing. The gray scale conversion processing unit 25 can also perform correction processing such as expansion processing and contraction processing on the detection result of the halftone dot background region. Thus, for example, in the detection result shown in FIG. 26, it is possible to remove a portion of the photographic area that protrudes by one pixel and to fill a portion that lacks the character area.

  Further, the gray scale conversion processing unit 25 checks whether the halftone dot area and the density difference between the characters on this area are within a predetermined range. For this reason, the gray scale conversion processing unit 25 first designates pixels whose complexity level given as the halftone dot feature amount exceeds the threshold value B (= 2450) as a halftone dot region, and labels each pixel in this region according to the procedure described above. To do. FIG. 27 is a schematic diagram showing a result of labeling the halftone dot region of the image data shown in FIG. FIG. 28 is a schematic diagram showing a halftone dot region in the image data shown in FIG. FIG. 29 is a schematic diagram showing a character area in the image data shown in FIG.

  In this example, labels 1 to 4 are attached to the pixels in the halftone dot area (see FIG. 27). However, since labels 1 to 4 are regarded as the same label, one group is included in the input image data. It can be determined that the halftone dot region is included. The gray scale conversion processing unit 25 calculates the average density (average pixel value) of the pixels in this halftone dot area (shown by hatching in FIG. 28), and obtains the average density = 137.

  Further, the gray scale conversion processing unit 25 sets a pixel whose edge degree given as a character feature amount exceeds a threshold A (= 500) as a character region (shown by hatching in FIG. 29), and is shown in FIG. The difference is calculated by comparing the density value of each pixel of the character area existing on the halftone dot area with the average density of the halftone dot area = 137, and the minimum difference is obtained. In this example, the difference from the character having the density value of 220 is the smallest, and the smallest difference is 83.

  The gray scale conversion processing unit 25 determines whether or not the minimum difference between the average density of the halftone dot area and the character density is within a predetermined range, and if it is within the predetermined range, the density correction of the halftone dot background area is performed. . In this example, when the minimum difference is smaller than 100, the grayscale conversion processing unit 25 performs density correction, and since the minimum difference is 83, the image data shown in FIG. Is called.

  The gray scale conversion processing unit 25 performs the density correction of the halftone dot background region using the above-described density correction formula 1 or the density correction formula 2, but the average density of the halftone dot background region for which the density correction is performed is higher than the predetermined density. Whether to use the density correction formula 1 or the density correction formula 2 is determined depending on whether it is high (dark) or low (light). In this example, when the average density value does not exceed 150, the gray scale conversion processing unit 25 performs high density (high) when the average density value does not exceed 150, and performs density correction according to the density correction formula 1, and when the average density value exceeds 150, It is assumed that the density is low (thin) and density correction is performed by density correction formula 2. 26 is 140 when the average density of the halftone dot background area shown in FIG. 26 is calculated. Since the density does not exceed the predetermined density 150 and the density is high (dark), the gray scale conversion unit 25 performs density correction according to the density correction formula 1. Do.

  FIG. 30 is a schematic diagram showing the density correction result of the halftone dot background region for the image data shown in FIG. The gray scale conversion processing unit 25 corrects the density value of each pixel in the halftone dot background region using the above-described density correction formula 1 (or a correction table created in advance based on FIG. 14). Thereby, since the density | concentration of a halftone dot background area | region can be made low (thin), the visibility of the character of a character area | region can be improved.

  The gray scale conversion processing unit performs processing such as the above-described halftone dot background region detection and density correction for each of the RGB components of the color image data, and performs gray scale conversion on the image data after the density correction. As a gray scale conversion method, various methods such as a method of using the G component of color image data as it is, a method of calculating and using a luminance value, and a method of calculating and using a weighted average of RGB pixel values are used. However, in this example, the grayscale conversion processing unit 25 performs conversion to grayscale using the G component as it is. In the case of this method, since the image data shown in FIG. 30 is of the G component, this image data is the result of gray scale conversion. The gray scale conversion processing unit 25 gives the converted gray scale image data to the spatial filter processing unit 26. Thereafter, the grayscale image data is subjected to appropriate image processing by the spatial filter processing unit 26, scaling processing unit 27, output tone correction unit 28, and tone reproduction processing unit 29, and is output by the image output device 12. Is output.

  In the image forming apparatus according to the present invention having the above-described configuration, a character feature amount indicating a character image feature, a halftone dot feature amount indicating a halftone image feature, and a photograph image feature amount indicating the feature of a photographic image from input image data. By calculating the feature quantity and detecting the halftone dot background area based on the calculation result of each feature quantity, the halftone dot background area can be detected with high accuracy, and the halftone dot background area can be detected. Processing such as density correction can be performed with high accuracy.

  Further, the image forming apparatus according to the present invention is configured to calculate the edge degree (edge inclination degree) as the character feature amount for each pixel of the input image data. It is possible to extract the possible ones without missing, and the character area can be detected with high accuracy, and the halftone dot background area using this can be detected with high accuracy.

  In addition, the image forming apparatus according to the present invention calculates pixel value differences in four directions including the main scanning direction of input image data, and calculates a complexity as a halftone dot feature amount from the sum of these differences. Therefore, it is possible not to detect only the presence or absence of halftone dots, so that even if the complexity of the halftone dots is small, it is possible to extract what may be halftone dots without missing, The detection of the ground area can be performed with high accuracy.

  In addition, the image forming apparatus according to the present invention is configured to obtain the local histogram by down-sampling the input image data and the calculated edge degree, thereby speeding up the processing because the calculation amount can be reduced. can do. Further, by adopting a configuration that calculates an entropy value as a photographic feature amount from this histogram, it is possible to accurately detect a photographic region in which the density change covers a wide range, and to detect a halftone dot background region with high accuracy. .

  In addition, the image forming apparatus according to the present invention has an edge degree calculated as a character feature amount that does not exceed the threshold value A, a complexity level calculated as a halftone dot feature amount exceeds the threshold value B, and an entropy calculated as a photographic feature amount. By adopting a configuration in which pixels whose values do not exceed the threshold value C are detected as pixels in the halftone dot background region, it is possible to accurately detect a dot region that is not a character region or a photo region as a halftone dot background region.

  In addition, the image forming apparatus according to the present invention is configured to perform density correction on the pixels in the halftone dot background area when the difference between the density of the halftone dot area and the density of the character on the halftone dot area is within a predetermined range. Thus, the contrast between the halftone dot background and the characters can be ensured, and the visibility of the characters in the formed image can be improved. In addition, the density correction is performed such that the density is reduced when the density of the halftone dot background area is high, and the density is increased when the density is low, so that the contrast between the halftone dot background and the character is more sure. Can be secured.

  In the present embodiment, the image forming apparatus is configured to perform density correction on the detected pixels in the halftone dot background region. However, the processing performed on the halftone dot background region is not limited to this. For example, A configuration may be used in which a strong smoothing filter is applied to the halftone dot background region, whereby it is possible to prevent the image from being moire and display the image with a smooth background color. In addition, for example, a process for removing a halftone dot background region may be performed, whereby the amount of toner or the like necessary for image formation can be reduced. Still other image processing may be performed.

  Further, a computer program for performing image processing such as the above-described halftone dot background region detection processing and density correction processing is created, and the program code (execution format program, intermediate code program, source program, etc.) 101 is stored in the memory card 100 or the like. The image forming apparatus may be configured to perform the above-described image processing by recording the program code 101 on the recording medium and reading and executing the program code 101. That is, each block in the image processing apparatus 20 shown in FIG. 1 may be configured as hardware or may be configured as software by executing the program code 101. The program code 101 is not limited to be recorded on the memory card 100, and the program code 101 is recorded on a recording medium fixed to a device such as a ROM, RAM, or hard disk such as an image forming apparatus or a computer. Alternatively, the program code 101 may be recorded on a portable recording medium that can be attached to and detached from the device. Examples of the removable recording medium include a tape medium such as a magnetic tape or a cassette tape, a magnetic disk medium such as a flexible disk or a hard disk, a disk medium such as a CD-ROM, MO, MD, or DVD, an IC card, or an optical card. The program code 101 may be stored in a card medium or a semiconductor memory such as a mask ROM, EPROM, EEPROM, or flash memory. Thus, the program code 101 for performing image processing such as halftone dot background region detection processing and density correction processing can be freely carried by the recording medium.

  Further, the program code 101 held in these recording media may be configured to be directly read and executed by the image forming apparatus or the computer or the like, and may be recorded on a hard disk or the like provided in the image forming apparatus or the computer or the like. A configuration may be adopted in which the program code 101 that is preinstalled in the medium and installed in the image forming apparatus or the computer is read and executed.

  In addition, a connection function may be installed in an image forming apparatus, a computer, or the like on a network such as the Internet and a LAN, and the program code 101 may be downloaded via the network. In this case, a program for performing the download process may be stored in advance in an image forming apparatus or a computer, or may be configured to install and execute a program held in various recording media. Note that the present invention can also be realized in the form of a computer data signal in which the program code 101 is embedded in a carrier wave embodied by electronic transmission. Further, the program code 101 is not for executing only the processing related to monochrome output including the detection processing of the halftone dot background area and the density correction processing, but is executed comprehensively including other processing of the image forming apparatus. It may be configured as described above.

  Further, as the image input device 11, for example, a device such as a flat head scanner, a film scanner, or a digital camera can be used. As the image output device 12, for example, an image display device such as a CRT display or a liquid crystal display, or an electrophotographic or inkjet printer that outputs a processing result to a recording paper such as paper can be used. Further, the image forming apparatus may include a modem or the like as communication means for connecting to a server apparatus or the like via a network.

  In the present embodiment, the image forming apparatus is configured to perform image processing such as halftone dot background area detection processing and density correction processing on RGB image data, but the present invention is not limited to this. Depending on the image processing configuration and the like, the image processing may be performed on image data in another color space such as CMYK or CIELAB. In this case, in a color space such as CMYK in which density values and color information are mixed, image processing may be performed by the same processing as in the case of RGB. In a color space such as CIELAB in which luminance and chromaticity are separated, processing may be performed with luminance as the center, and chromaticity information may be used as auxiliary information.

1 is a block diagram illustrating a configuration of an image forming apparatus according to the present invention. 1 is a block diagram illustrating a configuration of an image processing apparatus when an image forming apparatus performs color output. It is a block diagram which shows the structure of an area | region separation process part. It is a schematic diagram for demonstrating the calculation method of the edge degree by a character feature-value calculation part. It is a flowchart which shows the procedure of the character feature-value calculation process which a character feature-value calculation part performs. It is a schematic diagram for demonstrating the calculation method of the complexity by a halftone dot feature-value calculation part. It is a flowchart which shows the procedure of the halftone dot feature-value calculation process which a halftone dot feature-value calculation part performs. It is a schematic diagram for demonstrating the calculation method of the entropy value by the photograph feature-value calculation part. It is a schematic diagram for demonstrating the calculation method of the entropy value by the photograph feature-value calculation part. It is a flowchart which shows the procedure of the photograph feature-value calculation process which a photograph feature-value calculation part performs. It is a flowchart which shows the procedure of the process which a gray scale conversion process part performs. It is a flowchart which shows the procedure of the halftone dot area | region detection process which a gray scale conversion process part performs. FIG. 6 is a schematic diagram for explaining density correction formula 1; FIG. 6 is a schematic diagram for explaining density correction formula 1; FIG. 6 is a schematic diagram for explaining a density correction formula 2; FIG. 6 is a schematic diagram for explaining a density correction formula 2; It is a schematic diagram for demonstrating the density correction process which a gray scale conversion process part performs. It is a schematic diagram for demonstrating the density correction process which a gray scale conversion process part performs. It is a schematic diagram which shows an example of image data. This is an enlarged view of the periphery of the area A in the image data shown in FIG. It is a schematic diagram which shows the result of having calculated the edge degree as a character feature-value from the image data shown in FIG. It is a schematic diagram which shows the result of having calculated the complexity as a halftone dot feature amount from the image data shown in FIG. It is a schematic diagram which shows the calculation process which calculates a photograph feature-value from the image data shown in FIG. It is a schematic diagram which shows the result of having calculated entropy as a photograph feature-value from the image data shown in FIG. It is a schematic diagram which shows the halftone dot area | region in the image data shown in FIG. FIG. 21 is a schematic diagram showing a halftone dot background region in the image data shown in FIG. 20. FIG. 21 is a schematic diagram illustrating a result of labeling a halftone dot region of the image data illustrated in FIG. 20. It is a schematic diagram which shows the halftone dot area | region in the image data shown in FIG. It is a schematic diagram which shows the character area in the image data shown in FIG. FIG. 21 is a schematic diagram illustrating a density correction result of a halftone dot background region for the image data illustrated in FIG. 20.

Explanation of symbols

1 CPU
2 ROM
3 RAM
4 Operation unit 5 Display unit 6 Memory mounting unit 11 Image input device (input means)
12 Image output device (image forming means)
DESCRIPTION OF SYMBOLS 20 Image processing apparatus 21 A / D conversion part 22 Shading correction | amendment part 23 Input gradation correction | amendment part 24 Area | region separation | separation processing part 25 Gray scale conversion process part (Characteristic feature amount determination means, halftone dot feature amount determination means, photograph feature-value determination means , Halftone dot area detection means, image correction means)
26 Spatial Filter Processing Unit 27 Scaling Processing Unit 28 Output Tone Correction Unit 29 Tone Reproduction Processing Unit 41 Character Feature Amount Calculation Unit (Character Feature Amount Calculation Means)
42 Halftone dot feature value calculation unit (halftone feature value calculation means)
43 Photo feature amount calculation unit (photo feature amount calculation means)
100 memory card (recording medium)
101 Program code (computer program)

Claims (9)

In a halftone dot background detection method for detecting a background area composed of halftone dots from an image,
Calculating a character feature amount indicating a feature of the character image from the image;
Calculating a halftone dot feature amount indicating a characteristic of a halftone dot image from the image;
Calculate a photographic feature amount indicating the characteristics of the photographic image from the image,
It is determined whether the calculated character feature amount exceeds a predetermined amount,
It is determined whether the calculated halftone dot feature amount exceeds a predetermined amount,
Determine whether the calculated feature value of the photograph exceeds a predetermined amount,
A halftone dot region detection method, comprising: detecting a halftone dot region based on a combination of determination results. The halftone dot region detection method according to claim 1, wherein a gradient between a pixel value of each pixel and a pixel value of a peripheral pixel of the pixel is calculated from the image as the character feature amount. 3. The halftone dot region detection according to claim 1, wherein, as the halftone dot feature amount, a change degree of a pixel value of an adjacent pixel in one or a plurality of directions is calculated from the image. Method. The halftone dot region detection method according to any one of claims 1 to 3, wherein entropy with respect to the appearance frequency of a pixel value is calculated for each predetermined region of the image as the photographic feature amount. . An area where the halftone dot feature amount is determined to exceed a predetermined amount and the character feature amount and the photographic feature amount are determined not to exceed a predetermined amount is detected as the halftone dot background region. The halftone dot region detection method according to any one of claims 1 to 4. In an image processing apparatus comprising a halftone dot region detection means for detecting a background region composed of halftone dots from an input image,
A character feature amount calculating means for calculating a character feature amount indicating a feature of the character image from the image;
Halftone dot feature amount calculating means for calculating a halftone dot feature amount indicating a feature of a halftone dot image from the image;
Photographic feature amount calculating means for calculating a photographic feature amount indicating the characteristics of a photographic image from the image;
A character feature amount determining means for determining whether or not the character feature amount calculated by the character feature amount calculating means exceeds a predetermined amount;
Halftone dot feature amount determining means for determining whether the halftone dot feature amount calculated by the halftone dot feature amount calculating means exceeds a predetermined amount;
Photographic feature amount determination means for determining whether or not the photographic feature amount calculated by the photographic feature amount calculation means exceeds a predetermined amount;
The halftone dot background area detecting means detects a halftone dot background area based on a combination of determination results by the character feature amount determining means, the halftone dot feature amount determining means, and the photographic feature amount determining means. An image processing apparatus. An image processing apparatus according to claim 6;
Input means for inputting an image to the image processing apparatus;
An image correction unit provided in the image processing device, for performing image correction on a halftone dot background region detected by a halftone dot region detection unit of the image processing device;
An image forming apparatus comprising: an image forming unit that forms an image based on an image that has been subjected to image correction by the image correcting unit. In a computer program for causing a computer to perform a process of detecting a background area composed of halftone dots from an input image,
On the computer,
Calculating a character feature amount indicating a feature of the character image from the image;
Calculating a halftone dot characteristic amount indicating a characteristic of a halftone dot image from the image;
Calculating a photographic feature amount indicating a feature of a photographic image from the image;
Determining whether the calculated character feature amount exceeds a predetermined amount;
Determining whether the calculated halftone dot feature amount exceeds a predetermined amount;
Determining whether the calculated feature value of the photograph exceeds a predetermined amount;
And a step of detecting a halftone dot background region based on a combination of determination results.   A computer-readable recording medium in which the computer program according to claim 8 is recorded.

Images