JP2010278535A - Image processing apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing apparatus capable of generating a plurality of different character attribute layer images from an input image without deteriorating picture quality and capable of generating a highly compressed image file in which color irregularity or blur is not generated in a boundary part between a line drawing and a pattern. <P>SOLUTION: The image processing apparatus includes: a storage processing part 27 for subjecting a character attribute layer image generated by a character attribute discrimination part 21 to reversible compression, subjecting an input image to non-reversible compression, storing respective compressed images in storage parts 6, 7, and extending the character attribute layer image or the input image stored in the storage part; and a correction processing part 26 for correcting the character attribute layer image based on the input image and the character attribute layer image stored in respective storage parts 6, 7. A resolution conversion part 23 separates the input image stored in the storage part into a character color layer image and a background color layer image based on the corrected character attribute layer image and converts respective layer images into respective predetermined resolutions. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a character attribute determining unit that generates a character attribute layer image for identifying a character region and a background region other than the character region from the input image, and a character color layer image and a background color layer from the input image based on the character attribute layer image. A resolution conversion unit that separates each layer image into a predetermined resolution, and an image compression that performs lossless compression of the character attribute layer image after resolution conversion and lossy compression of the character color layer image and the background color layer image. The present invention relates to an image processing apparatus that includes a storage unit that stores a high-compression image file based on each of the compressed layer images, and a storage unit that stores the image.

  As described in Japanese Patent Application Laid-Open No. H10-260260, an ITU-T recommendation T.264 is used for high-compression encoding of an image in which a line drawing such as a character and a design such as a photograph are mixed so that the image quality does not deteriorate. MRC (Mixed Raster Content) defined in No. 44 is drawing attention.

  In MRC (Mixed Raster Content), image data in which a line drawing and a continuous tone pattern are mixed is separated into a character attribute layer indicating which of the foreground layer, the background layer, and the foreground layer and the background layer is selected. After the resolution is individually converted according to the characteristics of the layer, it is compressed and encoded.

  The foreground layer is a multi-value image representing line drawing color information, the background layer is a multi-value image representing a pattern other than the line drawing area, that is, background color information, and the character attribute layer is a binary image representing the shape of the line drawing. .

  The foreground layer and the background layer are compressed and encoded by a lossy compression method such as JPEG (Joint Photographic Experts Group), and the character attribute layer is compressed and encoded by a lossless compression method such as MMR (Modified Modified READ). As a result, the compression rate and the compressed image quality can be kept high as compared with the case of compressing image data in which a line drawing and a pattern are mixed by the JPEG method alone.

  Usually, when decoding an encoded image and returning it to the original resolution, bilinear interpolation or cubic interpolation is employed. For example, when the foreground layer is interpolated, the decoded foreground layer is decoded. Blur or color blur occurs in the outline of the line drawing.

  This is because pixel values outside the region are reflected by the interpolation processing in the outline portion of the line drawing in the foreground layer, and particularly in an image read by a scanner or the like, the outline of the line drawing is blurred, which becomes even more remarkable. In addition, when the lossy-compressed foreground layer is decoded, mosquito noise and ringing noise are generated, and these noise components are taken in by interpolation processing, which causes blurring and color blurring.

  Even if the foreground layer after the interpolation process is masked with the character attribute layer, the portion of the character outline that is muddy or blurred appears in the reproduced image. Such a phenomenon similarly occurs in the background layer.

  Particularly in the case of a line drawing, such color turbidity and blurring in the vicinity of the outline deteriorate the appearance of the reproduced image, and in the case of a line drawing including a thin line portion, the fine line portion becomes a color different from the thick line portion. There seemed to be problems such as the characters appearing to be separated on the way.

  Therefore, Patent Document 2 discloses the following image processing apparatus. That is, as shown in FIG. 10A, the image processing apparatus converts an original image 1000 into a mask plane 1040 representing a character / line drawing area, a foreground plane 1042 carrying character / line drawing color information, The background plane 1044 having image information other than the line drawing is separated. When the resolution of the foreground plane 1042 is reduced, the mask plane 1040 is referred to determine whether each pixel of the resolution reduction result is a pixel belonging to a character / line drawing. Interpolation is performed only from the values of pixels belonging to the character / line drawing on the foreground plane 1042. In the case of increasing the resolution of the reduced foreground plane 1052 and the reduced background plane 1054 that have been reduced in resolution, the same interpolation is performed with reference to the reduced masks 1053 and 1055.

  As shown in FIG. 10B, such an image processing apparatus includes a receiving unit 152 that receives MRC format data, a format decompression unit 154, a mask decompression unit 156, a mask reduction unit 158, a foreground decompression unit 160, and a background. The expansion unit 162, the foreground enlargement unit 164, the background enlargement unit 166, the single layer synthesis unit 168, and the image output unit 170 are provided.

  This will be described in detail below. The format decompression unit 154 decomposes the data in the MRC format received by the reception unit 152 into each plane (compressed data), and the mask decompression unit 156 converts the compressed data of each plane given from the format decompression unit 154 into the corresponding code. The data is expanded according to the conversion method to obtain a mask plane 1040, a reduced foreground plane 1052, and a reduced background plane 1054.

  The mask reduction unit 158 reduces the mask plane 1040 whose data has been expanded by the mask expansion unit 156 again, and creates reduced-resolution reduced mask planes 1053 and 1055 that match the resolutions of the reduced foreground plane 1052 and the reduced background plane 1054, respectively. The foreground enlargement unit 164 obtains the foreground plane 1062 by performing enlargement processing with reference to the reduction mask plane 1053 when enlarging the reduced foreground plane 1052 to the same resolution as the foreground plane 1042 before reduction. When enlarging the reduced background plane 1054 to the same resolution as the background plane 1044 before reduction, the background enlargement unit 166 executes enlargement processing with reference to the reduced mask plane 1055 to obtain the background plane 1064.

  The foreground plane 1062 and the background plane 1064 obtained in this way are masked with a mask 1040, thereby preventing or reducing the color turbidity of the characters and the outline of the picture portion of the reproduced image 1070.

JP-A-11-122479 JP 2004-187000 A

  Depending on the region separation algorithm, the edge of the pattern area other than the character area is extracted as noise in the character attribute layer image generated by the character attribute determination unit for the input image obtained via an image input device such as a scanner. In such a case, it is necessary to remove noise. Further, it may be required to generate a specific character region, for example, a character region of a specific color, as the character attribute layer image among the separated character regions.

  In such a case, it is necessary to generate the character attribute layer image again by the character attribute discriminating unit, but usually the input image is not saved as raw data from the viewpoint of the storage capacity, and the JPEG method or the like is used. Since the irreversibly compressed image is stored in the hard disk 7 via the memory 6 or the memory 6, it is necessary to decompress the input image after compression.

  Since such a decompressed image has deteriorated image quality, when a character attribute layer image is generated from the decompressed image by the character attribute discriminating unit 21, there has been a problem that a character region having a remarkable resolution deterioration cannot be properly separated.

  In the image processing apparatus described in Patent Document 2 described above, low-resolution reduced mask planes 1053 and 1055 that match the resolution of the reduced foreground plane 1052 and the reduced background plane 1054 are created, and the reduced foreground plane 1052 and the reduced background are generated. In enlarging the plane 1054, it is necessary to execute an enlarging process while referring to the reduced mask planes 1053 and 1055, so that it is necessary to provide the mask reducing unit 158, and the process becomes complicated.

  Furthermore, when the resolution of the mask plane 1040 is reduced to the same level as that of the reduced foreground plane 1052 and the reduced background plane 1054, the outline of the character is broken, and the reduced foreground plane 1052 and the reduced background plane 1054 are changed based on the broken outline of the character. Further improvement is demanded in that there is a problem that even if the reproduced image 1070 is obtained from the foreground plane 1062 and the background plane 1064 that have been subjected to the enlargement process, the outline of the character before the resolution reduction cannot be accurately reproduced.

  In view of the above problems, an object of the present invention is to enable generation of a plurality of different character attribute layer images from an input image without degrading the image quality, and to generate color unevenness and blur at the boundary between a line drawing and a pattern. An object of the present invention is to provide an image processing apparatus capable of generating a high-compression image file having no image.

  In order to achieve the above-described object, a first characteristic configuration of an image processing apparatus according to the present invention includes a character region and a background region other than the character region from an input image, as described in claim 1 of the claims. A character attribute determination unit that generates a character attribute layer image to be identified, and a resolution that separates an input image into a character color layer image and a background color layer image based on the character attribute layer image, and converts each layer image to a predetermined resolution. A conversion unit, an image compression unit that losslessly compresses the character attribute layer image whose resolution has been converted and irreversibly compresses the character color layer image and the background color layer image, and a high-compression image file based on each compressed layer image An image processing apparatus comprising a file generation unit for generating and a storage unit for storing an image, wherein the character attribute layer image generated by the character attribute determination unit is In addition to inverse compression, the input image is irreversibly compressed and stored in the storage unit, and a storage processing unit that decompresses the character attribute layer image or the input image stored in the storage unit, and the input image stored in the storage unit And a correction processing unit that corrects the character attribute layer image based on the character attribute layer image, and the resolution conversion unit converts the character color layer image from the input image stored in the storage unit based on the corrected character attribute layer image. And the background color layer image, and each layer image is converted to a predetermined resolution.

  Since the input image stored in the storage unit by the storage processing unit is irreversibly compressed, the image quality may deteriorate when the decompression process is performed, but the character attribute layer image generated based on the input image before compression is Since the image is reversibly compressed, the image quality does not deteriorate even when decompression is performed.

  Accordingly, the correction processing unit removes unnecessary character attribute information included in the character attribute layer image by correcting the character attribute layer image based on an image obtained by performing predetermined processing on the decompressed input image. can do.

  In the second feature configuration, as described in claim 2, in addition to the first feature configuration described above, the correction processing unit converts the input image stored in the storage unit into one of hue, saturation, and luminance. The character attribute layer image is corrected based on the converted input image.

  According to the above configuration, for example, if the luminance of the input image is converted, the black character becomes a low luminance image region and the pattern region becomes a relatively high luminance image region. Therefore, between the pixels of the luminance conversion image and the character attribute layer image. If only a low luminance area is extracted from the character attribute layer image by performing a logical operation, the edge noise of the picture area can be removed. Similarly, if saturation conversion is performed, for example, a high saturation region that is not used as a line drawing of characters or the like can be removed from the character attribute layer image.

  According to the third feature configuration, as described in claim 3, in addition to the first or second feature configuration described above, a character color from an input image based on a mask image generated based on a character attribute layer image. It is equipped with an index value replacement processing unit that extracts pixels and generates a character color layer image by converting pixels with the same value as the index value indicating ineffective pixels from the extracted character color pixels to a pixel value different from the index value, and resolution conversion The unit divides the character color layer image processed by the index value replacement processing unit into pixel blocks having a predetermined number of pixels set based on the resolution conversion rate, and refers to the mask image to belong to the mask image in the pixel block If a pixel exists, the average value of the pixel is used as the pixel value after resolution conversion. If there is no pixel belonging to the mask image in the pixel block, the index value is changed. It is configured to generate a reduced character color layer image by repeating the process of converting the pixel value later to all pixel blocks, and dilate the contour portion of the reduced character color layer image determined based on the index value. It is in.

According to the above-described configuration, the following specific operational effects are exhibited.
First, the reduced character color layer image generated by the resolution conversion unit is reduced based on the average value of only the character colors of the pixel blocks having a predetermined number of pixels, and therefore occurs when the bilinear interpolation method or the cubic interpolation method is used. No blurring of the color of the contour pixel occurs.

  In addition, the pixel value of the area corresponding to the character attribute becomes a pixel value indicating the character color, and the pixel value of the area corresponding to the background color that is not the character attribute becomes the index value, so without referring to the character attribute layer image, An area corresponding to the character attribute can be expanded based on the index value. In addition, the calculation load is significantly reduced compared to the expansion processing before the resolution conversion.

  In addition, since the character color layer image has been converted from a character color pixel to a pixel value that is the same as the index value indicating an ineffective pixel, the pixel value is different from the index value. The character color pixel does not change greatly, and the character color image does not deteriorate.

  Further, even if the reduced character color layer image obtained by expanding the highly compressed image file generated by the file generation unit is subjected to resolution conversion so as to expand to the original resolution, the pixel subjected to the expansion processing by the resolution conversion unit Therefore, the degree of blurring or blurring that occurs in the outline of the enlarged character color layer image is suppressed, and the character color layer image masked with the character attribute layer image has little effect.

  In the fourth feature configuration, as described in claim 4, in addition to the third feature configuration described above, when the expansion target pixel is an index value, the expansion processing uses the expansion target pixel as a central pixel. An average value of pixels other than the pixel indicating the index value is generated as the pixel value of the expansion target pixel among the predetermined number of pixel blocks to be processed, and when the expansion target pixel is not the index value, the predetermined value with the expansion target pixel as the central pixel Among the plurality of pixel blocks, an average value of pixel values whose difference from the pixel value of the central pixel is within a predetermined threshold is generated as the pixel value of the expansion target pixel.

  When expanding the outline of the reduced character color layer image or reduced background color layer image reduced by the resolution conversion unit, if the expansion target pixel is an index value, the pixel value of the expansion target pixel is the surrounding character color or Since it is generated as an average value of the background color, it is possible to secure a character color or background color around the expansion target pixel, and when the expansion target pixel is not an index value, the pixel value of the expansion target pixel is Since the difference from the pixel value of the expansion target pixel is generated as an average value of pixel values within a predetermined threshold, color mixing can be avoided when contour pixels of different colors are adjacent to the contour pixel. In any case, an unnatural color change that suddenly changes from the contour pixel to the expansion target pixel does not occur, and a smooth expansion processing effect can be obtained.

  The fifth feature configuration is that, as described in claim 4, in addition to the third or fourth feature configuration described above, the mask image is an image obtained by shrinking the outline of the character attribute layer image. .

  When the region separation unit extracts character color pixels from the input image, if an image obtained by shrinking the contour of the character attribute layer image is used as a mask image, the color unevenness of the contour portion of the character region included in the input image Can be surely eliminated.

  Then, the reduced character color layer image expanded by the subsequent resolution conversion unit becomes a clear image with no color unevenness in the contour portion.

  As described above, according to the present invention, it is possible to generate a plurality of different character attribute layer images from an input image without degrading image quality, and to prevent color unevenness and blurring from occurring at the boundary between a line drawing and a pattern. An image processing apparatus capable of generating a compressed image file can be provided.

Illustration of network system Multi-function machine perspective view Illustration of the main functional blocks of the MFP Functional block diagram around the image processing unit Block diagram of the compression / decompression image generator Illustration of resolution conversion rate Illustration of resolution conversion table Illustration of resolution conversion procedure Explanatory drawing of expansion processing procedure (A) is explanatory drawing of the image restoration in the conventional image processing apparatus, (b) is a functional block diagram of the conventional image processing apparatus

  Hereinafter, a multi-function machine which is an example of an image processing apparatus according to the present invention will be described. As shown in FIG. 1, a multifunction machine 1 includes PCs (personal computers) 91 and 92, a network printer 93, a FAX 94, and the like via a communication line 95 such as a public line or a LAN (local area network). Part of a network system connected according to

  In the network system, an image read by a scanner provided in the multifunction device 1 or FAX 94 is transmitted to the PCs 91 and 92, and an image created by an application of the PCs 91 and 92 is printed by the network printer 93 or the multifunction device 1. It is configured.

  2, 3, and 4, the multifunction device 1 includes an operation unit 2 that is a man-machine interface, a scanner 3 that reads a document image, a memory 6 that stores an image input from the scanner 3, An image processing unit 14 that processes an image stored in the memory 6 based on copy conditions such as copy density set in the operation unit 2 to generate an output image, and an output image generated by the image processing unit 14 A printer engine 5 that forms a toner image, transfers the toner image onto a sheet, heat-fixes and outputs the sheet, and an HDD (hard disk drive) that stores contents such as image data processed by the image processing unit 14 ) 7 and other functional blocks, and a control unit for controlling each functional block is provided.

  Specifically, an operation control unit 12 that controls the operation unit 2, a scanner control unit 13 that controls the scanner 3, a memory control unit 16 that controls the memory 6, and a printer engine control unit 15 that controls the printer engine 5, A network interface circuit 8 for controlling communication with the PCs 91 and 92, the network printer 93, etc. connected via the communication line 95, a FAX modem 90 for transmitting and receiving image data to and from the FAX 94, and an e-mail processing unit 80. Are connected via a command bus 18 for transmitting / receiving control signals to / from each other and an image bus 19 for transmitting / receiving image data.

  The system control unit 17 controls the above-described functional blocks connected by a bus, a microcomputer incorporating a CPU, a ROM storing a control program executed by the CPU and table data for control. A RAM for storing calculation data and control data is provided, and system programs and various processing programs stored in the ROM are read and executed by an input operation on the operation unit 2 and an input operation from a PC 91 or the like on the network. To control the multifunction device 1 in an integrated manner.

  The operation control unit 12, the scanner control unit 13, the printer engine control unit 15, the memory control unit 16, and the like each include a ROM that stores a CPU and a CPU operation program, a RAM that is a work area of the CPU, a peripheral circuit, and the like. Consists of. Each CPU reads out and executes an operation program from each ROM in accordance with an instruction from the system control unit 17, and operates in accordance with an algorithm defined in the operation program.

  The scanner 3 photoelectrically converts a document image placed on a document placement unit provided on the upper surface of the multifunction device 1 to generate a digital image as an input image, and stores the input image in the memory 6.

  The printer engine 5 is a functional block that prints an image on paper by an electrophotographic method and outputs it. The photosensitive member, a charging device that uniformly charges the surface of the photosensitive member, a laser beam is scanned based on the output image, and the photosensitive member is scanned. Image forming apparatus comprising: an exposure device that exposes a body surface to form an electrostatic latent image; a developing device that visualizes the formed electrostatic latent image as a toner image; and a transfer device that transfers the toner image onto a sheet. A thermal fixing device that melts and fixes the toner image transferred to the paper, a transport mechanism that transports the paper, etc., and the respective parts cooperate to execute a printing operation in response to a command from the printer engine control unit 15 To do.

  The operation unit 2 displays a mode setting key for setting the operation mode of the multifunction device 1, operation keys necessary for each operation mode, and a touch panel type liquid crystal display unit 2 a for displaying error information and the like, Various operation keys 2b such as a power key, a start key, and a stop key are provided.

  The mode setting key includes a copy key for selecting a copy mode in which an image read by the scanner 3 is printed by the printer engine 5, a FAX key for transmitting the image read by the scanner 3 to the outside via the FAX modem 90, and a scanner 3 includes a box key for storing the image read in 3 in a box which is a storage area partitioned in the hard disk 7.

  Regardless of which mode is set, if there is a print request from the PC 91, 92 or the like via the network interface circuit 8 or reception from an external FAX 94 via the FAX modem 90, the system control is performed. The unit 17 controls so that the corresponding image is printed out by the printer engine 5 or stored in the hard disk 7.

  In the box mode, which is shifted by operating the box key, either an input mode for storing an image on the hard disk 7 or an output mode using an image stored on the hard disk 7 is selected. When the input mode is selected, the scanner is selected. 3 is stored as an image file in a designated area partitioned on the hard disk 7, and when the output mode is selected, the image file in the designated area partitioned on the hard disk 7 is printed out by the printer engine 5. It is transmitted from the modem 90 or transmitted from the electronic mail processing unit 80 to the transmission destination as an attached file.

  The image processing unit 14 performs an operation on the input image stored in the setting memory 6 based on each operation mode set by the operation unit 12 and image processing information set via the liquid crystal display unit 2a in each operation mode. A circuit block including an ASIC for image processing or the like that executes predetermined processing and outputs the processed image to the printer engine 5 or stores it in the memory 6 or the hard disk 7 is provided.

  The image processing unit 14 includes an output image generation unit 31, a compression / decompression image generation unit 20, and the like. Each of the output image generation unit 31 and the compression / decompression image generation unit 20 reads a block image composed of a predetermined number of scanning lines from an image of one original stored in the memory 6 and performs a predetermined process. After the processing, the process of storing the processed block image in the memory 6 again is repeatedly performed over all the block images, thereby generating the desired processed image on the memory 6.

  The output image generation unit 31 reads an input image composed of monochrome or RGB color component pixels input from the scanner 3, the network interface circuit 8, or the FAX modem 90 and stored in the memory 6, and performs predetermined image processing. In this block, the generated image is written in the memory 6 and output to the printer engine 5 or the FAX modem 90. The color correction unit 31a for color adjustment, the enlargement / reduction processing unit 31b for magnification conversion, and the gradation correction unit for tone conversion 31c includes circuit blocks such as a color conversion unit 31d that converts RGB color components into an output image of YMCK color components corresponding to each toner color component.

  Based on the operation information of the operation unit 2, an output image is generated using any one or more of the color correction unit 31a, the enlargement / reduction processing unit 31b, the gradation correction unit 31c, and the color conversion unit 31d.

  The compression / decompression image generation unit 20 generates a high-compression image file, for example, a high-compression PDF file, based on an input image read by the scanner 3 or input via the network interface circuit 8 and stored in the memory 6. This is a block for expanding a highly compressed image file into an original image file.

  The highly compressed image file generated by the compressed / decompressed image generation unit 20 is stored in a box or transmitted to the outside as an attachment file of an e-mail by the e-mail processing unit 80.

  As shown in FIG. 5, the compression / decompression image generation unit 20 includes a character attribute determination unit 21, an index value replacement processing unit 22, a resolution conversion unit 23, an image compression / decompression unit 24, a file generation unit 25, A plurality of circuit blocks such as a correction processing unit 26, a storage processing unit 27, and an image composition unit 28 are provided.

  The character attribute discriminating unit 21 separates a character region and a background region other than the character region from the input image stored in the memory 6 based on a known region separation algorithm such as an edge extraction process, and the character region, that is, a character or line drawing. This is a circuit block that generates a character attribute layer image that identifies the area and the non-character area, i.e., the pattern or photo area. The character attribute layer image is a binary image in which the character area is 1 and the non-character area is 0.

  Note that the character attribute determination unit 21 may generate a character attribute layer image for all characters in the character area included in the input image stored in the memory 6, or character attributes only for characters desired by the user. A layer image may be generated. For example, the file size of the character attribute layer image can be reduced by not generating the character attribute layer image only for yellow characters, or by not generating the character attribute layer image for characters other than black (gray). It becomes possible to reduce.

  The index value replacement processing unit 22 is a circuit block that extracts a pixel having the same value as an index value indicating a predetermined ineffective pixel from the input image stored in the memory 6 and converts the extracted pixel to a pixel value different from the index value. is there. Specifically, when the RGB color components are represented by 8-bit gradations, for each color component, any one of (0, 0, 0) black pixels and (255, 255, 255) white pixels is used as an index value. Is adopted.

  When a black pixel of (0, 0, 0) is set as an index value, the index value replacement processing unit 22 searches for whether or not a black pixel of (0, 0, 0) exists in the input image. If there is a black pixel of (0, 0, 0), it is converted to (1, 1, 1). Similarly, if any pixel component of RGB is 0, it is converted to 1.

  When the white value of (255, 255, 255) is set as the index value, the index value replacement processing unit 22 searches for whether the white pixel of (255, 255, 255) exists in the input image. , (255, 255, 255), the pixel value of (254, 254, 254) is converted. Similarly, if any pixel component of RGB is 255, it is converted to 254.

  As described above, the index value replacement processing unit 22 replaces the pixel value having the same value as the index value with the pixel value having the smallest difference, so that the hue of the input image hardly changes.

  Based on the mask image generated from the character attribute layer image, the resolution conversion unit 23 generates and extracts a character color layer image obtained by extracting character color pixels from the input image replaced by the index value replacement processing unit 22 A circuit block is provided for resolution conversion so that the character color layer image has a predetermined resolution.

  More specifically, the resolution conversion unit 23 divides the character color layer image arranged in the row and column directions into pixel blocks having a predetermined number of pixels set based on the resolution conversion rate, and refers to the mask image to obtain pixels. If there is a pixel belonging to the mask image in the block, the average value of the pixel is the pixel value after resolution conversion, and if there is no pixel belonging to the mask image in the pixel block, the index value is the pixel value after resolution conversion. Is repeated for all the pixel blocks divided along the pixel arrangement to generate a reduced character color layer image.

  As shown in FIG. 8, for example, when generating a 37.5 DPI character color layer image from an input image with a resolution of 600 DPI, the resolution conversion rate is 1/16, and the 16 pixels of the input image are subjected to resolution conversion. In order to generate one pixel, the character color layer image is divided into 4 × 4 pixel blocks. Here, the character attribute layer image itself is a mask image, a pixel corresponding to 1 of the mask image is a character color layer image, and a pixel corresponding to 0 of the mask image is an ineffective pixel indicating a region other than the character region.

  Note that the pixels 0 and 1 shown in FIG. 8 indicate a mask image, and the resolution is actually converted for any pixel value of R, G, and B of 8 bits for each pixel. Therefore, both the character color layer image and the background color layer image are composed of three layers corresponding to the colors R, G, and B, and the resolution conversion process is executed for the layer image of each color.

  If there are pixels belonging to the mask image in the 4 × 4 pixel block, that is, there are pixels corresponding to 1 of the mask image, the average value of these pixels is used as the pixel value after resolution conversion, and the 4 × 4 pixel block If there is no pixel belonging to the mask image, that is, if there is no pixel corresponding to 1 in the mask image, the index value is used as the pixel value after resolution conversion. By applying such processing to the entire character color layer image, a reduced character color layer image is generated. The index value may be a white pixel or a black pixel.

  That is, resolution conversion is performed at the same time while separating and extracting the character color layer image from the replaced input image.

  Further, after that, the resolution conversion unit 23 performs an expansion process on the contour portion of the reduced character color layer image determined based on the index value with respect to the reduced character color layer image.

  More specifically, the resolution conversion unit 23 performs an n × n (n is a positive odd number) size filter process with the pixel of the reduced character color layer image arranged in the row direction and the column direction as the center pixel. , That is, all pixels along the column direction and the row direction.

  In the filter processing, when the central pixel that is the expansion target pixel is an index value, out of a predetermined number of pixel blocks corresponding to the filter size, that is, n × n pixel blocks, pixels other than the pixel indicating the index value, that is, character color pixels. When the average value is generated as the pixel value of the central pixel, and the central pixel is not an index value, that is, a character color pixel, the pixel of the central pixel in a predetermined number of pixel blocks corresponding to the filter size, that is, an n × n pixel block An average value of pixel values whose difference from the value is within a predetermined threshold is generated as the pixel value of the expansion target pixel. When all the pixel blocks are index values, the index value is set as a new pixel value.

  FIG. 9 illustrates this situation. The resolution conversion unit 23 uses the pixels P (l, m), P (1, m + 1), P (l, m + 2),... Of the reduced character color layer image arranged in the row direction and the column direction as central pixels. The expansion process and the smoothing process are executed by executing the 3 × 3 size filter process as follows along the pixel array.

  It is preferable to set the filter size to 3 × 3, and the boundary of the character color area of the reduced character color layer image is expanded by one pixel, and the character color area is smoothed. In order to expand two pixels, the filter size may be set to 5 × 5.

  Further, the resolution conversion unit 23 generates a background color layer image obtained by extracting background color pixels from the input image processed by the index value replacement processing unit, based on the inverted mask image generated based on the character attribute layer image. A circuit block for converting the resolution so that the extracted background color layer image has a predetermined resolution is provided.

  Specifically, the resolution conversion unit 23 divides the background color layer image arranged in the row direction and the column direction into pixel blocks having a predetermined number of pixels set based on the resolution conversion rate, and refers to the inverted mask image. If there is a pixel belonging to the inverted mask image in the pixel block, the average value of the pixel is the pixel value after resolution conversion, and if there is no pixel belonging to the inverted mask image in the pixel block, the index value is the pixel after resolution conversion. The reduced value background color layer image is generated by repeating the processing of values for all pixel blocks divided along the pixel array.

  For example, when a background color layer image having a resolution of 150 DPI is generated from an input image having a resolution of 600 DPI, the resolution conversion rate is 1/4, and the background color is generated in order to generate one pixel after resolution conversion from four pixels of the input image. The layer image is divided into 2 × 2 pixel blocks. Here, an image obtained by inverting the character attribute layer image is an inverted mask image, a pixel corresponding to 1 of the inverted mask image is a background color layer image, and a pixel corresponding to 0 of the inverted mask image is ineffective indicating a region other than the design area. It becomes a pixel.

  If a pixel belonging to the inverted mask image exists in the 2 × 2 pixel block, that is, there is a pixel corresponding to 1 of the inverted mask image, the average value of these pixels is used as the pixel value after resolution conversion, and 2 × 2 If there is no pixel belonging to the inverted mask image in the pixel block, that is, there is no pixel corresponding to 1 of the inverted mask image, the index value is used as the pixel value after resolution conversion. By performing such processing on the entire background color layer image, a reduced background color layer image is generated. The index value may be a white pixel or a black pixel.

  That is, resolution conversion is performed at the same time while separating and extracting the background color layer image from the replaced input image.

  Further, after that, the resolution conversion unit 23 performs an expansion process on the reduced background color layer image on the contour portion of the reduced background color layer image determined based on the index value. The expansion process here is the same as the expansion process for the reduced character color layer image described above.

  Further, the resolution conversion unit 23 converts the resolution so that the character attribute layer image has a predetermined resolution. More specifically, the pixel block is divided into pixel blocks having a predetermined number of pixels set based on the resolution conversion rate, and is converted to 1 if the average value of the pixel block is larger than a predetermined threshold, and to 0 if the average value is smaller.

  For example, when generating a character attribute layer image with a resolution of 600 DPI from a character attribute layer image with a resolution of 600 DPI, the resolution conversion rate is halved, and one pixel after resolution conversion is generated from two pixels of the input image. The background color layer image is divided into 1 × 2 or 2 × 1 pixel blocks. If the average value of the corresponding two pixels is 1, the resolution-converted pixel is 1, and if the average value is 0, the resolution is converted. If the subsequent pixel is 0, and the average value is 0.5, the pixel after resolution conversion is set to 1. That is, the threshold value is set to a value less than 0.5.

  The resolution of the reduced character color layer image is preferably about 37.5 DPI because the character color can be specified, and so high resolution is not required. The resolution of the reduced background color layer image is such that the background pattern can be properly reproduced. A medium resolution of about 150 DPI is preferable, and the resolution of the reduced character attribute layer image requires a relatively high resolution in order to ensure the reproducibility of the line drawing, and is preferably about 300 DPI.

  However, the values are not limited to these values, and are values that are appropriately set based on the resolution of the input image and the total capacity of the final highly compressed image file. For example, if the resolution of the input image is 300 DPI, it is only necessary to convert the resolution of the character attribute layer image to the same magnification, and it is not necessary to substantially convert the resolution.

  The image compression / decompression unit 24 reversibly compresses the character attribute layer image whose resolution has been converted by the resolution conversion unit 23 and irreversibly compresses the character color layer image and the background color layer image. Since the reduction processing has already been performed, a particularly high compression rate is not required.

  As a lossless compression method for character attribute layer images, an MR method, an MMR method, a JBIG method, etc. can be adopted. As a lossy compression method for character color layer images and background color layer images, an orthogonal transform coding method such as JPEG method is used. Can be adopted.

  The file generation unit 25 generates a highly compressed image file based on each compressed layer image. That is, a highly compressed image file in which each layer image is arranged following the header information including the file name, the file capacity, the compression parameter and resolution parameter of each layer image, and the identification code of each layer image is generated.

  The high compression image file is generated on the memory 6 by the compression / decompression image generation unit 20, and the high compression image file is stored in the box.

  When a printer engine prints an image corresponding to a high-compression image file input from a PC or the like connected to a network or received by an e-mail processing unit or a high-compression image file stored in a box, An output image is generated in the reverse procedure.

  In other words, each layer image is separated by the file generation unit 25, each layer image is expanded into a reduced layer image by the image compression / decompression unit 24, and reversely converted to a high resolution by the resolution conversion unit 23.

  In the image composition unit 28, the character attribute layer image generated in this way is used as a mask image or an inverted mask image, pixels in the character area are cut out from the character color layer image, and pixels in the pattern area are cut out from the background color layer image. As a result, an output image in which the pixels in the character area and the pixels in the picture area are combined is generated.

  A bilinear interpolation method or a cubic interpolation method can be employed as an enlargement processing algorithm performed by the resolution conversion unit 23. Such image conversion processing to high resolution may cause color unevenness or blurring in the contour portion of the character color layer image or the background color layer image. However, as a result of the expansion processing described above, expansion is performed more than the original contour portion. Color irregularities and blurs that occur in the specified area, and the original character color area and the background color area are cut out from the character color layer image and the background color layer image with reference to the character attribute layer image. Does not appear in the playback image.

  Further, the output image generation unit 31 described above generates an image to be output to the printer engine 5.

  The case where the character attribute layer image itself generated by the character attribute determination unit 21 is used as a mask image or a reversal mask image for reduction processing used in the resolution conversion unit 23 is shown. If an image obtained by shrinking the contour of the character attribute layer image generated by the determination unit 21 is used, blur and color unevenness in the contour portion of the character color area inherent in the input image can be effectively eliminated, and subsequent expansion processing can be performed. An image with a clear outline can be secured, and if an image obtained by inverting the outline of the character attribute layer image after being expanded is used as an inverted mask image, blurring and color unevenness in the outline of the background color area inherent in the input image are eliminated. Can be effectively eliminated, and an image with a clear outline can be secured by subsequent expansion processing. This is particularly effective when character color areas are mixed in the background color area.

  In other words, this can be realized by providing the resolution conversion unit 23 with a contraction processing circuit block for contracting the character attribute layer image and an expansion processing circuit block for expanding the character attribute layer image. For example, a contraction processing circuit block that sets one pixel at the outermost edge of the character attribute area of the character attribute layer image to 0, or an expansion processing circuit that sets one pixel outside the outermost edge of the character attribute area to 1 What is necessary is just to comprise a block.

  The correction processing unit 26 provided in the compression / decompression image generation unit 20 is a circuit block that corrects the character attribute layer image based on the input image and the character attribute layer image stored in the memory 6.

  The storage processing unit 27 reversibly compresses the character attribute layer image generated by the character attribute determination unit 21 using the MMR method or the JBIG method, and irreversibly compresses the input image using the JPEG method or the like. This is a circuit block that is stored in a storage unit such as the memory 6 or the hard disk 7, expands the character attribute layer image or the input image stored in the storage unit, and expands it in the memory 6.

  A character attribute layer image is generated by the character attribute determination unit 21 for an input image via the scanner 3 or the like, but depending on the region separation algorithm, an edge of a pattern region other than the character region may be extracted as noise. In such a case, it is necessary to remove noise. Further, it may be required to generate a specific character region, for example, a character region of a specific color, as the character attribute layer image among the separated character regions.

  In such a case, it is necessary to generate the character attribute layer image again by the character attribute discriminating unit 21, but normally, the input image is not saved as raw data from the viewpoint of the storage capacity, and the JPEG method or the like is used. Since the irreversibly compressed image is stored in the hard disk 7 via the memory 6 or the memory 6, it is necessary to decompress the input image after compression.

  Since the image quality of such a decompressed image is degraded, there is a problem that when a character attribute layer image is generated from the decompressed image by the character attribute discriminating unit 21, a character region whose resolution is significantly degraded cannot be separated properly.

  Therefore, the character attribute layer image generated by the character attribute determination unit 21 from the original input image is reversibly compressed by the storage processing unit 27 and stored in the storage unit.

  The correction processing unit 26 is provided corresponding to such a case, and the input image stored in the memory 6, that is, the input image decompressed after being stored by the storage processing unit 27 is converted into hue, saturation or The image processing apparatus includes a hue conversion unit, a saturation conversion unit, and a luminance conversion unit that convert to any one of luminance images, and a character attribute layer image that has been decompressed after being compressed and stored by the storage processing unit 27, and a hue, saturation, or luminance A correction circuit is provided for correcting the character attribute layer image based on the input image converted into any one of them.

  For example, in a saturation conversion image, a black character becomes a low saturation image region and a pattern region becomes a relatively high saturation image region. Therefore, by performing a logical operation between the pixels of the saturation conversion image and the character attribute layer image, If only the low saturation area is extracted from the attribute layer image, the edge noise in the pattern area can be removed.

  For example, in a luminance conversion image, a black character becomes a low luminance image region and a pattern region becomes a relatively high luminance image region. Therefore, by performing a logical operation between pixels of the luminance conversion image and the character attribute layer image, a character attribute layer is obtained. If only the low luminance region is extracted from the image, the edge noise in the pattern region can be removed.

  For example, when it is necessary to extract only the character area of a specific color from the character attribute area included in the character attribute layer image, between the pixels of the color conversion image extracted only by the hue conversion unit and the character attribute layer image. By performing the logical operation, it is possible to generate a character attribute layer image of only a character region of a specific color.

  Based on the character attribute layer image corrected by the correction processing unit 26, the resolution conversion unit 23 separates the input image after decompression stored in the memory 6 into a character color layer image and a background color layer image, Each is converted to a predetermined resolution.

  Even in such a case, degradation of image quality is suppressed by the expansion processing by the resolution conversion unit 23 described above. In particular, an image obtained by contracting the contour of the character attribute layer image generated by the character attribute discriminating unit 21 is used as the mask image, and an image that is inverted after the contour of the character attribute layer image is expanded is used as the inverted mask image. Degradation of image quality is greatly suppressed.

  The resolution conversion unit 23 displays the resolution conversion rates for the character attribute layer image, the character color layer image, and the background color layer image on the liquid crystal display unit 2a of the operation unit 2 when the input mode is selected in the box mode. The selection key is selected.

  Specifically, four types of selection keys “low”, “medium”, “high”, and “automatic” are displayed. The resolution corresponding to each key is as shown in FIG. Regardless of the selection key of “low”, “medium”, or “high”, a constant resolution conversion rate is set regardless of the resolution of the input image. Accordingly, the capacity of each layer image after resolution conversion varies depending on the resolution of the input image.

  When the “auto” selection key is selected, the resolution conversion rate is set so that the capacity of each layer image after resolution conversion falls within a predetermined capacity range in accordance with the resolution of the input image. That is, regardless of the resolution of the input image, the resolution of each layer image is converted so that the capacity of each layer image after resolution conversion falls within a predetermined capacity range.

  That is, as illustrated in FIG. 7, the resolution conversion unit includes a resolution conversion table in which the resolution of each layer image after resolution conversion is set in a predetermined capacity range according to the resolution of the input image. Based on the resolution conversion table, the resolution of each layer image is converted.

  For example, if the resolution of the input image is 600 DPI, the resolution conversion rate of the character attribute layer image is 1/2, the resolution conversion rate of the character color layer image is 1/16, and the resolution conversion rate of the background color layer image is 1/4. If the resolution of the input image is 300 DPI, the resolution conversion rate of the character attribute layer image is 1, the resolution conversion rate of the character color layer image is 1/8, and the resolution conversion rate of the background color layer image is 1/2. If the resolution of the input image is 1200 DPI, the resolution conversion rate of the character attribute layer image is 1/4, the resolution conversion rate of the character color layer image is 1/32, and the resolution conversion rate of the background color layer image is 1. / 8 is set.

  As a result, the resolution of the character attribute layer image after resolution conversion is converted to a constant resolution of 300 DPI, the resolution of the character color layer image is 37.5 DPI, and the resolution of the background color layer image is 150 DPI.

  Therefore, when the “automatic” selection key is selected, the high-compression image file generated by the file generation unit 25 is always a file having a substantially constant capacity. By changing the resolution conversion rate according to the resolution of the input image and limiting the final file capacity to a certain level, it is possible to avoid an increase in load when transmitting a highly compressed image file to the outside by e-mail transmission or the like. become able to.

Hereinafter, another embodiment will be described.
In the above-described embodiment, the example in which both the character color layer image and the background color layer image are expanded by the resolution conversion unit 23 has been described. However, if at least the character color layer image is expanded, the character area becomes clear. The reproduced image can be obtained.

  In the above-described embodiment, the multifunction peripheral as an example of the image processing apparatus of the present invention has been described. However, the present invention is also effective for a copying machine, a printer, and the like which are image processing apparatuses other than the multifunction peripheral. The present invention can also be applied to an image processing apparatus configured by installing software that implements the image processing algorithm according to the present invention in a personal computer or the like.

  In the embodiment described above, the case where the high-compression image file is a PDF file has been described. However, the high-compression image file is not limited to an image file in the PDF format, and can be applied to an image file in another format.

  Each of the above-described embodiments is merely an example of the present invention, and the scope of the present invention is not limited by the description. The specific configuration of each part is appropriately changed and designed within the scope of the effects of the present invention. It goes without saying that you can do it.

1: Image processing device (multifunction machine)
14: Image processing unit 20: Compression / decompression image generation unit 21: Character attribute determination unit 22: Index value replacement processing unit 23: Resolution conversion unit 24: Image compression / decompression unit 25: File generation unit 26: Correction processing unit 27: Storage Processing unit 28: Image composition unit

Claims (5)

A character attribute discriminating unit for generating a character attribute layer image for identifying a character area and a background area other than the character area from the input image, and separating the character color layer image and the background color layer image from the input image based on the character attribute layer image A resolution conversion unit that converts each layer image to a predetermined resolution, an image compression unit that reversibly compresses the character attribute layer image whose resolution has been converted, and an irreversible compression of the character color layer image and the background color layer image, and compression An image processing apparatus including a file generation unit that generates a high-compression image file based on each layered image, and a storage unit that stores the image,
The character attribute layer image generated by the character attribute determination unit is reversibly compressed, the input image is irreversibly compressed and stored in the storage unit, and the character attribute layer image or the input image stored in the storage unit is decompressed. A storage processing unit;
A correction processing unit for correcting the character attribute layer image based on the input image and the character attribute layer image stored in the storage unit;
The resolution conversion unit separates the character color layer image and the background color layer image from the input image stored in the storage unit based on the corrected character attribute layer image, and converts each layer image to a predetermined resolution. Configured image processing apparatus.   The correction processing unit converts the input image stored in the storage unit into an image of any one of hue, saturation, and luminance, and corrects the character attribute layer image based on the converted input image. Image processing device. A character color pixel is extracted from an input image based on a mask image generated based on a character attribute layer image, and a pixel value that is the same as the index value indicating an ineffective pixel is extracted from the extracted character color pixel. An index value replacement processing unit for generating a character color layer image converted into
The resolution conversion unit divides the character color layer image processed by the index value replacement processing unit into pixel blocks having a predetermined number of pixels set based on the resolution conversion rate, and refers to the mask image to generate a mask image in the pixel block. If there is a pixel belonging to the pixel block, the average value of the pixel is the pixel value after resolution conversion, and if there is no pixel belonging to the mask image in the pixel block, the index value is the pixel value after resolution conversion. 3. The image processing apparatus according to claim 1, wherein the image processing device is configured to generate a reduced character color layer image repeatedly and to expand the contour portion of the reduced character color layer image determined based on the index value.   In the expansion process, when the expansion target pixel has an index value, an average value of pixels other than the pixel indicating the index value is used as the pixel value of the expansion target pixel among a predetermined number of pixel blocks having the expansion target pixel as a central pixel. When the pixel to be expanded is not an index value, among the predetermined number of pixel blocks having the expansion target pixel as the central pixel, the average value of the pixel values whose difference from the pixel value of the central pixel is within the predetermined threshold is expanded The image processing apparatus according to claim 3, wherein the image processing apparatus is configured to generate a pixel value of a target pixel.   The image processing apparatus according to claim 3 or 4, wherein the mask image is an image obtained by performing contraction processing on an outline of a character attribute layer image.