JP2007019673A - Image processing apparatus, image compression method, image compression program, and data structure of compressed image - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To make disturbance of ruled lines inconspicuous, even if image data are compressed. <P>SOLUTION: An MFP includes a scanner section 103 for inputting image data; a region discrimination section 11 for extracting a ruled line region, including ruled lines from the input image data; a ruled line binary section 14 that generates ruled line binary data, including at least center pixels of the ruled lines from image data in the ruled line region and generates interpolated binary data for interpolating the ruled line data from the image data in the ruled line region; a reversible control circuit power supply 15 for respectively compressing the ruled line binary data and the interpolated binary data to generate first and second compression data; a reverse compression section 13 and an irreversible compression section 17 for generating third and fourth compression data, by compressing a region excluding the ruled line region from the image data; and a composite section 18 for compositing the first, the second and the third compression data to generate composite data. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an image processing apparatus, an image compression method, an image compression program, and a compressed image data structure, and more particularly to an image processing apparatus, an image compression method, an image compression program, and a compressed image data structure.

  In recent years, with the progress of computerization of information, there is an increasing demand for storing or transmitting documents in electronic form instead of paper. Further, when transmitting digitized data, it is common to compress the data for transmission efficiency and transmit it. When digitizing a document, an original is read by a scanner, and image data is obtained. If the entire image data is compressed in an irreversible manner, characters cannot be read if the amount of data after compression is reduced, and if the amount is compressed so that the characters can be read, the amount of data after compression increases. For this reason, a technique for compressing image data using different compression methods for a character area and a background area is known (for example, Patent Document 1). In this case, the character area is binarized and then compressed in a reversible manner, and the background area is compressed in an irreversible manner. According to this technique, characters can be compressed so that they can be read even if the compression rate is increased.

On the other hand, when a ruled line is represented on the document, it is desirable that the ruled line has a clear outline. Therefore, it is preferable to binarize the ruled line included in the image data in the same manner as the character. However, since it is difficult to set the original so that the original does not tilt when the original is read by the scanner, the original is often read in a tilted state. For this reason, if the ruled line portion read by being tilted is binarized, rattling of the ruled line is emphasized. If the ruled line is irreversibly compressed in the same manner as the background, rattling is eliminated but the outline of the ruled line is blurred.
JP 2005-500709 A

  The present invention has been made to solve the above-described problems, and one of the objects of the present invention is to provide an image processing apparatus and an image compression device that can make the ruled lines less noticeable even when image data is compressed. A method and an image compression program are provided.

  Another object of the present invention is to provide a data structure of a compressed image capable of storing data obtained by compressing image data so that the restored ruled line is not noticeable.

  In order to achieve the above object, according to one aspect of the present invention, an image processing apparatus includes an input unit that inputs image data, and a ruled line region extraction unit that extracts a ruled line region including a ruled line from the input image data. First generation means for generating ruled line data including at least the center pixel of the ruled line from the image data of the ruled line area, second generation means for generating interpolation data for interpolating the ruled line data from the image data of the ruled line area, and the ruled line First compression means for compressing data and interpolation data to generate first and second compressed data, and second compression means for compressing an area excluding the ruled line area from image data to generate third compressed data And combining means for combining the first compressed data, the second compressed data, and the third compressed data to generate combined data.

  The ruled line includes a straight line, a curve, an underline, a frame, a table, a straight line portion of a large character, a diagonal line, and a broken line. According to the present invention, ruled line data including at least the center pixel of the ruled line is interpolated by the interpolation data. Since the ruled line data includes the center pixel of the ruled line, the amount of data can be reduced, but the probability of occurrence of rattling is high. This backlash is interpolated by the interpolation data. The image data in the ruled line area is compressed as it is without dividing the ruled line data including the center pixel of the ruled line and the interpolation data for interpolating it. An image processing apparatus that can be eliminated can be provided.

  Preferably, the first generation means includes first extraction means for extracting a pixel whose brightness is not more than a first threshold value from pixels constituting the ruled line included in the image data of the ruled line region, and the second generation means The image processing apparatus according to claim 1, further comprising: a second extracting unit that extracts a pixel whose brightness exceeds a first threshold value from pixels constituting the ruled line included in the image data of the ruled line region.

  According to this invention, a pixel whose brightness is equal to or less than the first threshold is extracted as the center pixel of the ruled line, and a pixel whose brightness exceeds the first threshold is extracted as a pixel for interpolating the center pixel of the ruled line. For this reason, the interpolation data includes pixels in the vicinity of the center pixel of the ruled line, and the center pixel of the ruled line can be interpolated with the pixels in the vicinity thereof.

  Preferably, the first extraction means includes first binarization means for binarizing the image data of the ruled line area with a first threshold value to generate ruled line data.

  According to this invention, since it binarizes, the data amount of ruled line data can be reduced.

  Preferably, the apparatus further includes first ruled line color determining means for determining a first ruled line color of the ruled line included in the ruled line data from the ruled line data and the image data, and the combining unit associates the first compressed data with the first ruled line color. Includes ruled line color association means.

  According to the present invention, since the first ruled line color determined from the ruled line data and the image data is associated with the first compressed data, the center pixel of the ruled line can be restored with the first ruled line color.

  Preferably, the second extraction means includes second binarization means for binarizing the image data of the ruled line area inside and outside the first threshold value and the second threshold value to generate the interpolation data. Including.

  According to the present invention, since binarization is performed, the amount of interpolation data can be reduced.

  Preferably, a first ruled line color determining means for determining a first ruled line color of the ruled line included in the ruled line data from the ruled line data and the image data, and a second ruled line color included in the interpolation data from the interpolation data and the image data. And a second ruled line color determining unit that associates the first compressed data with the first ruled line color and includes a ruled line color associating unit that associates the second compressed data with the second ruled line color.

  According to this invention, the first ruled line color determined from the ruled line data and the image data is associated with the first compressed data, and the second ruled line color and the second compressed data determined from the interpolation data and the image data. Are associated with each other, the center pixel of the ruled line can be restored with the first ruled line color, and the pixel that interpolates the center pixel of the ruled line can be restored with the second ruled line color.

  Preferably, the first generation means includes binarization means for binarizing the image data of the ruled line area, and resolution conversion means for converting the binarized data into ruled line data having a low resolution, and the second generation The means includes binarization means for binarizing the image data of the ruled line area, and difference extraction means for generating interpolation data by removing ruled line components included in the ruled line data from the binarized data.

  According to the present invention, ruled line data is generated by converting the binarized data of the ruled line area image data to a low resolution, and the ruled line component included in the ruled line data is converted from the binarized image data of the ruled line area. The interpolation data is generated by removing. Since the resolution of the ruled line data is reduced, the amount of data is reduced. However, the ruled line becomes unstable. Since the ruled line component included in the ruled line data is removed from the binarized data, the interpolation data reduces the data amount. The ruled line shape of the image data in the ruled line area can be restored by the ruled line data and the interpolation data. As a result, it is possible to provide an image processing apparatus that can make the ruled lines less noticeable even when the image data is compressed.

  Preferably, the image processing apparatus further includes a dividing unit that collects ruled lines whose hue and saturation sets are approximated from the image data of the ruled line area, and divides the image data of the ruled line area into sets of hue and saturation to generate a plurality of divided data. Each of the first and second generation means receives a plurality of divided data, and extracts ruled line data and interpolation data for each set of hue and saturation.

  According to the present invention, even when color image data is compressed, it is possible to make the ruled lines less noticeable.

  According to another aspect of the present invention, an image compression method includes a step of inputting image data, a step of extracting a ruled line region including a ruled line from the input image data, and a center pixel of the ruled line from the image data of the ruled line region Generating ruled line data including at least a ruled line data, a step of generating interpolation data for interpolating the ruled line data from the image data of the ruled line region, and compressing the ruled line data and the interpolated data, respectively. A step of generating, a step of compressing a region excluding the ruled line region from the image data to generate third compressed data, a first compressed data, a second compressed data, and a third compressed data being combined and combined Generating data.

  According to the present invention, it is possible to provide an image compression method capable of making ruled lines less noticeable even when image data is compressed.

  According to still another aspect of the present invention, an image compression program includes a step of inputting image data, a step of extracting a ruled line area including a ruled line from the input image data, and the center of the ruled line from the image data of the ruled line area. A step of generating ruled line data including at least pixels; a step of generating interpolation data for interpolating ruled line data from image data of a ruled line region; and the first and second compressed data by compressing the ruled line data and the interpolation data, respectively Generating a third compressed data by compressing a region excluding the ruled line region from the image data, combining the first compressed data, the second compressed data, and the third compressed data, Generating combined data.

  According to the present invention, it is possible to provide an image compression program capable of making the ruled lines less noticeable even when image data is compressed.

  According to still another aspect of the present invention, the data structure of the compressed image is a first compressed data obtained by extracting a center pixel of a ruled line from a ruled line area included in the image data and compressing ruled line data including at least the center pixel. Interpolated data for interpolating ruled line data from the ruled line area included in the image data, second compressed data obtained by compressing the interpolated data, and third compressed data obtained by excluding the ruled line area from the image data. Compressed data.

  According to the present invention, it is possible to provide a data structure of a compressed image that can store data obtained by compressing image data so that the restored ruled line is not noticeable.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description, the same parts are denoted by the same reference numerals. Their names and functions are also the same. Therefore, detailed description thereof will not be repeated.

<First Embodiment>
FIG. 1 is a block diagram showing a hardware configuration of the MFP according to the first embodiment of the present invention. Referring to FIG. 1, an MFP (Multi Function Peripheral) 100 includes a scanner unit 103, an image forming unit 106 that forms an image on a recording medium such as paper, and a CPU (Central Processing Unit) each connected to a bus 111. ) 101, storage unit 102, input image processing unit 104 that processes image data output from the scanner unit 103, output image processing unit 105 that performs processing suitable for output to image data, and network interface (I) / F) 107, modem 108, operation panel 109, and card I / F 110.

  CPU 101 controls the entire MFP 100. The storage unit 102 includes a semiconductor memory such as a ROM (Read Only Memory), a RAM (Random Access Memory), and an EEPROM (Electrically Erasable / Programmable ROM), and a magnetic storage device such as a hard disk drive (HDD). The CPU 101 loads the image compression program recorded in the ROM into the RAM and executes it.

  The scanner unit 103 includes a document table and a photoelectric conversion device such as a CCD (Charge Coupled Device). The scanner unit 103 optically reads a document placed on the document table and inputs image data as electronic data to the input image processing unit 104. Output. The input image processing unit 104 performs color conversion processing, color correction processing, and the like on the input image data, and stores the processed image data in the storage unit 102.

  The output image processing unit 105 reads the image data from the storage unit 102, performs screen control processing, smoothing processing, and pulse width modulation (PWM), and outputs the processed image data to the image forming unit 106. The image forming unit 106 is a laser printer, and visualizes the image data input from the output image processing unit 105 on a recording medium such as paper. In the case of color printing, the image forming unit 106 forms an image with toners of four colors, yellow, magenta, cyan, and black. The image forming unit 106 may be an ink jet printer.

  An IC card 110A is attached to the card I / F 110. The CPU 101 can access the IC card 110 </ b> A via the card I / F 110. Network I / F 107 connects MFP 100 to network 112. The network I / F 107 outputs data to the network according to a predetermined communication protocol, and receives data from the network 112 according to a predetermined communication protocol. The CPU 101 can communicate with other computers connected to the network 112 via the network I / F 107. For this reason, for example, image data can be transmitted and received using e-mail.

  Modem 108 connects MFP 100 to telephone line 113. The modem 108 enables facsimile communication of data according to a facsimile communication protocol. The modem 108 stores the data received by facsimile in the HDD of the storage unit 102. The modem 108 also includes an NCU (Network Control Unit) to enable communication with a computer connected to the telephone line 113. Data received from another computer by the modem 108 is stored in the HDD of the storage unit 102.

  Operation panel 109 includes an input unit 109A and a display unit 109B. Input unit 109 </ b> A is an input device such as a touch panel, a keyboard, or a mouse for receiving an operation input by a user of MFP 100. The display unit 109B is a liquid crystal display device or an organic EL (Electro-Luminescence) display panel. When a touch panel made of a transparent member is used as the input unit 109A, an instruction of a button displayed on the display unit 109B can be detected by placing the touch panel on the display unit 109B. Thereby, various operations can be input.

  Note that the image compression program executed by the CPU 101 is not limited to the one stored in the ROM, but may be stored in the EEPROM. If stored in the EEPROM, the image compression program can be rewritten or additionally written. For this reason, another computer connected to the network 112 can rewrite the image compression program stored in the EEPROM of the MPP 100, or can additionally write a new image compression program. Further, MFP 100 can download an image compression program from another computer connected to network 112 and store the image compression program in the EEPROM. Further, the image compression program may be stored in the IC card 110A, and the image compression program recorded in the IC card 110A attached to the card I / F 110 may be loaded into the RAM and executed.

  The recording medium for storing the image compression program is not limited to the IC card 110A, but a flexible disk, cassette tape, optical disk (CD-ROM (Compact Disc-ROM) / MO (Magnetic Optical Disc / MD (Mini Disc) / MD) / A medium carrying a fixed program such as a semiconductor memory such as a DVD (Digital Versatile Disc), an optical card, a mask ROM, an EPROM, or an EEPROM may be used.

  The program here includes not only a program directly executable by the CPU 101 but also a program in a source program format, a compressed program, an encrypted program, and the like.

  FIG. 2 is a functional block diagram showing an outline of the functions of the MFP according to the first embodiment. Referring to FIG. 2, image data (image data) read and output by the scanner unit 103 is processed and input to the CPU 101 by the input image processing unit 104. The CPU 101 includes a preprocessing unit 10, an area determination unit 11 to which image data processed by the preprocessing unit 10 is input, and a character binarization unit to which image data of a character area output from the region determination unit 11 is input. 12, a reversible compression unit 13 connected to the character binarization unit 12, a ruled line binarization unit 14 to which image data of a ruled line region output from the region determination unit 11 is input, and a ruled line binarization unit 14 The connected lossless compression unit 15, the resolution reduction unit 16 to which the image data of the background region output from the region determination unit 11 is input, the lossy compression unit 17 connected to the resolution reduction unit 16, and the lossless compression Parts 13 and 15 and a connecting part 18 connected to the lossy compression part 17.

  The image data input to the CPU 101 is not limited to the image data output from the scanner unit 103, and may be image data received from another computer. In this case, the image data is input via the network 112 or the telephone line 113 or via the IC card 110A.

  The preprocessing unit 10 performs image format conversion processing, resolution conversion processing, and background removal processing on input image data. These processes are processes for converting the image data into data suitable for processing by the region discriminating unit 11, and are not limited to these processes, and these processes may not be executed. May be executed. What is necessary is just to combine suitably according to the performance of the scanner part 103, or the input image data.

  The area determination unit 11 receives preprocessed image data. The image data input here is full-color image data having three RGB planes in which one pixel is 24 bits. The area determination unit 11 converts full-color image data into image data in which one pixel has an 8-bit brightness plane. Hereinafter, the image data simply indicates image data having a brightness plane, and full-color image data having an RGB plane is referred to as “RGB image data”.

  The area discriminating unit 11 extracts each area of a figure, a photograph, and a background from the image data. The figure, photograph, and background areas are output to the resolution reduction unit 16 as background areas. Further, an edge image is generated from the image data, and a ruled line area including a ruled line is extracted from the edge image. The ruled line includes a straight line, a curve, an underline, a frame, a table, a straight line portion of a large character, a diagonal line, and a broken line. The ruled line area is output to the ruled line binarization unit 14. Then, a character block is detected from the edge image obtained by removing the ruled line area. Further, the character color is calculated from the pixel value of the detected RGB image data of the character block. Character blocks having similar colors in the vicinity are integrated from the distance between the blocks of the plurality of character blocks and the difference between the calculated character colors to form a character region. This is because the amount of data can be reduced by reducing the number of character blocks. The character area is output to the character binarization unit 12.

The character binarization unit 12 receives a character area and image data. The character binarization unit 12 binarizes the character area of the image data. And the color of a character area is determined from the pixel value in the RGB image data of the pixel made into the character in a character area. The average value or median value of the RGB colors of the pixels that are the characters of the RGB image data is determined as the color of the character region. The character binarization unit 12 outputs the binarized data of the character region to the lossless compression unit 13 and outputs the color of the character region to the combining unit 18. The lossless compression unit 13 losslessly compresses the binarized data input from the character binarization unit 12. Here, as a lossless compression method, MMR (Modified
Modified Read) encoding is used. The lossless compression unit 13 outputs third compressed data obtained by lossless compression of the binarized data to the combining unit 18.

  A ruled line region, image data, and RGB image data are input to the ruled line binarization unit 14. The ruled line binarizing unit 14 obtains the hue and saturation of a plurality of input ruled line areas from the RGB image data, and generates divided data obtained by collecting and integrating the ruled line areas having similar hues and saturations. Thereby, the divided data of the number of sets of hue and saturation is generated. The ruled line binarization unit 14 interpolates ruled line data and ruled line binarized data including at least the center pixel of the ruled line from the image data of each of the plurality of ruled line areas included in the divided data, for each of the plurality of divided data. Interpolated binary data is generated and output to the lossless compression unit 15. The ruled line binarization unit 14 determines the first ruled line color from the pixel values of the RGB image data specified by the ruled line binarized data, outputs the first ruled line color to the combining unit 18, and specifies the RGB specified by the interpolated binarized ruled line data. The second ruled line color is determined from the pixel value of the image data and output to the combining unit 18.

  The lossless compression unit 15 receives ruled line binarized data and interpolated binarized data. The ruled line binarized data and the interpolated binarized data are input for each set of hue and saturation (each divided data). The reversible compression unit 15 outputs the first compressed data obtained by reversibly compressing the ruled line binarized data input for each set of hue and saturation to the combining unit 18 and inputs the first compression data for each set of hue and saturation. The second compressed data obtained by reversibly compressing the interpolated binary data is output to the combining unit 18.

  The resolution reduction unit 16 receives the background area and the RGB image data, and reduces the resolution of the background area of the RGB image data. This is to reduce the amount of data and increase the compression efficiency by lowering the resolution. The resolution reduction unit 16 outputs the background area of the RGB image data whose resolution has been reduced to the lossy compression unit 17. The irreversible compression unit 17 performs irreversible compression on the RGB image data of the background area with the resolution reduced from the resolution reduction unit 16. Here, JPEG (Joint Photographic Experts Group) encoding is used as a lossy compression method. Note that other than JPEG may be used as long as it is an irreversible compression method.

  The combining unit 18 includes the third compression data from the lossless compression unit 13, the first and second compression data from the lossless compression unit 15, the fourth compression data from the lossy compression unit 17, and the character from the character binarization unit 12. As the color of the area, the first ruled line color and the second ruled line color are input from the ruled line binarization unit 14. The combining unit 18 includes first related data in which the first compressed data and the first ruled line color are associated, second related data in which the second compressed data and the second ruled line color are associated, the third compressed data, and the character area. The third related data in which the colors are associated with the first to fourth compressed data are combined to generate combined compressed data for output. The output may be either storage in the storage unit 102 or the IC card 110A, or transmission from the network I / F 107 or the modem 108 to another computer. The combined data may be in the format of PDF (Portable Document Format).

  FIG. 3 is a functional block diagram illustrating detailed functions of the ruled line binarization unit in the first embodiment. Referring to FIG. 3, ruled line binarization unit 14 binarizes image data of a ruled line area with a first threshold value and extracts ruled line binarized data, A second binarization unit 23 that binarizes the image data of the ruled line area inside and outside the first and second threshold values and extracts interpolated binarized data, and RGB specified by the ruled line binarized data A first ruled line color determination unit 22 that determines the first ruled line color from the pixel values of the image data, and a second ruled line color determination that determines the second ruled line color from the pixel values of the RGB image data specified by the interpolated binarized data. Part 24.

  Here, a case will be described in which a ruled line area included in one piece of divided data specified by a set of hue and saturation is input. Actually, when there are a plurality of divided data, ruled line binarized data and interpolated binarized data are generated for each of the plurality of divided data, and the first ruled line color and interpolated binary corresponding to the ruled line binarized data are generated. A second ruled line color corresponding to the digitized data is determined. In other words, the ruled line binarization unit 14 generates ruled line binarized data and interpolated binarized data for each set of hue and saturation, and the first ruled line color and interpolation corresponding to the ruled line binarized data. A second ruled line color corresponding to the binarized data is determined.

  Here, determination of the first threshold value and the second threshold value will be described. FIG. 4 is a diagram showing a brightness histogram in the vicinity of the ruled line region of the image data. Referring to FIG. 4, the horizontal axis indicates a dimension in which the brightness decreases from the right side to the left side of the drawing. The vertical axis indicates the number of pixels. The mountain on the left side (low lightness side) of the histogram corresponds to the ruled line portion, and the mountain on the right side (high lightness side) corresponds to the background region. A first threshold value is set at a position with higher brightness than the peak of the left mountain. A pixel having a lightness lower than the first threshold corresponds to a portion where the ruled line is dark, in other words, a pixel near the center of the ruled line. The second threshold value has a higher brightness than the first threshold value. Pixels having brightness between the first threshold value and the second threshold value mainly correspond to pixels in a portion where the ruled line density is low, in other words, an image near the edge of the ruled line.

The first threshold value and the second threshold value are determined by the following procedure from the ruled line region histogram.
(1) In order to remove noise in the histogram, the number of lightness gradations is reduced from 256 gradations to 64 gradations.
(2) The peak value of the ruled line region of the histogram is obtained.
(3) Of the two brightness values corresponding to 10% of the peak value of the number of pixels in the histogram, a brightness value higher than the brightness value of the peak value is determined as the second threshold value.
(4) The lightness that can include the upper 70% of the number of pixels having the lightness equal to or higher than the second threshold value in the ruled line region is determined as the first threshold value.

  The parameters that define the first threshold value and the second threshold value may be appropriately changed according to the shape of the histogram, the number of pixels to be processed, and the like.

  Returning to FIG. 3, the first binarization unit 21 binarizes the image data of the ruled line area with the first threshold value to generate ruled line binarized data. In other words, the first binarization unit 21 sets the pixel whose brightness is equal to or less than the first threshold among the pixels of the image data in the ruled line region to “1” and sets the others to “0”. Generate data. The ruled line binarized data is binary data in which the pixel near the center of the ruled line in the image data of the ruled line area is “1” and the others are “0”. For this reason, when the image data in the ruled line area includes a tilted ruled line or a curved line, the ruled line is not a straight line, and the ruled line is conspicuous.

  RGB image data and ruled line binarized data are input to the first ruled line color determination unit 22. The first ruled line color determination unit 22 determines the average value or the median value of the RGB values of the pixels of the RGB image data corresponding to the pixel whose pixel value is “1” in the ruled line binarized data as the first ruled line color. To do. The first ruled line color is the ruled line color of the pixel whose pixel value is “1” in the ruled line binarized data. The first ruled line color determining unit 22 outputs the first ruled line color to the combining unit 18.

  The second binarization unit 23 binarizes the image data of the ruled line area inside and outside the first and second threshold values, and generates interpolated binarized data. In other words, the second binarizing unit 23 sets the pixels whose brightness is higher than the first threshold value and lower than or equal to the second threshold value among the pixels of the image data in the ruled line region to “1”, and the other “ Interpolated binarized data set to “0” is generated. For this reason, in the image data of the ruled line region, the interpolated binarized data is binary data in which a pixel having high brightness near the edge of the ruled line is set to “1” and the others are set to “0”. Such a pixel often exists in the vicinity of a shaky portion of a ruled line binarized by the first binarization unit 21.

  The second ruled line color determination unit 24 receives RGB image data and interpolated binary data. The second ruled line color determining unit 24 determines, as the second ruled line color, the average value or the median value of the RGB values of the pixels of the RGB image data corresponding to the pixel whose pixel value is “1” in the interpolated binarized data. To do. This second ruled line color is the ruled line color of the pixel having the pixel value “1” in the interpolation binarized data. The second ruled line color determining unit 24 outputs the second ruled line color to the combining unit 18.

  FIG. 5 is a diagram showing the relationship between ruled line binarized data and interpolated binarized data. FIG. 5A shows ruled line binarized data, and FIG. 5B shows ruled line binarized data and interpolated binarized data. Referring to FIG. 5A, when the document is tilted and read by the scanner, the ruled lines are not perpendicular or parallel to the main scanning direction of the scanner. The ruled line binarized data is image data in which the pixel corresponding to the vicinity of the center of the ruled line is set to “1”. Therefore, the line is not a straight line as shown in FIG. To do.

  With reference to FIG. 5B, the ruled lines included in the ruled line binarized data are indicated by fine hatching, and the ruled lines included in the interpolated binarized data are indicated by rough hatching. The ruled line included in the interpolated binarized data is located in the vicinity of the shaky part of the ruled line included in the ruled line binarized data. Since the lightness of the ruled line included in the interpolated binarized data is higher than the lightness of the ruled line included in the ruled binarized data, the shaky part is visually relaxed by the ruled line included in the interpolated binarized data.

  FIG. 6 is a flowchart showing the flow of the area-specific compression processing executed in the MFP. Referring to FIG. 6, in the region-by-region compression processing, first, the process is in a standby state until RGB image data is input (NO in step S01), and when RGB image data is input, the process proceeds to step S02. In step S02, the brightness of each pixel of the RGB image data is calculated, and the RGB image data is converted into image data in which each pixel has an 8-bit brightness plane. The reason why the lightness is calculated and converted into image data is that the lightness is processed in order to improve the determination accuracy and reduce the calculation load in the region determination processing to be executed later.

  In the next step S03, the image data is smoothed. Thereby, noise is removed from the image data, and erroneous detection can be reduced. Then, a photographic block is detected (step S04). In the detection of the photographic block, the position of the rectangular photographic block included in the image data is detected by binarizing the image data and labeling the binarized data. The reason why the photo block is rectangular is that the photo determination processing executed thereafter determines for each rectangle.

  In the next step S05, it is determined whether the photographic block is a figure region, a photographic region, a base region, or another region according to the characteristics of the image in the photographic block. The determination result in step S05 is used for determination of characters and ruled lines. For example, in order to determine a character or ruled line in a photographic area as a character or ruled line, it is necessary to use a threshold value different from the threshold value used for determining a character or ruled line on the white background. This is because the ruled line separation process can be used not only to detect the ruled lines of the entire image but also to perform ruled line detection specialized for the area determined as the figure area.

  Then, it is determined whether or not the next photo block exists (step S06). If it exists, the process returns to step S05, and if it does not exist, the process proceeds to step S07.

  In the next step S07 and step S08, the RGB image data is converted into image data having a brightness plane and smoothed in the same manner as in step S02 and step S03. Then, an edge is extracted from the smoothed image data (step S09). Edge extraction is performed, for example, by filtering image data using a primary differential filter or a secondary partial filter to generate an edge image. The reason for extracting the edge without using the binarized data as in the photographic area is to detect characters and inverted characters in the high brightness area.

  Then, a ruled line is extracted using the edge image. The ruled line is extracted not only for the entire image, but also only for the area determined as the figure area in step S05 and only the area determined as the photo area. By extracting ruled lines from the entire image, it is possible to extract underlines and the like that modify characters. Further, a ruled line can be accurately extracted from a region determined as a figure attribute or a photo region. Furthermore, ruled lines can be detected in consideration of vertical and horizontal continuity.

  Here, the ruled line separation process will be described in detail. FIG. 7 is a flowchart showing a ruled line separation process executed in step S10 of FIG. Referring to FIG. 7, first, the inclination of image data is detected (step S41). In order to detect the inclination, the edge image is first reduced in resolution. The resolution is reduced to an appropriate resolution according to the processing speed and detection accuracy. Then, an inclination detection area is extracted. Since the inclination detection has high accuracy in a region with many characters, a region with many edges is extracted as a detection region. In the detection area, a histogram of edge pixels is created for each line. The detection area is rotated, and a histogram is created at each rotation angle. Then, the rotation angle when the shape of the histogram becomes the shape when the document is not inclined is detected as the inclination angle.

  In the next step S42, the edge image is rotated by the inclination angle detected in step S41. Then, a ruled line is detected (step S43). To detect ruled lines, first create a histogram in the row direction of the rotated edge image, compare the shape and peak value of the histogram with the threshold value, and detect the peak value line as a ruled line area if a predetermined condition is met. To do. The detected ruled line area is separated from the edge image. If there are a plurality of rows of peak values, the plurality are detected as ruled line regions. Next, a histogram in the column direction is created. Then, the peak value column is separated as a ruled line region. If there are a plurality of columns of peak values, the plurality are separated. The detected ruled line area is separated from the edge image. Then, the detected ruled line area and the edge image from which the ruled line area is separated are rotated by the same angle in the opposite direction to that rotated in step S42.

  Returning to FIG. 6, in step S11, a character block is detected from the edge image from which the ruled lines are separated. By labeling the edge image, a character block as large as one word is detected. By making the character block about one word in size, the processing time can be reduced and the character determination system can be improved. The size of the character block is not limited to this, and may be larger, for example, about one line, or smaller, for example, about one character.

  Then, it is determined using the image data of the character block whether or not the character block has character characteristics (step S12). That is, it is determined whether or not it has character characteristics using the brightness of the pixels included in the character block in the image data. A character block having character characteristics is regarded as a character area, otherwise the character block is not regarded as a character area.

  In the next step S13, halftone dot removal processing is performed on the image data in the character area. This is to prevent the halftone dot from being binarized because when the halftone dot exists in the background of the character area, the halftone dot becomes noise when binarized. In the next step S14, the image data of the character area is binarized (step S14). Thereby, the character and the background are separated, and the data amount is reduced. Pixels that are converted to characters by binarization are set to “1”, and pixels that are set to background are set to “0”.

  Then, the character color is determined (step S15). The character color is determined for each character region, and one color is determined for the character region. Specifically, the average value or median value of the pixel values in the RGB image data of the pixels set as characters in the character area is used as the character color of the character area. Based on the distance between the character areas of the plurality of character areas and the difference between the determined character colors, the character areas in the vicinity and similar colors are integrated to reduce the number of character areas. Thereby, the data amount is reduced.

  Next, it is determined whether or not there is a character block to be processed (step S16). If there is such a character block, the process returns to step S12, and if not, the process proceeds to step S17. In step S17, an image of the background region obtained by extracting the background region from the RGB image data by using the region excluding the character region determined as the character in step S12 and the ruled line region extracted in step S10 from the entire image data as a background region. Data (background area data) is generated. Then, the background area data is converted to a low resolution (step S18) and compressed by the lossy compression method (step S19). As a result, fourth compressed data is generated. Here, the image is compressed by the JPEG method.

  In step S20, character area data in which character areas of similar colors are integrated is generated (step S20). All the integrated character area data has the same color. The same color may be the average or median of the RGB values of the pixels made into characters. Here, it is assumed that character area data of each color of black, red, blue and green is generated. Then, the character area data generated for each color is reversibly compressed (step S21). Thereby, the 3rd compression data according to color are produced | generated. Here, the third compressed data obtained by compressing the black character region, the third compressed data obtained by compressing the red character region, the third compressed data obtained by compressing the blue character region, and the third compressed data obtained by compressing the green character region. 3 compressed data is generated.

  In the next step S22, ruled lines of the same color are integrated (step S22). From the RGB values of the pixels included in the ruled line area detected in step S13, a set of hue and saturation for each ruled line area is determined. Then, the ruled line regions that approximate the hue and saturation pairs are integrated. Thereby, a ruled line area is specified for each set of hue and saturation. Here, for the sake of explanation, a case where the set of hue and saturation in the ruled line region is four colors of black, red, green, and blue will be described as an example. Here, the ruled line region is specified for each set of hue and saturation, but the process from step S23 to step S28 may be executed for each ruled line separated in step S10.

  Steps S <b> 23 to S <b> 28 are performed for all of the ruled line regions having the same hue and saturation set among the ruled line regions integrated for each set of hue and saturation. In step S23, the image data of the ruled line area is binarized with the first threshold value. Thereby, ruled line binarized data is generated. The ruled line binarized data is data in which, in the image data of the ruled line area, a pixel whose brightness is lower than the first threshold value is set to “1”, and other pixels are set to “0”. In the ruled line binarized data, a pixel having a pixel value “1” indicates a ruled line. Then, the generated ruled line binarized data is reversibly compressed (step S24). As a result, first compressed data obtained by reversibly compressing the ruled line binarized data is generated. Then, the first ruled line color is determined from the pixel values of the RGB image data specified by the ruled line binarized data (step S25). The first ruled line color is, for example, the average value or the median value of the RGB values of the pixels of the RGB image data corresponding to the pixels whose pixel value is “1” in the ruled line binarized data.

  In step S26, the image data of the ruled line area is binarized inside and outside the first threshold value and the second threshold value. Thereby, interpolated binary data is generated. The interpolated binarized data is data in which, in the image data of the ruled line region, pixels that are higher than the first threshold value and lower than or equal to the second threshold value are set to “1”, and other pixels are set to “0”. In the interpolated binarized data, the pixel having the pixel value “1” indicates a ruled line. Then, the generated interpolated binary data is reversibly compressed (step S27). As a result, second compressed data obtained by reversibly compressing the interpolated binary data is generated. Then, the second ruled line color is determined from the pixel value of the RGB image data specified by the interpolation binary data (step S25). The second ruled line color is, for example, the average value or median value of the RGB values of the pixels of the RGB image data corresponding to the pixel whose pixel value is “1” in the interpolated binarized data.

  In the next step S29, it is determined whether or not there is a ruled line region of a set of unprocessed hue and saturation. If such a ruled line area exists, the process returns to step S23, and if not, the process proceeds to step S30. As a result, the processing from step S23 to S28 is executed for all the ruled line regions integrated for each combination of hue and saturation in step S22. Accordingly, the ruled line binarized data and the interpolated binarized data are generated by the number of divided data obtained by integrating the ruled line regions for each set of hue and saturation and are reversibly compressed. In other words, the first compressed data and the second compressed data are generated for each set of hue and saturation. Here, the set of hue and saturation is black, red, green, and blue. Therefore, for each of these four sets of divided data, the first compressed data, the first ruled line color, the second compressed data, and the first Two ruled line colors are required.

  In step S30, the first compressed data for each set of hue and saturation, the second compressed data for each set of hue and saturation, the third compressed data for each color, and the fourth compressed data are integrated into one data and To do. At this time, the related data that associates the first compressed data with the first ruled line color, the related data that associates the second compressed data with the second ruled line color, the third compressed data, and the color of the character area are associated with each other. Combined with related data. Here, the hue and saturation pairs are black, red, green, and blue, respectively, and the first compressed data and the second compressed data are generated. Therefore, the hue and saturation pairs are black, red, green, and blue, respectively. And the related data associating the first compressed data with the first ruled line color, the related data associating the second compressed data with the second ruled line color, and the black color and the third compressed data. , Related data that associates the red color with the third compressed data, related data that associates the blue color with the third compressed data, and the green color and the third compressed data. The related data associated is stored.

  FIG. 8 is a diagram illustrating a data structure of combined data. Referring to FIG. 8, the combined data includes a header, a body, a cross reference table, and a trailer. The body includes character information including a creation date, data blocks for each of the first to last pages, a child page dictionary corresponding to each page, a parent page dictionary, and a catalog dictionary.

  Each of the data blocks of the first page to the last page includes fourth compressed data obtained by compressing the background area, third compressed data obtained by compressing the black character area, third compressed data obtained by compressing the red character area, and blue The third compressed data obtained by compressing the character area, the third compressed data obtained by compressing the green character area, the first compressed data of the first ruled line color (dark black) among the combination of hue and saturation, and Second compressed data of the second ruled line color (light black) in which the set of hue and saturation is black; and first compressed data of the first ruled line color (dark red) in which the set of hue and saturation is red; Second compressed data of the second ruled line color (light red) out of the hue and saturation set of red; and first compressed data of the first ruled line color (dark green) out of the hue and saturation set of green; The second compressed data of the second ruled line color (light green) out of the hue and saturation set in green, and the hue And the first compressed data of the first ruled line color (dark blue) out of the set of blue and saturation, the second compressed data of the second ruled line color (light blue) out of the set of hue and saturation, and the layer Information. The layer information includes four related data in which the first ruled line color and the first compressed data are associated with each of the hue and saturation pairs of black, red, green, and blue, and the hue and saturation pairs of black, red, Four related data that associate the second ruled line color and the second compressed data with respect to each of green and blue, and further, related data that associates the black color with the third compressed data, the red color, and the third color Related data that associates the compressed data with each other, associated data that associates the blue color with the third compressed data, and relevant data that associates the green color with the third compressed data.

  FIG. 9 is another diagram showing the relationship between ruled line binarized data and interpolated binarized data. FIG. 9A shows ruled line binarized data, and FIG. 9B shows ruled line binarized data and interpolated binarized data. With reference to FIG. 9A, when the document includes curved ruled lines, the ruled lines are not perpendicular or parallel to the main scanning direction of the scanner. Since the ruled line binarized data is image data in which the pixel corresponding to the vicinity of the center of the ruled line is “1”, the curve is a set of a plurality of short straight lines as shown in FIG. Exists.

  With reference to FIG. 9B, the ruled lines included in the ruled line binarized data are indicated by fine hatching, and the ruled lines included in the interpolated binarized data are indicated by rough hatching. The ruled line included in the interpolated binarized data is located in the vicinity of the shaky part of the ruled line included in the ruled line binarized data. Since the lightness of the ruled line included in the interpolated binarized data is higher than the lightness of the ruled line included in the ruled binarized data, the shaky part is visually relaxed by the ruled line included in the interpolated binarized data.

  As described above, in the MFP according to the first embodiment, the ruled line region in the RGB image data is converted into the ruled line binarized data of the first ruled line color for each set of hue and saturation, and the brightness of the ruled line color is higher than that of the first ruled line color. The compressed data is divided into interpolation binary data having a high second ruled line color. The shakiness of the ruled line generated by reducing the data amount of the ruled line binarized data of the first ruled line color is interpolated with the interpolated binarized data of the second ruled line color. For this reason, even if the image data is compressed, the rattling of the ruled lines can be made inconspicuous.

<Second Embodiment>
Next, an image processing apparatus according to the second embodiment will be described. Since the hardware configuration of the image processing apparatus in the second embodiment is the same as that of the image processing apparatus in the first embodiment, description thereof will not be repeated here. The image processing apparatus according to the second embodiment makes the ruled line rattle generated when the ruled line is reduced in resolution and compressed to be inconspicuous.

  FIG. 10 is a functional block diagram showing an outline of the functions of the MFP in the second embodiment. Referring to FIG. 10, the difference from FIG. 2 is a ruled line binarization unit 14A. Since other functions are the same as those shown in FIG. 2, the description will not be repeated here.

  FIG. 11 is a functional block diagram illustrating detailed functions of the ruled line binarization unit according to the second embodiment. Referring to FIG. 11, ruled line binarization unit 14A in the second embodiment receives a ruled line area and image data, and binarizes 31 to binarize the image data of the ruled line area. A ruled line color determining unit 34 that inputs the binarized data and the RGB image data and determines the ruled line color from the pixel value of the RGB image data specified by the binarized data, and the binarized data. A resolution conversion unit 32 for converting to a lower resolution and outputting ruled line binarized data, binarized data and ruled line binarized data are input, and the binarized data is converted into ruled line binarized data. A difference extraction unit 33 that removes ruled line components and outputs interpolated binary data.

  The binarization unit 31 receives the ruled line area and the image data, and binarizes the image data in the ruled line area. The image data is high-resolution data, and is 600 dpi (dot per inch) here. Since the ruled line area is 300 dpi, the resolution of the ruled line area is converted to 600 dpi, and the image data of the ruled line area is specified. The ruled line color determining unit 34 calculates the average value or median value of the RGB values of the pixels of the RGB image data corresponding to the pixel (ruled line) whose pixel value is “1” in the binarized data as the first ruled line color. Determine as. The ruled line color is set to the ruled line color of the pixel having the pixel value “1” in the ruled line binarized data, and the ruled line color of the pixel having the pixel value “1” in the interpolated binarized data. The ruled line color determining unit 34 outputs the ruled line color to the combining unit 18.

  The resolution conversion unit 32 receives the binarized ruled line area image data, converts the binarized data to a low resolution, and outputs the ruled line binarized data to the lossless compression unit 15. Here, the ruled line binarized data is 300 dpi. The difference extraction unit 33 receives data obtained by binarizing the image data in the ruled line region and data obtained by restoring the resolution of the ruled line binarized data, and data and ruled lines obtained by binarizing the image data in the ruled line region. The difference between the binarized data and the original resolution is output as interpolated binarized data. In other words, interpolated binarized data is extracted from the data obtained by binarizing the image data of the ruled line region, by removing the components that are ruled by the ruled line binarized data.

  Data obtained by binarizing the image data in the ruled line area has the same resolution of 600 dpi as that of the image data, but the ruled line binarized data is converted into 300 dpi by the resolution conversion unit 32. On the other hand, the interpolated binarized data is a difference between the data (600 dpi) obtained by restoring the resolution of the ruled line binarized data (300 dpi) from the binarized data (600 dpi) of the image data. For this reason, the interpolated binarized data remains at the same 600 dpi resolution as the image data. That is, the interpolated binarized data includes only the ruled line portion that has been deleted by the resolution conversion unit 32 reducing the resolution among the ruled lines included in the ruled line region.

  Note that the resolution is shown by way of example and is not limited to this. In addition, since the difference extraction unit 33 here reduces the vertical and horizontal resolutions by half by the resolution conversion unit 32, the region of 2 × 2 size in the vertical and horizontal sizes among the binarized image data of the ruled line region. The pixel value of a pixel in which all the pixels are “1” may be set to “0”. Here, the ruled line binarized data is reduced in resolution after binarizing the image data of the ruled line area, but may be binarized after reducing the resolution of the image data in the ruled line area. .

  Returning to FIG. 10, the first compression data and the second compression data are input to the combining unit 18 from the lossless compression unit 15, and the ruled line color is input from the ruled line binarization unit 14 </ b> A. The combining unit 18 generates related data in which the first compressed data is associated with the ruled line color, and related data in which the second compressed data is associated with the ruled line color, and includes them in the combined compressed data as layer information.

  FIG. 12 is a flowchart showing the flow of the area-specific compression process executed by the MFP according to the second embodiment. In the image processing apparatus according to the second embodiment, processing similar to the region-by-region compression processing shown in FIG. 6 is executed, but the processing from step S23 to step S28 is different, and instead of these steps, steps S41 to S46 are performed. Executed. Referring to FIG. 12, steps S <b> 41 to S <b> 46 are performed for all ruled line regions having the same hue and saturation group among the ruled line areas integrated for each hue and saturation group. In step S41, the image data of the ruled line area is binarized with a predetermined threshold value. Here, since the ruled line area has a resolution of 300 dpi and the image data has a high resolution of 600 dpi, the ruled line area is converted to 600 dpi to specify the image data of the ruled line area. The data obtained by binarizing the image data of the ruled line area is data in which the pixels having lightness lower than a predetermined threshold are set to “1” and the other pixels are set to “0” in the image data of the ruled line area. In the binarized data, the pixel with the pixel value “1” indicates a ruled line in the high-resolution image data. Then, the ruled line color is determined from the pixel values of the RGB image data specified by the binarized data (step S42). The ruled line color is, for example, the average value or median value of the RGB values of the pixels of the RGB image data corresponding to the pixels whose pixel values are “1” in the binarized data.

  Then, ruled line binarized data is generated (step S43). The ruled line binarized data is generated by converting the binarized data to a low resolution. Here, the resolution of the ruled line binarized data is 300 dpi. The generated ruled line binarized data is reversibly compressed (step S44). As a result, first compressed data obtained by reversibly compressing the ruled line binarized data is generated. Next, interpolated binary data is generated (step S45). The interpolated binarized data is a difference between the data binarized in step S41 and the data obtained by restoring the resolution of the ruled line binarized data. In other words, interpolated binarized data is the data binarized in the ruled line binarized data from the data binarized in step S41. The interpolated binarized data remains at the same 600 dpi resolution as the image data, and includes only the ruled line portion that cannot be included in the low-resolution ruled line binarized data among the ruled lines included in the ruled line region. Then, the generated interpolated binary data is reversibly compressed (step S46). As a result, second compressed data obtained by reversibly compressing the interpolated binary data is generated.

  FIG. 13 is still another diagram showing the relationship between ruled line binarized data and interpolated binarized data. 13A shows a ruled line portion of image data, FIG. 13B shows ruled line binarized data, and FIG. 13C shows interpolated binarized data. Referring to FIG. 13A, ruled line binarized data is data obtained by binarizing data obtained by reducing the resolution of image data to 1/2 in the vertical and horizontal directions. It can be seen that by reducing the resolution of the image data, a part of the ruled line of the image data is lost and the shape of the ruled line is deformed. Even if the ruled line binarized data is restored, it does not have the same shape as the ruled line existing in the image data. Referring to FIG. 13C, the interpolated binarized data is set to “1” for pixels lost from the image data by reducing the resolution of the image data to generate ruled line binarized data. Is data in which “0” is set. For this reason, it can be understood that the ruled line shape shown in FIG. 13A is obtained by combining the restored ruled line binarized data and the interpolated binarized data. Therefore, the same result as that obtained by binarizing the ruled line area of the image data can be obtained. The ruled line binarized data has a reduced data amount because the resolution is reduced to 1/2 in the vertical and horizontal directions. For this reason, the data amount of the ruled line binarized data and the interpolated binarized data is smaller than the binarized ruled line area of the image data.

  According to MFP 100 in the second embodiment, ruled line binarized data obtained by binarizing data obtained by reducing the resolution of image data in the ruled line area for each color from the ruled line area in the RGB image data, and the ruled line area The image data is divided into binary binarized data obtained by removing ruled line components included in the ruled line data from the binarized data and compressed. By reducing the resolution of the image data in the ruled line area, the data amount is reduced. On the other hand, the irregularity of the ruled line shape becomes conspicuous, but the irregularity of the ruled line is alleviated by the interpolated binary data. Since the interpolated binarized data is data obtained by removing the ruled line components included in the ruled line data from the binarized image data of the ruled line area, the data amount is reduced. For this reason, the image data of the ruled line region can be restored, and even if the image data is compressed, the rattling of the ruled line can be made inconspicuous.

  In the present embodiment, an MFP is described as an example of an image processing apparatus. However, the present invention is not limited to an MFP, and any apparatus capable of processing image data may be used. For example, a computer may be used. In the present embodiment, the MFP has been described as an example. However, the invention can be captured as an image compression method and an image compression program for causing a computer to execute the compression processing illustrated in FIGS. 6 and 7. Needless to say.

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

<Appendix>
(1) The image processing apparatus according to (1), wherein the first compression unit compresses by a lossless compression method.
(2) In (1), the second compression means compresses by a lossy compression method.
(3) The image processing apparatus according to claim 1, wherein the input unit includes a reading unit that outputs image data obtained by reading a document.

1 is a block diagram showing a hardware configuration of an MFP according to a first embodiment of the present invention. 2 is a functional block diagram illustrating an outline of functions of an MFP according to the first embodiment. FIG. It is a functional block diagram which shows the detailed function of the ruled line binarization part in 1st Embodiment. It is a figure which shows the histogram of the brightness in the ruled line area vicinity of image data. It is a figure which shows the relationship between ruled line binarization data and interpolation binarization data. 6 is a flowchart showing a flow of area-specific compression processing executed in the MFP. It is a flowchart which shows the flow of the ruled line separation process performed by step S13 of FIG. It is a figure which shows the data structure of combined data. It is another figure which shows the relationship between ruled line binarization data and interpolation binarization data. FIG. 10 is a functional block diagram illustrating an outline of functions of an MFP according to a second embodiment. It is a functional block diagram which shows the detailed function of the ruled line binarization part in 2nd Embodiment. 10 is a flowchart showing a flow of compression processing by area executed by the MFP according to the second embodiment. It is another figure which shows the relationship between ruled line binarization data and interpolation binarization data.

Explanation of symbols

  10 Pre-processing unit, 11 Area discriminating unit, 12 Character value conversion unit, 13, 15 Lossless compression unit, 14, 14A Ruled line binarization unit, 16 Resolution reduction unit, 17 Lossless compression unit, 18 Joining unit, 21 1 binarization unit, 22 first ruled line color determination unit, 23 second binarization unit, 24 second ruled line color determination unit, 31 binarization unit, 32 resolution conversion unit, 33 difference extraction unit, 34 ruled line color Determination unit, 102 storage unit, 103 scanner unit, 104 input image processing unit, 105 output image processing unit, 106 image forming unit, 107 network I / F, 108 modem, 109 operation panel, 109A input unit, 109B display unit, 110 Card I / F, 110A IC card, 111 bus, 112 network, 113 telephone line.

Claims (11)

Input means for inputting image data;
Ruled line area extracting means for extracting ruled line areas including ruled lines from the input image data;
First generation means for generating ruled line data including at least a central pixel of the ruled line from the image data of the ruled line region;
Second generation means for generating interpolation data for interpolating the ruled line data from the image data of the ruled line region;
First compression means for compressing the ruled line data and the interpolation data to generate first and second compressed data;
Second compression means for compressing an area excluding the ruled line area from the image data to generate third compressed data;
An image processing apparatus comprising: a combining unit configured to combine the first compressed data, the second compressed data, and the third compressed data to generate combined data. The first generation means includes first extraction means for extracting pixels whose brightness is equal to or less than a first threshold value from pixels constituting a ruled line included in the image data of the ruled line region,
The said 2nd production | generation means contains the 2nd extraction means which extracts the pixel in which the lightness exceeds the said 1st threshold value from the pixels which comprise the ruled line contained in the image data of the said ruled line area | region. The image processing apparatus described.   3. The image processing according to claim 2, wherein the first extraction unit includes a first binarization unit that binarizes image data of the ruled line region with the first threshold value to generate the ruled line data. apparatus. A first ruled line color determining means for determining a first ruled line color of a ruled line included in the ruled line data from the ruled line data and the image data;
The image processing apparatus according to claim 3, wherein the combining unit includes a ruled line color association unit that associates the first compressed data with the first ruled line color.   The second extraction means includes second binarization means for binarizing the image data of the ruled line area inside and outside the first threshold value and the second threshold value to generate the interpolation data. The image processing apparatus according to claim 3, further comprising: First ruled line color determining means for determining a first ruled line color of a ruled line included in the ruled line data from the ruled line data and the image data;
A second ruled line color determining means for determining a second ruled line color included in the interpolation data from the interpolation data and the image data;
The image processing apparatus according to claim 5, wherein the combining unit includes a ruled line color association unit that associates the first compressed data with the first ruled line color and associates the second compressed data with the second ruled line color. . The first generation means includes binarization means for binarizing the image data of the ruled line area;
Resolution conversion means for converting the binarized data into ruled line data having a low resolution,
The second generation means includes binarization means for binarizing the image data of the ruled line area;
The image processing apparatus according to claim 1, further comprising: a difference extracting unit that generates interpolation data by removing ruled line components included in the ruled line data from the binarized data. Further comprising a dividing unit that collects ruled lines whose hue and saturation sets are approximated from the image data of the ruled line areas, and divides the image data of the ruled line areas into sets of hues and saturations to generate a plurality of divided data;
2. The image processing compression apparatus according to claim 1, wherein each of the first and second generation units receives the plurality of divided data and extracts the ruled line data and the interpolation data for each set of hue and saturation. Inputting image data;
Extracting a ruled line area including a ruled line from the input image data;
Generating ruled line data including at least a central pixel of the ruled line from the image data of the ruled line region;
Generating interpolation data for interpolating the ruled line data from the image data of the ruled line region;
Compressing the ruled line data and the interpolation data, respectively, to generate first and second compressed data;
Compressing an area excluding the ruled line area from the image data to generate third compressed data;
An image compression method comprising: combining the first compressed data, the second compressed data, and the third compressed data to generate combined data. Inputting image data;
Extracting a ruled line area including a ruled line from the input image data;
Generating ruled line data including at least a central pixel of the ruled line from the image data of the ruled line region;
Generating interpolation data for interpolating the ruled line data from the image data of the ruled line region;
Compressing the ruled line data and the interpolation data, respectively, to generate first and second compressed data;
Compressing an area excluding the ruled line area from the image data to generate third compressed data;
An image compression program that causes a computer to execute a step of combining the first compressed data, the second compressed data, and the third compressed data to generate combined data. First compressed data obtained by extracting a center pixel of a ruled line from a ruled line region included in image data and compressing ruled line data including at least the center pixel;
Generating interpolated data for interpolating the ruled line data from the ruled line area included in the image data, and second compressed data obtained by compressing the interpolated data;
A compressed image data structure including third compressed data obtained by compressing an area obtained by removing the ruled line area from the image data.

Images