JP4035456B2 - Image compression method and image compression apparatus - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image compression method for compressing a color document image, an image compression apparatus, a program, and a recording medium.
[0002]
[Prior art]
In recent years, with the widespread use of color printers, color scanners, etc., the number of colorized documents has increased, and there has been an increased opportunity to capture and store them as electronic files or send them to third parties via the Internet. Yes. However, if the full color data is used as it is, the load on the storage device and the line is large, so it is necessary to reduce the amount of data handled by a method such as compression processing.
[0003]
Conventionally, as a method for compressing a color image, for example, a method of compressing a binary image having a pseudo gradation by error diffusion or the like, a method of compressing in a JPEG format, and a ZIP compression by converting to an 8-bit palette color And a method of performing LZW compression (see, for example, Patent Document 1).
[0004]
[Patent Document 1]
JP 2002-077631 A
[0005]
[Problems to be solved by the invention]
However, although the size (data amount) is small in the pseudo gradation binary image, the color information is lost. Further, when compression is performed in the JPEG format, a trade-off between the compression size and the character quality occurs due to mosquito noise peculiar to JPEG.
[0006]
Furthermore, the method of converting to palette colors and performing ZIP compression or LZW compression is efficient when the color document image is held in multiple bits because the color distribution of most of the color document image is local rather than discrete. However, as a result of compressing it, the efficiency is naturally inferior.
[0007]
Furthermore, according to the method described in Patent Document 1, high quality can be obtained for a normal character region by combining region determination, binary compression by MMR, lossless compression by ZIP, and lossy compression by JPEG. However, for example, it is difficult to determine the area of a portion of a manuscript that has been corrected by handwriting, and there is a problem that mosquito noise occurs due to JPEG compression.
[0008]
The present invention has been made in view of such circumstances, and provides an image compression method, an image compression apparatus, a program, and a recording medium capable of suitably compressing a color document image with high compression efficiency. Objective. Another object of the present invention is to be able to cope with compression of an image in which character areas and non-character areas are mixed, in addition to the above-described image of only a color document.
[0009]
[Means for Solving the Problems]
In order to solve the above-described problem, the present invention provides an image compression method for compressing a color image, wherein each pixel of the color image is converted into an index assigned corresponding to a color value, and an indexed color is converted. An index conversion step for generating color information including data on the number of pixels and color for each pixel, an index image obtained by converting each pixel into an index, and a color value corresponding to a specific index in the index image as a background of the color image. A background color determination step to be a color, a compression order determination step for determining a compression order when the color image is compressed for each index based on the color information and the background color, and each index from the index image A binary image generating step for generating a binary image and the binary according to the compression order A compression step of compressing an image; and a generation step of generating compressed image data by integrating background data including the size of the color image and the color value of the background color and compressed data of the binary image for each index. It is characterized by having.
[0010]
DETAILED DESCRIPTION OF THE INVENTION
In the following embodiments, a color image compression technique that can be mounted on, for example, a color copying machine will be described.
[0011]
The functions of the color copying machine include, for example, a color copy function, a color print function, and a color scanner function. The compression technology described in this embodiment is applicable to the color copy function and the color scanner function. is there. Specifically, it is a compression technique used when compressing color image data obtained by reading a color original.
[0012]
As the color scanner function, for example, a data transmission function for compressing color image data obtained by reading a color original and transmitting the compressed data to the outside, and the color image data are compressed and stored in a storage unit inside the copier. This is a compression technology that can be installed in both functions.
[0013]
In the present embodiment, a plurality of compression modes are provided as described below. These modes are selected depending on the content of the color image (color document) to be compressed. This selection may be automatic or manual by the user.
[0014]
<< first compression mode >>
First, as an embodiment related to the first compression mode, a compression method described in detail below can be executed. This compression mode is suitable for compressing a color image composed of a small number of colors. Therefore, this compression mode can be automatically selected by determining the color distribution of the color image or determining the character and photographic area and determining the ratio. Alternatively, the user may select manually.
[0015]
Hereinafter, a configuration or method for implementing the first compression mode will be described in detail with reference to the drawings.
[0016]
FIG. 7 is a schematic diagram for explaining data constituting a color document image to be compressed in the present embodiment. As shown in FIG. 7, in the present embodiment, an original color document in which red letters and black letters are formed on a white background and handwritten correction is added with a blue ballpoint pen is used. The color image (that is, the original image) captured by the scanner is divided into white, black, red, and blue partial binary images by performing color reduction processing, same color determination processing, intermediate color reduction processing, and the like.
[0017]
Here, when the number of pixels for each of the divided colors is compared, the white color of the background is usually the largest, so this color is determined as the background color. The white color determined as the background color is not stored as image data, but is stored according to the actual size and color value of the original. The remaining black, red, and blue colors are displayed as shown in FIG. Each is divided into partial (binary) images and compressed and stored. In this way, unlike a natural image such as a color photograph, a color document image has a local color distribution and therefore has high compression efficiency.
[0018]
Here, FIG. 8 is a schematic diagram for explaining the foreground color image, the intermediate color image, and the complementary intermediate color image. That is, since the partial binary image that has not been reduced by the intermediate color reduction processing is as shown in FIG. 8B, the inside is complemented to reduce the edge amount as in the complementary intermediate color image shown in FIG. The image is expanded in the order of the complementary intermediate color image and then the foreground color image. By doing so, the complemented portion is overwritten by the foreground color image that is the upper layer, and the same image as when the complement processing was not performed can be reproduced and the compression size can be reduced.
[0019]
FIG. 9 is a schematic diagram for explaining the principle of the complementary processing in the present embodiment. In FIG. 9, when the portion indicated by the horizontal line 91 of the image pattern “A” is expressed as a layer, it is as shown in FIG. Here, the intermediate color is represented by a pattern with 8 edges, but if it is drawn ahead of the foreground color, even if complement processing is performed as shown on the right, if viewed from above, there is no complement processing. Looks the same. Here, when the number of edges of the intermediate color subjected to the complement processing is examined, it is four as shown in the lower right FIG. 93, which is reduced to half before the complement processing. This increases the redundancy of the partial binary image and improves the compression efficiency.
[0020]
Here, as for the compression order, if a color having a pattern like the intermediate color image shown in FIG. 8B is selected at the top, the compression size becomes small, but actually checking the image pattern increases the processing load. . In general, a foreground color in a document image is a highly visible color such as black on a white background or yellow on a blue background, and an intermediate color in question is, for example, gray between white and black. Thus, the same effect as when the image pattern is examined can be obtained by sorting the colors remaining after the color reduction by values such as luminance and color difference and determining the compression order based on the extracted background color.
[0021]
FIG. 1 is a block diagram for explaining a detailed configuration of a compression apparatus for compressing a color document image in the first compression mode of the present invention. In FIG. 1, reference numeral 101 denotes an original image (color image) obtained by reading a color document with a scanner or the like. Reference numeral 102 denotes a color reduction processing unit, which performs simple color reduction processing on an input original image 101 to a predetermined number of colors and indexes it.
[0022]
Reference numeral 103 denotes color information, which includes data representing the number of pixels, the color centroid, and the color distribution range for each color indexed by the color reduction processing unit 102. Reference numeral 104 denotes an index color image to which an image reduced in color by the color reduction processing unit 102 is output.
[0023]
A color information sorting unit 105 sorts the data of the color information 103 according to the number of pixels. The same color integration unit 106 compares the color information 103 sorted by the color information sorting unit 105, determines the same color, performs integration processing, and updates the color information.
[0024]
Reference numeral 107 denotes an intermediate color reduction unit that performs a process of combining the color information in the vicinity of the color information 103 to reduce the number of colors located in the halftone. After processing here, the value of the top color centroid of the remaining color information becomes the background color data 108. A compression order determination unit 113 sorts the color information reduced by the intermediate color reduction unit 107 by luminance, and determines the order of compression based on the background color information.
[0025]
Reference numeral 109 denotes a binary image creation / compression unit, which creates a binary image for each remaining color information other than the background color and performs compression in accordance with the order determined by the compression order determination unit 113. At this time, the color information image having a high compression order is complemented by the image data having a low order.
[0026]
Reference numeral 110 denotes binary image compression data, which is composed of data groups created by the binary image creation compression unit 109, and color information is added to each. A data integration unit 112 integrates the background color data 108 and the binary image compressed data 110 to create a compressed image 112.
[0027]
That is, in the present embodiment, as will be described in detail below, an image compression method for compressing a color image (particularly a color document image) will be described. The procedure is as follows. First, the color reduction processing unit 102 converts each pixel of the color image into an index assigned corresponding to the color value, and the color information 103 including the number of pixels for each indexed color and each pixel. And an index image (index color image 104) converted into an index. Then, the color value corresponding to the predetermined index in the index image is set as the background color of the color image by the color information sorting unit 105 or the like. In addition, the compression order determination unit 113 determines the compression order when the color image is compressed for each index based on the color information and the background color. Further, the binary image creation compression unit 109 generates a binary image for each index from the index image, and compresses the binary image according to the compression order. Then, the data combining unit 111 integrates the background data 108 including the size of the color image and the color value of the background color and the compressed data of the binary image for each index to generate compressed image data (compressed data 112). It is characterized by.
[0028]
FIG. 2 is a flowchart for explaining the flow of color image compression processing in the first compression mode of the present invention.
[0029]
First, when a color image that is the original image 101 is input, the color reduction processing unit 102 performs color reduction processing of the original image 101 to a predetermined number of colors, and outputs color information 103 and an index color image 104. Here, in the present embodiment, the indexing is characterized by generating an index image (index color image 104) in which one index is associated with a predetermined range of color values and the number of bits of the color image is reduced. To do.
[0030]
Note that the color reduction processing in the present embodiment is to reduce the full-color RGB 24-bit data to R-G-B, such as 2-2-2, 3-2-2, 3-3-3 bits, etc. . The method of selecting the number of bits can be arbitrarily selected depending on the accuracy of the color determination. In the following embodiment, description will be made assuming that 2-2-2 bits are selected. The color information 103 output from the color reduction processing unit 102 includes the number of pixels for each reduced color, the color centroid (average value of colors), and coordinate data representing the color distribution range.
[0031]
That is, in the image compression method according to the present embodiment, in the index conversion process described above, in addition to the number of pixels, the color centroid and color distribution range for each indexed color of the color image are counted as color information. It is characterized by doing.
[0032]
Next, the color information 103 obtained by the color reduction processing unit 102 is input to the color information sorting unit 105, and is sorted by a value that is weighted according to the number of pixels and the index number (step S202).
[0033]
Sorting is basically positioned higher as the number of pixels is higher, but the weighting, that is, the coefficient, is adjusted so that the color close to the primary color is higher when comparing the color information of the same number of pixels. Value. FIG. 3 is a diagram illustrating an example of a sorting weight coefficient in the color information sorting unit 105 in the first compression mode. As shown in FIG. 3, in this embodiment, coefficients of 0.9 to 1.2 are assigned to each of the four gradations of RGB. This value is arbitrary, and may be changed according to the number of color reductions in the color reduction processing unit 102 and the priority of colors.
[0034]
Further, the color information sorted by the color information sorting unit 105 is input to the same color integration unit 106. The same color integration unit 106 compares the values of the color centroids of the color information, and if the distance is short, the color information is integrated as the same color (step S203). This process is a process for returning a color separated into a plurality of colors to one color because any RGB value is close to the threshold value in the first color reduction process, although they are originally the same color. For the color information to be integrated, the sorted upper color information remains, and the data of the number of pixels for each color, the color centroid, and the color distribution range are recalculated.
[0035]
That is, the image compression processing according to the present embodiment is characterized by performing the same color integration processing for integrating color information having close color centroids in color information. Further, when the same color integration process integrates the same color into a sorting higher color, the calculated number of pixels, the color distribution range, and the centroid of the color value are recalculated. Furthermore, the same color integration process performs sorting based on the product of coefficients determined in advance according to the number of pixels for each color in the color image and the position in the color space, and determines the same color based on the sorting order. It is characterized by performing.
[0036]
Further, the intermediate color reduction unit 107 converts the color centroid value of the color information updated by the same color integration unit 106 into a luminance color difference value, compares the values, and performs a process of combining neighboring color information. (Step S204). Thereby, the number of colors located in the halftone is reduced. The purpose of this processing is to remove many gradation components generated from white to black at the boundary between the white of the background and the black of the character part when the original document image is a black and white document, for example. That is. That is, gray near white is white and gray near black is black. The other colors are processed in the same manner. In this way, sorting is performed according to the product of the coefficients corresponding to the number of pixels in the color and the position in the color space, and the color centroid value is not recalculated when performing the same color integration processing and intermediate color reduction processing according to the order. This improves the color reproducibility.
[0037]
In the same color integration unit 106, the color information to be reduced and combined is recalculated and updated similarly to the update of the color information in step S203, but only the value of the color centroid is left as it is. This is because the value of the color centroid is the final display color, and if the intermediate colors are integrated to obtain the color centroid, for example, if it is black, gray will be mixed and become a lighter color, and white will be darker This is to prevent it from becoming. As the representative color of the combined color information, the color centroid of the color information having a high sorting order is adopted. Here, the reason why the weighting coefficient is used in the sorting as well as the number of pixels is to give priority to the primary colors that will be used in the color document original.
[0038]
In other words, the image compression processing according to the present embodiment is characterized in that the color information is subjected to intermediate color reduction processing for combining the color information having similar luminance color differences with each other to reduce the intermediate color. The intermediate color reduction process is characterized in that intermediate colors are reduced based on the sorting order. In addition, when the intermediate color reduction process integrates similar colors into sorting upper colors, the number of counted pixels and the color distribution range are recalculated, and only the color centroid is not recalculated.
[0039]
Then, the value of the color centroid of the most significant color information obtained as a result of processing by the intermediate color reduction unit 107 is output as the background color data 108 (step S205). That is, the image compression processing according to the present embodiment is characterized in that the centroid of the color value of the color positioned at the top of the sorting is extracted as the background color.
[0040]
Further, the color information is sorted, and the compression order is determined based on the background color obtained in step S205 using the result (step S206). For example, if the color information reduced by the intermediate color reduction unit 107 is five colors with identification numbers 1 to 5 and the background color identification number is 4, the result of sorting by luminance is 2, 3, 5, 1, 4 , The compression order is 1, 5, 3, 2 in this order. In the case of a document using many white characters, the order is 4, 1, 5, 3, 2, and the compression order is 1, 5, 3, 2. Furthermore, when it is estimated that white characters and normal characters are mixed, such as 2, 3, 4, 5, 1, the priority of white characters that are considered to be less frequently used in the document is set. In order to make it low, the compression order is set to 5, 1, 3, 2.
[0041]
Then, a binary image for each color is created using the color information remaining in the intermediate color reduction unit 107, the result of the compression order determination unit 113, and the index color image 104, and is compressed by a method such as MMR (step S207). ). Note that the binary image created here has a size corresponding to the color distribution range held in the color information, and if it exists only in a part of the document, only that part is compressed and stored.
[0042]
That is, in the image compression method according to this embodiment, in determining the background color, the color value corresponding to the index of the maximum number of pixels is used as the background color of the color image. Then, determination of the compression order for the index is not performed, and generation and compression of a binary image are not executed.
[0043]
For example, assuming that the index number of the index color image 104 is 0 to 63 and the color information of the indexes 60 and 62 is integrated into the index 63, the binary image is an index image. This is an image obtained by ORing the data of 60 and 62, and is drawn with the value of the color centroid of the color information that the index 63 has. This is the basic part of the binary image for each color, and complement processing is performed according to the order determined in step S206.
[0044]
In this complementing process, the background color identification number is 1, and the order of the remaining colors is 2, 3, and 4. The identification number is a temporary identifier representing a set of integrated index numbers. For example, when the arrangement of 8 pixels is 1, 2, 3, 4, 2, 3, 3, 1, first, the identification number 2 having the highest priority is first viewed. If attention is paid only to the identification number 2, the bit pattern is 01001000, but complement processing is performed to obtain a bit pattern of 01111000. The identification numbers 3 and 4 are 00100110 and 00010000, respectively.
[0045]
The result is binary image compression data 110, which is a data group comprising color values, position and size, and MMR compression data for each color information data. Then, the data combining unit 111 creates the compressed data 112 by combining the background color data 108 and the binary image compressed data 110 and outputs them (step S208). That is, the compression processing according to the present embodiment is characterized in that the binary image is MMR compressed.
[0046]
FIG. 4 is a diagram for explaining the configuration of the compressed data 112 in the first compression mode. First, information such as the size of the input document image (original), the color value of the background color, and the resolution is stored in the header portion. Here, since the color having the largest number of pixels is basically selected as the background color, for example, when the original is printed on color paper such as red, a red value is entered. However, it is generally considered that the background is often white, so the white color of the background color may be determined, and if the background color is determined to be white, the background color value may be omitted. In the white determination, for example, when each value of RGB is equal to or larger than a certain value and the difference between the values is within a certain value, it is regarded as white.
[0047]
Further, as shown in FIG. 4, the header portion is followed by compressed data for each color. The compressed data is composed of the color value, the position coordinates and size where the color exists, and the MMR compressed data body. For example, when the number of colors remaining excluding the background color is N, data having the same structure exists for the number of colors. This storage order is assumed to follow the compression order determination unit 113. Note that if the input image is a single-color original such as a blank sheet, data for this portion is not created.
[0048]
If the input image is a black and white document, the number of color compressed data is 1, so the data amount is almost equivalent to that of a binary image. Here, if the black pixel is only a part of the document, the MMR compressed data is compressed only in that part, so that it becomes smaller than the normal MMR compression.
[0049]
Next, the compressed data 112 is returned to the original image. The entire area of the original is written with the background color stored in the header portion shown in FIG. This is possible by expanding the MMR image and overwriting the stored position with a color value using the image as a mask. Here, the storage order follows the order of the compression order determination unit 113, and the decompression is performed in this order. Therefore, the binary image creation compression unit 109 has the same decompression result regardless of the presence or absence of the complement processing. Thus, the compression efficiency is improved by complementing the partial binary image according to the compression order. In addition, by determining the compression order based on the background color, it is possible to quickly check the order in which the compression efficiency is improved when complement processing is performed.
[0050]
FIG. 5 is a flowchart for explaining in detail the operation procedure of the intermediate color reduction unit 107 shown in step S204 in the compression processing of FIG.
[0051]
First, a list is created in the sorting order of color information (step S501). Then, intermediate color reduction processing is performed along this list. Next, the color centroid value of the color information is converted from RGB to YCrCb luminance color difference signals and data is added (step S502). The reason for converting to a color difference is that it is suitable for reducing intermediate colors by integrating similar colors with close luminance differences.
[0052]
In steps S503, S504, and S505, a comparison is made as to whether or not the difference between each component of Y, Cr, and Cb added to each color information is within a predetermined threshold value. Reduce the number of colors by integrating. FIG. 6 is a flowchart for explaining details of the processing in steps S503 to S505 in FIG. Note that the processing of the flowchart shown in FIG. 6 is different from the processing in steps S503 to S505 only in the threshold for comparing YCrCb.
[0053]
First, color information I is selected from the color information list given from the processing of step S501 in FIG. 5 (step S601). Next, the YCrCb value of the color information I is stored in Y′Cr′Cb ′ (step S602). Then, the lowest color information in the list is selected as the color information J (step S603).
[0054]
Here, the respective values of Y, Cr, and Cb of the color information I and J are compared, and it is checked whether or not the difference is within the threshold value given in each of steps S503 to S505 in FIG. 5 (step S604).
[0055]
As a result, if it is determined that the value is within the threshold (YES), the process proceeds to step S606, and if not (NO), the process proceeds to step S608. However, if the signs of the Cr or Cb values of the color information I and J are different, the absolute value of the Cr or Cb value of the color information J is checked, and this condition is not satisfied if it is greater than a predetermined value. Like that. For example, even if the difference between the Cr values of the color information I and J is within the threshold value given in the flowchart shown in FIG. 5, if the signs are different, the absolute value of the color information J is compared with a predetermined value. If it is smaller, the process proceeds to step S605. If it is larger, the process proceeds to step S608.
[0056]
The same process is performed for the value of Cb. This is because, depending on the size of the threshold set in the flow of FIG. 5, when the differences are simply compared and integrated, the color information of different colors such as light blue and light red is integrated. Is to prevent. Conversely, if the threshold value is set so that different colors such as light blue and light red are not integrated, it is impossible to integrate similar colors such as blue and light blue.
[0057]
In step S606, the color information J is integrated with the color information I. As a result, the data such as the number of colors and the distribution range of the color information I are updated, but the value of the color centroid is not updated. Instead, the color centroid when integrated is calculated and reflected in Y′Cr′Cb ′ stored in step S602.
[0058]
Next, the color information J data is removed from the list (step S607). In step S608, it is checked whether or not the color information located at the top of the color information J list is I. As a result, if it is not I (NO), the process proceeds to step S609, and if it is I (YES), the process proceeds to step S610.
[0059]
In step S609, the color information located one level above color information J is set as color information J, and the process returns to step S604. On the other hand, in step S610, the stored Y′Cr′Cb ′ value is returned to YCrCb of color information I, and the reference values for threshold comparison in the flowcharts shown in FIGS. 5 and 6 are integrated in steps S604 to S609. The updated color information is updated to the original color centroid value.
[0060]
In step S611, it is checked whether the color information immediately below color information I is at the lowest level, or is deleted from the list in step S607 and does not exist. The process proceeds to S612. In step S612, the color information immediately below the color information I is set to I, and the process returns to step S602. On the other hand, if not (YES) in step S611, the process is terminated, and the updated color information list is returned to the flowchart of FIG.
[0061]
In the present embodiment, the threshold value set in each of steps S503 to S505 described with reference to FIG. 6 is set slightly larger than the color difference CrCb for the purpose of integrating similar colors. In step S503, the color information of similar colors with a small threshold and a relatively small difference in luminance color difference is integrated, and the threshold is increased in steps S504 and S505, and the similar color having a relatively large difference in luminance color difference. Integrate color information. The value of YCrCb, which is the temporary color centroid added in step S502, is recalculated in the flowchart of FIG. 6 so that the colors can be integrated without setting the threshold value after step S504 to a very large value. It is to do.
[0062]
For example, it is assumed that bright black is integrated with black in the process of the first step S503. In this case, the value of YCrCb, which is the temporary color centroid, moves in the bright black direction. In such a case, since the distance from the dark gray is closer, the dark gray can be integrated into the black simply by increasing the threshold value after the next step S504 to some extent.
[0063]
In addition, as shown in steps S503 to S505, multi-stage processing is performed because it is possible to integrate from the closest color, and the size of the compressed data 112 can be controlled by changing the number of stages. is there.
[0064]
As can be seen from FIG. 4, one of the factors that determine the size of the compressed data 112 is the number of color compressed data. That is, if the number of colors finally remaining is controlled, the compression size can be controlled. For example, if the number of colors is increased, the quality is close to that of the original color image, and if the number of colors is decreased, the image is closer to a simple binary image. Therefore, the number of stages may be determined according to the image quality to be obtained and the compression size.
[0065]
Thereafter, in step S506, the updated color information is returned to the processing shown in the flowchart of FIG.
[0066]
In other words, in the image compression method according to the present embodiment, the intermediate color reduction process has a plurality of predetermined thresholds, converts the centroid of the color value into a luminance color difference signal, and performs integrated reduction of the colors in the threshold range according to the sorting order. It is characterized by performing. Further, the centroid of the color value is recalculated every time integration reduction is performed using the value converted into the luminance color difference signal as the temporary color centroid. Further, the plurality of threshold values are changed depending on whether the compression rate priority or the image quality priority is given.
[0067]
As described above, in view of the fact that many color document images have one or more characters or the like formed on a specific background color, there is a partial binary image for each color when this is compressed. Therefore, it can compress efficiently. Also, the image quality and size can be controlled by using a plurality of threshold values in the intermediate color reduction processing and controlling the number and values of the threshold values.
[0068]
In this compression mode, each pixel constituting the color document image is classified into one of a predetermined number of limited colors, a color corresponding to the background is specified from the limited colors, and the luminance value of the background color is determined. In accordance with the brightness value of each other limited color, the overwriting relationship of each limited color is ranked, and for each limited color, a connected pixel composed of a pixel corresponding to the limited color and its adjacent pixels A group is generated, and for a connected pixel group corresponding to a certain limited color A (for example, equivalent to an intermediate color in FIG. 9), an adjacent pixel included in the connected pixel group is higher than the limited color A (overwritten later) as an overwrite relationship. For example, if there is a pixel at a position where a limited color corresponding to the foreground color in FIG. 9 is present, the corresponding adjacent pixel is replaced with the limited color A and obtained here. The procedure to compress the connected pixel group In the background (e.g., corresponding to the background 9 color) in the case of processing considered to be the lowest priority override relationship, many color document image is compressed efficiently.
[0069]
(Modification)
In the description of the first compression mode described above, the interpolation processing is performed only in the primary direction (raster direction). However, since the binary image compression method employed also performs the interpolation in the secondary direction, If an effect can be expected, it may be performed.
[0070]
In the above description, the sorting is performed based on the luminance. However, the sorting may be performed based on the distance in the color space between the background color and the remaining colors. That is, in this case, the compression order is determined based on the distance in the color space between the determined background color and the remaining colors.
[0071]
Further, in the above description, the complementing method performs complementation based on whether or not the blank portion of the identification number of the color of interest is filled with the lower identification number, but this is complemented depending on whether the previous pixel is 1 or not. You can go.
[0072]
Furthermore, as in the above description, the identification number is 1, the order of the remaining colors is 2, 3, 4, and the arrangement of 8 pixels is 1, 2, 3, 4, 2, 3, 3, 1. In this case, the identification number 2 is found at the second pixel, 01, the next is identification number 3, the previous pixel is 0, the next is 011, the next is 4, and so on. 01111110. The identification numbers 3 and 4 are 00110110 and 1000010000, respectively. Therefore, although the bit pattern obtained is different from that of the first compression mode described above, since the edge amount is the same, roughly the same compression size can be obtained and only the pixels are sequentially viewed, so the processing is light. High-speed processing is possible.
[0073]
<< second compression mode >>
Next, as the second compression mode, a compression method described in detail below can be executed. This compression mode is suitable for compression of color images in which characters and non-characters (for example, photographic images) are mixed. Therefore, this compression mode can be automatically selected by judging the color distribution of the color image. Alternatively, the user may select manually.
[0074]
Hereinafter, a configuration or a method for implementing the second compression mode will be described.
[0075]
First, an outline of the steps performed by the mode will be described. First, a luminance histogram is generated for the entire image to be encoded, binarized, and several character areas are extracted. Next, character cut processing is performed for each character area, and it is determined from the result whether the area should be treated as a character area again. As a result, if it is not an area to be handled as a character area, it is determined whether or not the object in the area is a single color. If it is a single color, it is a target for MMR compression, and if it is not a single color, it is a target for JPEG compression.
[0076]
On the other hand, if it is determined that the image is to be handled as a character area, the colors constituting the area are reduced by a predetermined color reduction process. As a result of this color reduction processing, when one color is obtained, a palette indicating that color (for example, (R, G, B) = (20, 30, 40)) and a binary image are associated with each other to perform MMR compression. set to target. As a result of the color reduction processing, when it can be expressed by a predetermined number of colors (for example, 4 colors) or less, for each character cut, a pallet indicating each color and a multi-value image indicating the pixel position of each color are associated with the ZIP. Target of compression. If it cannot be represented by a predetermined number of colors, the original image before the color reduction process is set as a target for JPEG compression processing.
[0077]
FIG. 10 is a block diagram for explaining the processing for carrying out the image compression method in the second compression mode and the intermediate image. In FIG. 10, reference numeral 1101 denotes an original image. Reference numeral 1102 denotes an image binarization unit that inputs an original image and performs optimum binarization of the image. Reference numeral 1103 denotes an entire binarized image binarized by the image binarizing unit. Reference numeral 1104 denotes a character area detection unit that inputs the entire binarized image 1103 and detects a character area to create character area coordinates 1112.
[0078]
1108 inputs character area coordinates 1112, calculates the original image color of the black portion of the binary image while referring to the original image and the binary image within the coordinates, creates a plurality of palettes 1114, and accordingly, It is a character color extraction unit that performs color reduction processing.
[0079]
Reference numeral 1105 denotes a region detected as a character by the character region detection unit 1104 and a black region of the binary image 1103 of the region where the character color is less than M by the character color extraction unit 1108 is extracted from the original image. This is a character portion fill portion that fills the surrounding color and creates an image A.
[0080]
Reference numeral 1106 denotes a reduction unit that inputs an image A and reduces it to create an image B. Reference numeral 1107 denotes a JPEG compression unit that inputs an image B and compresses it to create a compression code X (113).
[0081]
Reference numeral 1109 denotes a color-reduced image of a plurality of character areas reduced by the character color extraction unit 1108. Reference numeral 1110 denotes an MMR compression unit that inputs a reduced color image and performs MMR compression to generate a plurality of compressed codes Y (1115) when the reduced color image 1109 is 1 bit. Reference numeral 1111 denotes a ZIP compression unit that inputs a reduced color image and performs ZIP compression to create a plurality of compressed codes Z (1116) when the reduced color image 1109 is 2 bits or more. Finally, the data from 1112 to 1116 summarized in 1A are combined to become compressed data.
[0082]
<Character area detection processing>
FIG. 11 is a flowchart for explaining the character area detecting process in the character area detecting unit 1104. First, a color image is input, and luminance conversion is performed while reducing the resolution by thinning out, thereby creating a luminance image J (step S1301). For example, if the original image is RGB 24-bit 300 dpi, the vertical direction and the horizontal direction are every 4 pixels,
Y = 0.299R + 0.587G + 0.114B
When a new image J is created by performing the above calculation, the image J becomes a Y8-bit 75 dpi image. Next, a histogram of luminance data is taken and a binarization threshold T is calculated (step S1302).
[0083]
Next, the luminance image J is binarized by the threshold value T to create a binary image K (step S1303). Further, the black pixel outline is traced, and all black areas are labeled (step S1304). Next, a character-like area in the black area is determined (step S1305). Then, the objects to be combined from the shape and position are combined (step S1306).
[0084]
<Character color extraction processing for character area>
FIG. 12 is a flowchart for explaining the character color extraction processing in the character color extraction unit 1108. Here, the entire binarized image 1103 is used as the binary image. However, the present invention is not limited to this. For example, only the coordinates of the character area and the color image are input, and the result of binarizing the color image again is obtained. The representative color calculation process may be performed by using it.
[0085]
The process shown in the flowchart of FIG. 12 is performed as described below for all areas determined to be characters by the character area detection unit 1104.
[0086]
(Step S3001: Re-binarization process)
First, re-binarization is determined in step S3001. The entire binarized image 1103 is not necessarily a binarized image of all character areas. Even if the binary image is too dark or too thin, the resulting image quality is adversely affected. Therefore, ideally, it is desirable to perform optimum binarization for each character area.
[0087]
Specifically, the character area detection unit 1104 scans a binary image of an area determined as a character, and performs pattern matching with an isolated point filter. Then, it is determined whether or not an isolated point is present in the area over the threshold. If the isolated point is over the threshold, a luminance histogram of the area is taken, an optimum threshold is calculated, and re-binarization is performed. In the case of a normal character region, a better binary image can be obtained by partially passing through a luminance histogram, but in rare cases, the result is worse than before (ie, binary values that are darker than before after re-binarization). There will be cases).
[0088]
Therefore, in order to prevent such a phenomenon, in the re-binarization, the binarization threshold value used when obtaining the entire binarized image is input, and compared with the threshold value for the re-binarization, compared with the previous one. If a dark result can be obtained, exception processing such as re-binarization is not provided.
[0089]
(Step S3002: Character cutting process)
In step S3002, character cut information is created. In the character-only processing, the processing contents vary depending on whether the character area is horizontal writing or vertical writing. This horizontal writing or vertical writing information is determined by the character area detection unit 1104 based on the arrangement of black clusters. As a result, in the case of horizontal writing, projection of black pixels of the binary image is first taken in the main scanning direction. Then, after detecting line breaks, projection of black pixels is taken in the sub-scanning direction for each line to obtain information for each character. On the other hand, in the case of vertical writing, line segmentation is performed in the sub-scanning direction and character segmentation is performed in the main scanning direction. At this time, in order to withstand a slight inclination of the image, it is preferable to divide into three in the row direction in order to take a projection of the row cutout. By this processing, coordinate information of each line and character coordinate information existing in each line can be obtained.
[0090]
On the other hand, in the character determination process (step S3003 described later), the character limit information is used to determine whether each of the black objects in the area determined as a character by the character area detection unit 1104 is further a character. . Specifically, it is determined whether the character is a character based on the size and shape of the character. For example, in view of image quality and compression, there is no need to stick to “character” in order to convert to a single color or multiple colors. As an example, a mark or the like expressed in a single color is better in image quality compression rate if expressed in a single color MMR than expressed in JPEG. However, as a problem of probability, since regions other than characters are often expressed in gradation, it is important to determine whether or not they are characters.
[0091]
(Step S3003: Character determination process)
Next, character determination is performed in step S3003. Here, information on character cutting (step S3002) is input, and the average character size of the line is calculated for each line. At this time, if the information of extremely small characters is ignored, a better result is obtained. A character rectangle that is extremely larger than the average size is determined not to be a character. Furthermore, regardless of the average, it is determined that the character is not a character when the shape is clearly not a character from the aspect ratio information.
[0092]
As a result of the determination, if m characters exist in the area, and if it is determined that all m characters are not characters, the character area detection unit 1104 outputs a result that this area is an image.
[0093]
However, when n characters (m> n, n> = 0) of m characters are not characters, that is, when a rectangle that is a character remains, the black object determined not to be a character is deleted from the binary image. The result that this area is a character is output.
[0094]
Further, in consideration of the final image quality, the present embodiment adds the following exception processing. For example, when there are 10 characters in the area, 5 characters that are skipped are monochromatic as characters, and JPEG compression is performed assuming that the remaining characters are not characters. Become. Therefore, in this character determination process, in the case where the determination of characters and images is frequently switched, it is arranged so that all rectangle characters are aligned or all rectangle images are aligned based on the arrangement and frequency of rectangles determined to be characters. To do.
[0095]
In the above-described character determination process, the process proceeds to step S3004 for an area determined as a character, and the process proceeds to step S3005 for an area determined not to be a character.
[0096]
(Step S3005: monochrome determination processing)
In step S3005, monochrome determination is performed. An area that passes through this process is an area that is determined not to be a character by character determination although it is determined to be a character by the character area detection unit 1104. As described above, regardless of whether it is a character or not, if the region is expressed in a single color, it is better to use a single color and perform MMR in terms of image quality compression rate. Therefore, it is determined whether or not this area is a single color.
[0097]
As a specific example, a histogram of each RGB level of a color image pixel corresponding to a black portion of a binary image is taken, and if the variance values of all the histograms are less than or equal to a threshold value, this region is determined to be a single color. As a result, if it is a single color, the process proceeds to one color extraction processing in step S3006, and if it is a plurality of colors, DOJPEG is returned.
[0098]
(Step S3006: one color extraction process)
One color extraction processing in step S3006 will be described with reference to the flowchart of FIG. That is, FIG. 13 is a flowchart for explaining the one-color extraction process in the second compression mode. First, thinning processing of a binary image referred to by the character coordinates is performed, black pixels corresponding to a change portion from the background to the character portion at the time of reading the scanner are reduced, and a new binary image newbi is created (step S1201).
[0099]
Next, a histogram of the RGB values of the original image corresponding to newbi black pixels is taken (step S1202). Of course, this processing may be performed in another color space such as YUV. Then, representative values for each of RGB are calculated (step S1203). This representative value is, for example, the largest value. Alternatively, the representative value may be obtained by reducing the number of steps in the histogram and obtaining the largest value in the rough histogram and then obtaining the largest value in the fine histogram existing in the histogram.
[0100]
By taking the latter method, a true representative value 1401 can be obtained from the histogram as shown in FIG. 14 without being confused by the noise 1402. FIG. 14 is a diagram for explaining a representative value calculation method in the one-color extraction process. As a fine histogram, for example, a 256-stage histogram as shown in FIG. 14 is obtained from 8-bit R data. In this case, the maximum value is 1402, which is not a true representative value. Therefore, the histogram is divided into 64 overlapping widths, and the 8-level histogram is recalculated from the 256-level histogram. This is shown from A to H. A and H are only 32 widths. This recalculation indicates that the representative value exists in G, and the maximum value in G can be searched to obtain 1401. By repeating the above processing for all character coordinates, one representative color is calculated for each character coordinate.
[0101]
(Step S3004: Color reduction processing)
In step S3004, a color reduction process is performed on the character. In the color reduction processing unit 1082, even when the original document is expressed in a single color with respect to the color of the character portion, there is a color transition portion from the background to the character portion at the time of scanning with the scanner.
[0102]
FIG. 15 is a diagram for explaining a transition part (gradation) of a character part generated by a scanner. In FIG. 15, for simplicity, only R of RGB is described, and GB is omitted. The letter A was originally composed of R = 32 level monochrome, but when read by the scanner, the data varies as shown by the enlarged pixels. In FIG. 15, there are only three pixels 3201, 3202, and 3203 that have reached black near the original level R = 32, and the other pixels are between the background color (for example, white) and R = 32 As a result, the character is expressed by gradation with a transition portion.
[0103]
FIG. 16 is a diagram showing a transition portion (gradation) of the character portion generated by the scanner as a three-dimensional histogram. In FIG. 16, it is assumed that the background color is white indicated by 3501 and the character color is black indicated by 3502. A transition portion at this time is indicated by 3503. Here, there is no need to strictly represent a transition portion that is a variation due to reading of a character portion that was originally expressed in a single color. In other words, if only the representative color can be expressed, the image quality is good and the data amount is small. However, even if the binary image is thinned, it is difficult to completely remove the color of the transition portion of the character portion from this background.
[0104]
Therefore, by utilizing the fact that one character is often expressed in a single color, the character cut information is used to limit the image to one color and aim to improve the image quality and compression rate. However, when it is desired to compress characters and the like expressed in gradation from the beginning with higher image quality, exception processing such as determination of whether or not the characters are expressed in multiple colors may be added. That is, if the character cut information is used to make each character one color, it is possible to remove a transition portion that occurs as a variation due to reading of a character image originally expressed in a single color.
[0105]
<Character part filling part 1105>
An example of processing performed by the character portion filling unit 1105 will be described with reference to FIGS. 17 and 18. FIG. 17 is a diagram for explaining the outline of the character filling process. FIG. 18 is a flowchart for explaining the flow of the character portion filling process.
[0106]
As shown in FIG. 17A, a gradation image is used as a background, and as an example, an image in which a blue character ABC is drawn near the center is used as an original image. Assume that a binary image of one character area as shown in FIG. 17B is obtained from this original image. As the character portion filling process, first, in step S11010, the entire image is divided into, for example, 32 × 32 regions (hereinafter, parts), and the process is performed for each part. FIG. 17C shows a state where the parts are divided. This figure shows a state of being divided into 4 × 5 parts for simple explanation. The numbers on the upper left of each area in FIG. The number of area divisions is not limited to this, and may be other division numbers.
[0107]
In step S11020, it is determined whether there is an unprocessed part. For an unprocessed part, the process proceeds to step S11030, and it is determined whether a character fill target area exists in the part. It should be noted that even if the area is determined to be a character area by the character area detection unit 1104, the area for which DOJPEG is returned by the character color extraction unit 1108 is not a target area to be painted.
[0108]
As shown in FIG. 17C, the parts 00 to 04, 10, 14, 20, 24, and 30 to 35 are determined to have no character filling target area in step S 11030, and the next part is not processed. Proceed to For a part (for example, part 11) in which a character filling target area exists, the process proceeds to step S11040, the corresponding binary image is referred to, and the RGB value (or YUV, etc.) of the color image corresponding to the white portion of the binary image is referred to. Average value ave_color may be calculated. In step S11050, the corresponding binary image is referred to, and the density data of the pixel corresponding to the black pixel is set as ave_color. The above processing is repeated for the parts where the character filling target area exists (here, parts 12, 13, 21, 22, 23). In this way, the average value of surrounding pixels can be filled in the portion where the character exists.
[0109]
The filled image obtained in this way is reduced in the reduction unit 1106. In this embodiment, simple thinning processing is performed as an example. Note that the order of the reduction process and the character part filling process may be reversed. In that case, it is necessary to pay attention to the displacement between the binary image and the color image.
[0110]
Furthermore, if necessary, a format in which five character region coordinates 1112, a palette 1114, a compression code X (1113), a compression code Y (1115), and a compression code Z (1116) are collected is created.
[0111]
As an example of a format that combines the five, Adobe (registered trademark) PDF or the like can be considered. Adobe's PDF is a format that can be displayed by Adobe Reader (registered trademark), which is distributed free of charge by Adobe, and since there is no application that created the document, troubles such as the inability to open the file on the receiver side are avoided. be able to. Other formats include XML. XML is a description language for exchanging and distributing documents and data via a network.
[0112]
(Modification)
In the second compression mode, the binary image is created with the entire single threshold value. However, the present invention is not limited to this. For example, an optimal threshold value is calculated for each character area detected by the character area detection 104 to 2 A value image may be created. In that case, the re-binarization determination in step S3001 in the flowchart of FIG. 12 is not necessary.
[0113]
Further, the same binary image is used in the character portion filling unit 1105 and the character color extracting unit 1108, but the present invention is not limited to this, and the optimum binarization unit is selected as the character portion filling unit 1105, the character color extracting unit, respectively. It may be owned inside 1108.
[0114]
<< Other Embodiments >>
Note that the present invention can be applied to a system composed of a plurality of devices (for example, a host computer, an interface device, a reader, a printer, etc.), but a device (for example, a copier, a facsimile machine, etc.) composed of a single device. You may apply to.
[0115]
Also, an object of the present invention is to supply a recording medium (or storage medium) in which a program code of software that realizes the functions of the above-described embodiments is recorded to a system or apparatus, and a computer (or CPU or CPU) of the system or apparatus. Needless to say, this can also be achieved when the MPU) reads and executes the program code stored in the recording medium. In this case, the program code itself read from the recording medium realizes the functions of the above-described embodiment, and the recording medium on which the program code is recorded constitutes the present invention. Further, by executing the program code read by the computer, not only the functions of the above-described embodiments are realized, but also an operating system (OS) running on the computer based on the instruction of the program code. It goes without saying that a case where the function of the above-described embodiment is realized by performing part or all of the actual processing and the processing is included.
[0116]
Further, after the program code read from the recording medium is written into a memory provided in a function expansion card inserted into the computer or a function expansion unit connected to the computer, the function expansion is performed based on the instruction of the program code. It goes without saying that the CPU or the like provided in the card or the function expansion unit performs part or all of the actual processing, and the functions of the above-described embodiments are realized by the processing.
[0117]
When the present invention is applied to the recording medium, program code corresponding to the flowchart described above is stored in the recording medium.
[0118]
【The invention's effect】
As described above, according to the present invention, a color document image can be suitably compressed with high compression efficiency. Further, the present invention is not limited to the above-described image of only a color document, and can support compression of an image in which character areas and non-character areas are mixed.
[Brief description of the drawings]
FIG. 1 is a block diagram for explaining a detailed configuration of a compression apparatus for compressing a color document image in a first compression mode of the present invention.
FIG. 2 is a flowchart for explaining the flow of color image compression processing in the first compression mode of the present invention;
FIG. 3 is a diagram illustrating an example of a sorting weighting factor in the color information sorting unit 105 in the first compression mode.
FIG. 4 is a diagram for explaining a configuration of compressed data 112 in a first compression mode.
5 is a flowchart for explaining in detail an operation procedure of an intermediate color reduction unit 107 shown in step S204 in the compression processing of FIG. 2;
6 is a flowchart for explaining details of processing in steps S503 to S505 in FIG. 5;
FIG. 7 is a schematic diagram for explaining data constituting a color document image to be compressed in the present embodiment.
FIG. 8 is a schematic diagram for explaining a foreground color image, an intermediate color image, and a complementary intermediate color image.
FIG. 9 is a schematic diagram for explaining the principle of complement processing in the same embodiment;
FIG. 10 is a block diagram for explaining processing and an intermediate image for carrying out an image compression method in a second compression mode.
FIG. 11 is a flowchart for explaining character region detection processing in a character region detection unit 1104;
FIG. 12 is a flowchart for explaining a character color extraction process in a character color extraction unit 1108;
FIG. 13 is a flowchart for explaining one-color extraction processing in a second compression mode.
FIG. 14 is a diagram for explaining a representative value calculation method in one-color extraction processing;
FIG. 15 is a diagram for explaining a transition part (gradation) of a character part generated by a scanner.
FIG. 16 is a diagram showing a transition portion (gradation) of a character portion generated by a scanner as a three-dimensional histogram.
FIG. 17 is a diagram for explaining an outline of character filling processing;
FIG. 18 is a flowchart for explaining a character filling process;
[Explanation of symbols]
101 Original image
102 Color reduction processing unit
103 Color information
104 Index color information
105 Color information sorting section
106 Same color integration unit
107 Intermediate color reduction unit
108 Background color data
109 Binary image creation compression unit
110 Binary compressed image data
111 Data combiner
112 Compressed data
113 Compression order determination unit

Claims (18)

An image compression method for compressing a color image,
An index image in which each pixel of the color image is converted into an index assigned corresponding to a color value, and includes color information including data regarding the number of pixels for each indexed color and the color, and each pixel is converted into an index. An index conversion process for generating
A background color determination step in which a color value corresponding to a specific index in the index image is used as a background color of the color image;
A compression order determination step for determining a compression order when compressing the color image for each index based on the color information and the background color;
A binary image generation step of generating a binary image for each index from the index image;
A compression step of compressing the binary image according to the compression order;
And a generation step of generating compressed image data by integrating background data including the size of the color image and the color value of the background color and compressed data of the binary image for each index. Compression method. In the background color determination step, a color value corresponding to the index of the maximum number of pixels is set as a background color of the color image,
The image compression method according to claim 1, wherein the compression order determination step, the binary image generation step, and the compression step do not execute determination of a compression order for the index and generation and compression of a binary image.   3. The image compression method according to claim 1, wherein the index conversion step generates an index image in which one index corresponds to a color value in a specific range and the number of bits of the color image is reduced. .   The index conversion step further includes, as the color information, counting a color centroid and a color distribution range for each indexed color of the color image. The image compression method as described. 5. The image compression method according to claim 4 , further comprising the same color integration step of integrating color information having the color centroids close to each other in the color information.   6. The image compression method according to claim 1, further comprising an intermediate color reduction step of combining the color information having luminance color differences close to each other to reduce intermediate colors in the color information.   In the same color integration step, sorting is performed using a value obtained by multiplying a predetermined coefficient according to the number of pixels for each color in the color image and the position in the color space, and the same color determination is performed based on the sorting order 6. The image compression method according to claim 5, wherein:   8. The image compression according to claim 5, wherein the same color integration step recalculates the counted number of pixels, the color distribution range, and the centroid of the color value when integrating the same color into a sorting higher color. Method.   7. The image compression method according to claim 6, wherein in the intermediate color reduction step, intermediate colors are reduced based on a sorting order.   10. The method according to claim 6, wherein the intermediate color reduction step recalculates the counted number of pixels and the color distribution range when integrating similar colors into sorting upper colors, and does not recalculate only the color centroid. Image compression method.   11. The image compression method according to claim 1, wherein the background color determination step extracts a centroid of a color value of a color positioned at the top of sorting as the background color. 7. The image according to claim 6, wherein the intermediate color reduction step has a plurality of predetermined threshold values, converts the centroid of the color value into a luminance color difference signal, and performs integrated reduction of the colors in the threshold range according to the sorting order. Compression method.   13. The image compression method according to claim 12, wherein the centroid of the color value is recalculated every time integration reduction is performed using a value converted into a luminance color difference signal as a temporary color centroid. The plurality of threshold values, the image compression method of claim 12 Symbol mounting and changing the value depending compressibility priority or image quality priority.   The image compression method according to any one of claims 1 to 14, wherein in the compression step, the binary image is subjected to MMR compression.   16. The image compression method according to claim 1, wherein the determination of the compression order is based on a distance in a color space from the color extracted in the background color extraction step. The image compression method enables execution of image compression in a plurality of image compression modes;
In the first compression mode, for the color image to be encoded, the index conversion step, the background color determination step, the compression order determination step, the binary image generation step, the compression step, Performing the generating step,
In the second compression mode, for a color image to be encoded,
An area identifying step for identifying a character area and a non-character area of the color image;
A second compression step of compressing the image of the character region with a binary image algorithm;
The image compression method according to claim 1, further comprising: performing a third compression step of compressing the image of the non-character region using a multi-value image algorithm. An image compression apparatus for compressing a color image,
An index image in which each pixel of the color image is converted into an index assigned corresponding to a color value, and includes color information including data regarding the number of pixels for each indexed color and the color, and each pixel is converted into an index. Index conversion means for generating
A background color determining means that uses a color value corresponding to a specific index in the index image as a background color of the color image;
Compression order determining means for determining a compression order when compressing the color image for each index based on the color information and the background color;
Binary image generation means for generating a binary image for each index from the index image;
Compression means for compressing the binary image according to the compression order;
Image data comprising: generating means for generating compressed image data by integrating background data including a size of the color image and a color value of the background color and compressed data of the binary image for each index Compression device.

Images