JP5983210B2 - Image compression processing method and apparatus, and image forming apparatus - Google Patents

Description

  The present invention relates to an image compression processing method and apparatus for reducing the number of colors by compressing an input image area, and an image forming apparatus including such an image compression processing apparatus.

  Conventionally, multi-function image forming apparatuses called multifunction peripherals or MFP (Multi Function Peripherals) have been used. In an image forming apparatus, an image (image data) is usually input via a scanner or a network. The input image is often subjected to image compression processing for processing inside the apparatus or for storage.

  Conventionally, in image compression, image data (hereinafter, simply referred to as “image”) is divided into a plurality of matrix areas, and pixels in the areas are represented by representative colors (approximate colors). The number of colors in the area is reduced and the amount of data is reduced.

  For example, a method has been proposed in which a small box is placed in the color space, the color of the pixel of the image is put in each box, and the central point of N boxes is finally used as the representative color value ( Patent Document 1).

  Further, a method has been proposed in which the number of gradations is reduced and the number of colors is reduced for each small region obtained by dividing an input image based on image data included therein (Patent Document 2).

Japanese Patent Laid-Open No. 10-40364 Japanese Patent Laid-Open No. 2002-300412

  However, the conventional image compression processing method described above has the following drawbacks. That is, in the method of Patent Document 1, since the central point of each box is used as a representative color, the dynamic range of the color gradation of the original image is reduced by this process. In particular, the influence of the reduction of the dynamic range is large in a halftone original having many colors.

  For example, in the subsequent processing, when different processing between the processing for further reducing the dynamic range and the processing for maintaining the dynamic range is performed in two adjacent regions, the original processing is performed in the region where the dynamic range is further reduced. There is a problem in that there is almost no gradation change, and a gap is generated between the two areas, which is recognized by the user as a texture.

  In the method of Patent Document 2, since the number of gradations is changed according to the characteristics of the input image, the compression rate varies depending on the image, and variable length compression is performed. Therefore, there is a problem that it is difficult to implement this compression processing method in hardware.

  The present invention has been made in view of the above-described problems, and an object of the present invention is to maintain the dynamic range of the color gradation of pixels included in an input image area as much as possible without reducing the number of colors. And

  According to the method of an embodiment of the present invention, the dynamic range for the gradation of each color component in the color space is obtained for the pixels included in the region of the input image, and the dynamic range of each color component is set to L, M, Alternatively, the rectangular parallelepiped corresponding to the dynamic range of each color component is divided into small rectangular parallelepiped blocks by dividing each into N (L, M, and N are integers of 1 or more), and pixels included in the respective blocks A pre-processing for reducing the number of colors in the image area by replacing the gray level of each color component with the gray level of the nearest vertex of the block, and the distance in the color space for the pre-processed pixel This processing for reducing the number of colors in the image area is performed by grouping together those that are close to each other and determining a representative color of each set.

  According to the apparatus of one embodiment of the present invention, DR calculation means for obtaining a dynamic range for the gradation of each color component in the color space for the pixels included in the region of the input image, and the dynamic range of each color component, A block dividing means for dividing a rectangular parallelepiped corresponding to the dynamic range of each color component into small rectangular parallelepiped blocks by dividing each of them into L, M, or N (L, M, N are integers of 1 or more); Pixel replacement processing means for performing processing for reducing the number of colors in the area of the image by replacing the gradation of each color component of the pixels included in each block with the gradation of the nearest vertex of the block; and the pixel replacement For pixels that have been processed by the processing means, those that are close to each other in the color space are grouped together to represent each set. Having a representative color determining means for reducing the number of colors in the region of the image by determining the.

  In addition, a rectangular parallelepiped, a block, a vertex, division | segmentation, etc. are virtual. Therefore, dividing a rectangular parallelepiped into blocks means, for example, obtaining the gradation or coordinates of the vertices of a virtual block obtained by the division, and it is not always necessary to perform actual division or division in space.

  In the representative color determining means, various methods can be adopted as a method for grouping and a method for determining the representative color of the set.

  According to the present invention, the dynamic range of the color gradation of the pixels included in the input image area can be maintained as much as possible without reducing, and the number of colors can be reduced.

1 is a block diagram illustrating an example of a configuration of an image forming apparatus according to an embodiment of the present invention. It is a block diagram which shows the example of a structure of an image compression processing apparatus. It is a figure which shows the example of the area | region divided with the input image. It is a figure for demonstrating the image compression process in an image compression processing apparatus. It is a figure for demonstrating the pre-processing in an image compression processing apparatus. It is a figure for demonstrating the processing operation of DR calculation part in a pre-processing. It is a figure for demonstrating the pixel replacement process in a pre-processing. It is a figure for demonstrating the example of the block division process in a pre-processing. It is a figure which shows the example of this process in case pre-processing is not performed. It is a figure for demonstrating the example of the division | segmentation method according to a frequency component. It is a figure for demonstrating the example of the division | segmentation method according to a dynamic range. It is a figure for demonstrating the example of the division | segmentation method according to the density of gradation. It is a figure for demonstrating the example of the division | segmentation method when the number of colors is smaller than a threshold value. It is a flowchart which shows the example of the outline of a compression process. It is a flowchart which shows the example of the flow of pre-processing. It is a flowchart which shows the example of a division number determination process. It is a flowchart which shows the example of a block division process.

[Entire configuration of image forming apparatus]
FIG. 1 is a block diagram illustrating an example of a schematic configuration of an image forming apparatus 1 according to an embodiment of the present disclosure.

  In FIG. 1, an image forming apparatus 1 includes a CPU (MPU) 10a, a RAM 10b, a ROM 10c, a hard disk 10d, a control circuit 10e, an operation panel 10f, a communication interface 10g, an engine (printing apparatus) 10h, and a scanner (image reading apparatus). 10i or the like.

  The control circuit 10e is a circuit for controlling the hard disk 10d, the operation panel 10f, the communication interface 10g, the engine 10h, the scanner 10i, and the like.

  The operation panel 10f is a touch panel display panel, and displays a screen for giving a message or an instruction to the user, a screen for inputting a processing type and processing conditions desired by the user, and the like. Further, the user can give an instruction to the image forming apparatus 1 by touching a predetermined position of the operation panel 10f.

  The communication interface 10g is a circuit for communicating with a user terminal or a personal computer via a communication line.

  The scanner 10 i is a so-called image scanner, and is a device that reads a figure, a photograph, or the like from a document and stores it as an image in a memory in the image forming apparatus 1.

  One or a plurality of CPUs 10a are provided to perform overall control of the image forming apparatus 1 and necessary processing.

Various programs and data are stored in the ROM 10c and the hard disk 10d, which are read to the RAM 10b as necessary, and the read programs are executed by the CPU 10a.
[Outline of image compression processor]
FIG. 2 shows an example of the functional configuration of the image compression processing device 3 provided in the image forming apparatus 1.

  The function of the image compression processing apparatus 3 is a part of the image processing function provided in the image forming apparatus 1. In the image compression processing device 3, compression processing (image compression processing) is performed on color images in order to reduce the data amount.

  The compression processing of the image compression processing device 3 is fixed-length compression, and includes processing by the preprocessing unit MS (preprocessing) and processing by the processing unit HS (main processing). In the preprocessing, the image of the region ER is subjected to a process of reducing the number of colors while maintaining the color characteristics without reducing the dynamic range of the color gradation of the pixels included therein. In this processing, the pre-processed image of the region ER is compressed by grouping the pixels included therein to create several sets and replacing each set with one representative color.

  In the compression processing, color components in a predetermined color space are extracted for each pixel of the color image. In the present embodiment, the RGB space is used as the color space, and R, G, and B color components are used as the color components. However, the present invention is not limited to this. For example, a YMC space may be used as the color space. The case of 256 gradations will be described as an example of the number of gradations of each color component. However, the present invention is not limited to this. For example, 128 gradations, 64 gradations, and the like may be used.

  Note that the functions of the image compression processing device 3 can be realized by a program for compression processing being executed by the CPU 10a and in combination with an appropriate hardware circuit.

  2, the image compression processing device 3 includes a partition processing unit 21, a compression control unit 22, a comparison unit 23, a DR calculation unit 24, a division number determination unit 25, a block division unit 26, a pixel replacement unit 27, and a pixel concentration unit. 28, a representative color determining unit 29, and the like.

  The DR calculation unit 24, the division number determination unit 25, the block division unit 26, and the pixel replacement unit 27 constitute a preprocessing unit MS, and the pixel concentration unit 28 and the representative color determination unit 29 constitute a main processing unit HS.

  In the image compression processing device 3, an image output from the scanner 10i, an image input via the communication interface 10g, or the like is subjected to preliminary processing as necessary, and input as an input image FN1.

FIG. 3 shows an example of the input image FN1 and the partitioned area ER. FIG. 4 shows the state of the compression process for the image FER in each region ER along the flow of the process. The preprocessing unit MS converts the image FER in FIG. 4A to the image FEM in FIG. 4B. The processing unit HS converts the image FEM in FIG. 4B to the image FEH in FIG. As will be described later, when the number of colors NE of the image FER is smaller than the threshold value X, conversion from the image FER in FIG. 4A to the image FEM in FIG. Without processing, the processing by the processing unit HS is performed on the image FER in FIG.
[Input image area section]
With reference to FIG. 3, the partition processing unit 21 sequentially partitions the input image FN1 into a matrix-like region (block) ER along the main scanning direction and the sub-scanning direction. Each region ER is composed of a plurality of pixels GS arranged in an m × n matrix. The size of the region ER is, for example, 8 × 8 = 64 pixels or 4 × 4 = 16 pixels. The image FER in each region ER or the pixel GS included in each region ER may be referred to as “image FER”.

  The color number NE of the image FER in the region ER is the number of pixels GS having different colors. If all the pixels GS have different colors, the color number NE is equal to the number of pixels GS. Therefore, when the region ER is 8 × 8 = 64 pixels, the maximum value of the number of colors NE is 64. Usually, in a small region ER, the pixel GS often has the same color, and is often smaller than 64. When all the pixels GS have the same color, the color number NE is 1.

  Since the color of the pixel GS is represented by a combination of R, G, and B gradations that are the color components, it can be said that the colors are the same when the gradations of these color components are all the same. However, it is also possible to widen the same color range by allowing a slight difference in gradation.

  In FIG. 4A, the pixels GS included in the region ER are indicated by black dots, and a state in which the pixels GS are distributed in the RGB space is shown. Since the pixels GS of the same color are at the same position, the number of black spots is equal to the number of colors NE. A rectangular parallelepiped KE surrounding each pixel GS indicates the dynamic range DR of each color component of the image FER, as will be described later.

  Now, the image FER of each region ER partitioned by the partition processing unit 21 is sequentially input to the compression control unit 22 and the comparison unit 23.

  2 and 4, the comparison unit 23 checks the gradation of each pixel GS of the image FER and obtains the number of colors NE. In obtaining the color number NE, for example, the counter is incremented each time a pixel GS having a different gradation is input. The count value of the counter is the number of colors NE.

  The comparison unit 23 compares the number of colors NE with the threshold value X and sends the result to the compression control unit 22. The threshold value X is a preset value, for example, set in consideration of the maximum number of representative colors in the processing unit HS. In this case, the threshold value X is the same as the maximum value of the number of representative colors in the processing unit HS, or a predetermined value or a value larger than the predetermined value.

  The compression control unit 22 controls the input image FER so that the content of the compression process is changed according to the number of colors NE. That is, depending on the color number NE and the threshold value X, whether the preprocessing by the preprocessing unit MS is performed before the main processing by the main processing unit HS or only the main processing is performed without performing the preprocessing. Is determined.

  That is, when the color number NE is larger than the threshold value X, the compression control unit 22 performs control so that the processing by the main processing unit HS is performed after the processing by the preprocessing unit MS is performed on the image FER. . When the number of colors NE is smaller than the threshold value X, control is performed so that only the processing of the main processing unit HS is performed on the image FER without performing processing by the preprocessing unit MS.

  The reason is that when the number of colors NE of the image FER is close to the number of representative colors in the main processing unit HS, a remarkable effect cannot be obtained depending on the processing of the preprocessing unit MS.

By doing in this way, when there is no remarkable effect in the processing by the pre-processing unit MS, the processing by the main processing unit HS is performed directly without performing the processing, and the processing time is shortened.
〔Preprocessing〕
Next, preprocessing by the preprocessing unit MS will be described.

  First, the DR calculation unit 24 obtains the dynamic range DR for the gradation of each color component in the color space for the image FER. The dynamic range DR is obtained as a value obtained by subtracting the minimum value from the maximum value for the gradation of each color component.

  FIG. 6 shows an example in which an image FERb of B (blue) component is extracted from the image FER. For the extracted B component image FERb, the maximum value Max and the minimum value Min of the gradation of each pixel are obtained, and the difference (Max−Min) is obtained as the B component dynamic range DRb.

  For example, when the maximum value Max of the gradation of the image FERb is 210 and the minimum value Min is 34, the dynamic range DRb is 176 (= 210−34).

  The dynamic ranges DRr and DRg are similarly obtained for the R and G components that are other color components. When the dynamic ranges DRr, DRg, and DRb for each color component are obtained, as shown in FIG. 5A, a rectangular parallelepiped KE having sides with lengths corresponding to the sizes can be defined.

  FIG. 5A basically has the same contents as FIG. 4A described above, and the distribution of the pixels GS constituting the image FER in the RGB space is indicated by black dots. The number of black spots is equal to the number of colors NE of the image FER. FIG. 5A further shows the length of each side of the rectangular parallelepiped KE, that is, the dynamic ranges DRr, DRg, and DRb. Accordingly, all the pixels GS of the image FER are included in the inside of the rectangular parallelepiped KE (including the surface of the rectangular parallelepiped KE).

  The division number determination unit 25 determines the division numbers L, M, and N (L, M, and N are integers of 1 or more) according to the characteristics of the color components of the pixels included in the region ER. Various methods are used for determining the number of divisions.

  For example, when any one of the color components has a frequency component, the division number determination unit 25 sets the division number for the color component to be larger than that of the color component having no frequency component. Further, the number of divisions for a color component having a large dynamic range DR is made larger than that for a color component having a small dynamic range DR.

  Further, each color component may be divided equally (equally) or may be divided unevenly. For example, each color component is finely divided in a partial area where the gradation of the color component of the pixel is high, and is roughly divided in a partial area where the density is low. Details will be described later.

  The block division unit 26 divides the gradation (number of gradations) in the dynamic range DR of each color component into the number of divisions L, M, or N determined by the division number determination unit 25, so that each color component The rectangular parallelepiped KE corresponding to the dynamic range DR is divided into small rectangular parallelepiped blocks BKE.

  In FIG. 5B, the rectangular parallelepiped KE is divided into eight blocks BKE, with each side divided into two equal parts. That is, in this example, the division numbers L, M, and N are all 2, and are divided equally.

  The pixels GS in the rectangular parallelepiped KE are divided so as to be included in any of the eight blocks BKE. Each block BKE is a rectangular parallelepiped of the same size, and has eight vertices. Since the vertices of adjacent blocks BKE are common, there are 27 vertices TT as a whole. Since each vertex TT exists in the RGB space, it has a gradation for the RGB color components.

  FIG. 8 shows another example of the number of divisions and the division method.

  In FIG. 8A, the division numbers L, M, and N are 6, 3, and 2, respectively, and are divided equally. Each block BKE is a rectangular parallelepiped of the same size, has 8 vertices, and has 84 vertices TT as a whole.

  In FIG. 8B, the division numbers L, M, and N are 5, 3, and 3, respectively, and some are divided unevenly. The block BKE includes rectangular parallelepipeds having different sizes, each having eight vertices, and 96 TTs as a whole.

  The pixel replacement unit 27 performs a process of reducing the number of colors NE of the image FER by replacing the gradation of each color component of the pixel GS included in each block BKE with the gradation of the nearest vertex TT of each block BKE. Do.

  As shown in FIG. 7, in each block BKE, all the pixels GS included in the block BKE move to the nearest vertex TT. In determining the destination vertex TT of each pixel GS, here, the vertex TT having the smallest linear distance between the pixel GS and the vertex TT is selected. Then, the gradation of each pixel GS is replaced with the gradation of the destination vertex TT.

  That is, by the processing of the pixel replacement unit 27, the pixel GS of the image FER shown in FIG. 5B moves to the vertex TT closest to each pixel GS among the vertices TT of each block BKE. .

  As a result, as shown in FIG. 5C, each pixel GS is aggregated to one of the vertices TT, and as shown in FIG. 5D, converted into an image FEM composed of the pixels GS aggregated at the vertices TT. Is done. As the pixels GS are aggregated at the vertex TT, the color number NEM of the image FEM is smaller than the color number NE of the image FER.

  Note that the image FEM illustrated in FIG. 5D corresponds to the image FEM illustrated in FIG.

  In this way, the image FEM in which the color number NE is reduced to the color number NEM is output from the preprocessing unit MS.

  In the example shown in FIG. 5, the maximum value of the number of colors NE in the original image FER is 64, but the maximum value of the number of colors NE in the image FEM output from the preprocessing unit MS is reduced to 27.

  Moreover, the size of the rectangular parallelepiped KE is not changed while the image FER is converted into the image FEM by the preprocessing unit MS. That is, since the pixels GS of the original image FER are collected at the vertex TT of each block BKE, the dynamic range DR of the original image FER is maintained without reducing the dynamic range DR of each color component.

  That is, since the dynamic range DR of the original image FER is determined by the gradation of the pixels GS constituting the image FER, at least one pixel GS exists on all the outer peripheral surfaces of the rectangular parallelepiped KE. When the pixels GS existing on the outer peripheral surface of the rectangular parallelepiped KE are aggregated to the vertex TT, the pixels GS are aggregated to any vertex TT existing on the outer peripheral surface of the rectangular parallelepiped KE. Therefore, the rectangular parallelepiped KE defined by the dynamic range DR is There is no change before and after aggregation. Therefore, there is no reduction in the dynamic range DR.

  Further, since the pixels GS of the original image FER move to the vertex TT of the block BKE, it can be said that the distribution of the pixels GS is unevenly distributed in the obtained image FEM.

Note that, as described above, when the number of colors NE is smaller than the threshold value X, this processing is performed directly on the image FER without performing preprocessing. In this case, that the original image FER is input to the main processing unit HS is the same in the compressed content as the image FER input to the preprocessing unit MS is directly output to the main processing unit HS as an image FEM. It is. Therefore, not performing pre-processing is equivalent to making the number of divisions L, M, and N equal to the number of gradations in the dynamic ranges DRr, DRg, and DRb of each color component in the compressed content.
[Main processing]
Next, the main process in the main processing unit HS will be described. The configuration and processing content of the processing unit HS can use conventionally known compression processing, and can be various other than those described below.

  The pixel concentration unit 28 collects the pixels GS of the image FEM that have been preprocessed and having a short distance in the color space into a set.

  That is, the pixels GS of the image FEM shown in FIG. 4B are gathered as shown in FIG. 4C to generate a set YG of pixels GS. FIG. 4C schematically shows eight generated sets YG. In FIG. 4C, the set YG with more pixels GS is shown to have a larger size (volume).

  The representative color determining unit 29 reduces the number of colors of the input image FEM by determining the representative color of each set YG generated by the pixel collecting unit 28. That is, since one representative color is determined for each set YG, the image FEH output from the representative color determination unit 29 has the same number of colors NEH as the number of sets YG.

  It should be noted that in determining the representative color, various methods, for example, using the average value or repeat value of the gradation of each color component of the pixel GS included in each set YG, or the pixel GS near the center position of the set YG It is possible to use gradation.

  By the way, when the set YG is generated in the pixel collecting unit 28, the smaller the number of colors NEM of the input image FEM, the smaller the spread of the pixels GS, and thus the size of each set YG tends to be small. Therefore, the smaller the number of colors NEM of the image FEM, the smaller the gradation change of the pixels GS included in the image FEM when the representative color determining unit 29 determines the representative color, and the dynamic range DR is easily maintained.

  Here, the processing when the preprocessing is not performed, that is, when the image FER is input to the processing unit HS will be described with reference to FIG.

  When the image FER shown in FIG. 9A is input to the pixel collecting unit 28, the pixel collecting unit 28 collects the pixels GS as shown in FIG. 9B to generate a set YG. The set YG generated here has a wide distribution of the pixels GS of the image FER, so that the size of the set YG becomes large. Therefore, as shown in FIG. 9C, when the representative color determining unit 29 determines the representative color, the gradation change of the pixel GS included in the image FER is likely to be large. For this reason, the dynamic range DR is easily reduced, and the dynamic range DR of the original image FER is not maintained.

  In particular, when the image FER is a halftone image, the number of colors NE increases, and the distribution of pixels GS, that is, the distribution of gradations, is wide. For this reason, the size of the set YG tends to increase, and the change in gradation becomes large, so that the dynamic range DR tends to be reduced as a whole image.

  As described above, when the number of colors NE of the image FER is small, particularly when the number of colors NE is smaller than the threshold value X, the dynamic range can be directly applied to the image FER. There is almost no reduction in DR.

  Note that only the number of sets YG, that is, the number of colors NEH of the output image FEH, depends on the amount of output image data, and the number of divisions by preprocessing does not depend on the output image data.

The compression control unit 22 has the functions described below in addition to the functions described above.
[Correction of number of divisions by frequency component]
That is, although not shown, the compression control unit 22 is provided with a frequency analysis unit that analyzes the frequency component of the input image FER. The frequency analysis unit analyzes whether or not there is a frequency component for each color component of the image FER, the size of the frequency component, and the like.

  For example, when the image FER is a halftone image, it often has a frequency component. The frequency component occurs when the gradation is repeated in one color component or the same color component appears intermittently.

  Further, when there is no specific color component for all the pixels GS of the image FER, and when the gradation of a specific color component gradually changes from large to small or from small to large, there is no frequency component. I can say that.

  The compression control unit 22 instructs the division number determination unit 25 to correct the division numbers L, M, and N depending on whether or not the color component has a frequency component.

  For example, in FIG. 10, the R component in the image FER has a frequency component, the G component has a monotone change, and thus has no frequency component, and the B component has only a single gradation and has no frequency component. It is shown.

  In this case, when the frequency component is not considered, the division numbers L, M, and N of the respective color components are all four. When the frequency component is considered, since the R component has a frequency component, a command is issued from the compression control unit 22 to the division number determination unit 25 so as to increase the division number L of the R component. The number L is increased and corrected to 7.

As a result, the change in gradation of the R component having a frequency component in the compression process is reduced, and the reproducibility of the image FER is improved.
[Change of processing contents and correction of division number by dynamic range]
Further, the compression control unit 22 determines the size of the dynamic range DR calculated by the DR calculation unit 24. That is, the dynamic range DR of each color component is compared with the threshold value Y, and when the dynamic range DR is smaller than the threshold value Y, control is performed so that only this processing is performed without performing preprocessing.

  The reason is that when the dynamic range DR is smaller than the threshold value Y, the gradation change in the color component is very small, so that the dynamic range DR is reduced even if only the main processing by the main processing unit HS is performed. This is because it hardly occurs.

  In addition, when the dynamic range DR of each color component varies greatly depending on the color component, the length of each side of the divided block BKE, that is, the level for each color component, when the division numbers L, M, and N are the same. The change in key is greatly different.

  In that case, if the state is left as it is, in the pixel replacement process in the pixel replacement unit 27, the variation value of the gradation greatly varies depending on each color component, which may cause the color balance of the pixel to be greatly different and the color balance to be lost. is there.

  In order to prevent this, the compression control unit 22 instructs the division number determining unit 25 to change the division numbers L, M, and N in accordance with the dynamic range DR of the color component.

  That is, the division number determination unit 25 issues a command to correct the division numbers L, M, and N so that the intervals between the divided gradations of the components are the same.

  For example, in FIG. 11, the R component in the image FER has a larger dynamic range DR than the G component, and the G component has a larger dynamic range DR than the B component. Are determined to be in the order of the R component, the G component, and the B component, respectively, 7, 5, and 4.

As a result, the color balance of the pixels does not vary greatly, and the color balance is maintained well.
[Modification of division method by gradation dispersion]
Further, the compression control unit 22 instructs the division number determination unit 25 to correct the division numbers L, M, and N according to the gradation distribution (distribution) of each color component.

  That is, the gradation of each color component may be unevenly distributed within each dynamic range DR. In that case, the partial region having a low gradation density does not need to be finely divided unlike the partial region having a high gradation density. In addition, for a partial region having a high gradation density, the gradation variation in the pixel replacement unit 27 can be reduced by increasing the number of divisions.

  For this reason, the compression control unit 22 calculates gradation dispersion values of the respective color components, and if the gradation values are extremely unevenly distributed, it is not divided evenly but is a partial region having a high gradation density. The division number determining unit 25 is instructed to divide each of the partial areas, which are finer and more coarse for the partial area having a low gradation density.

  For example, FIG. 12 shows that for a certain color component gradation, the gradation density is high in a partial region having a high gradation. In this case, the number of divisions is increased for a high-density partial region, and the number of divisions is decreased for a low-density partial region.

Thereby, the gradation variation in the pixel replacement unit 27 can be reduced.
[Change of processing contents depending on the number of colors]
The compression control unit 22 omits preprocessing according to the number of colors NE in the image FER.

  That is, as described above, when the color number NE is smaller than the threshold value X, the compression control unit 22 performs control so that only the main processing is performed without performing preprocessing on the image FER. As a result, the processing time may be shortened.

  For example, in FIG. 13, since the number of colors NE is small, the gradation of each color component of each pixel GS is maintained as it is.

  That is, for example, a document created using a personal computer or a document once copied often has a small number of colors NE of the image FER. In that case, there is a case where the set YG is not joined in the present process and the lossless compression is possible. For this reason, when the number of colors NE of the image FER is equal to or less than the threshold value X, the gradation of each color component is made equal to one gradation, so that the input image FER is output as it is. can do. In this case, for example, the division numbers L, M, and N may be set as the number of gradations in the dynamic range DR.

The threshold value X can be set in advance as a fixed value by the user. It is also possible for the user to change the compression rate by changing the threshold value X in various ways.
[Explanation by flowchart]
Next, the flow of compression processing in the image compression processing device 3 will be described using a flowchart.

  FIG. 14 shows an example of a schematic flow of compression processing.

  In FIG. 14, it is determined whether or not pre-processing is necessary (# 1). If yes, pre-processing is performed in step # 2, and if no, the process jumps to main processing in step # 3. Whether or not pre-processing is necessary is determined based on the number of colors NE of the image FER, the size of the dynamic range DR, and the like as described above. Further, as described above, even when the preprocessing is not performed, the image FER may be input to the preprocessing unit MS and passed therethrough.

  FIG. 15 shows an example of preprocessing.

  In FIG. 15, the dynamic range DR of each color component is obtained (# 11), the division number determination process is performed (# 12), and the block division process is performed (# 13). Then, the process of calculating the gradation value of the vertex TT is performed (# 14), the process of searching for the closest vertex TT for each pixel GS is performed (# 15), and the vertex TT searched for the gradation of each pixel GS. The process of replacing with the gradation is performed (# 16).

  FIG. 16 shows an example of the division number determination process.

  In FIG. 16, for each color component of the input image FER, the number of colors NE is counted (# 21), and it is determined whether the count is equal to or greater than a threshold value X (# 22). If no in step # 22, the process proceeds to step # 30, where the number of divisions is the same as the gradation of the dynamic range DR and no preprocessing is performed.

  If yes in step # 22, the frequency components of each color component of the image FER are analyzed (# 23). When there is no frequency component (No in # 24), the process proceeds to Step # 30. If there is a frequency component (Yes in # 24), the dynamic range DR is obtained in Step # 25. It is determined whether or not the dynamic range DR is greater than or equal to the threshold value Y (# 26). If no, the process proceeds to step # 30.

  When the dynamic range DR is equal to or greater than the threshold value Y, the gradation variance is calculated (# 27), and the division numbers L, M, and N are determined based on the variance (# 28, 29).

  FIG. 17 shows an example of block division processing.

  In FIG. 17, the division number (here, L) and the dynamic range DR of the color component are input (# 41). The dynamic range DR is divided by the division number L (# 42). Assuming that the variable n is 1 to L, (n · DR / L) is calculated (# 43). Each obtained value is set as the gradation value of each vertex TT (# 44).

  That is, in this way, in the block division processing, the rectangular parallelepiped KE is not actually divided, but the gradation of each vertex TT of the block BKE obtained by the division may be obtained.

  According to the image forming apparatus 1 of the embodiment described above, the dynamic range DR of the color gradation of the pixel GS included in the region ER of the input image FN1, that is, the pixel GS included in the image FER is reduced. The number of colors can be reduced while maintaining as much as possible.

  In the embodiment described above, the pre-processing unit MS, the main processing unit HS, the compression control unit 22, the image compression processing device 3, or the configuration of each unit or the entire image forming device 1, circuit, number, processing content, processing order The processing timing can be appropriately changed in accordance with the gist of the present invention.

DESCRIPTION OF SYMBOLS 1 Image forming apparatus 3 Image compression processing apparatus 21 Partition processing part 22 Compression control part 23 Comparison part 24 DR calculation part 25 Division number determination part 26 Block division part 27 Pixel replacement part 28 Pixel concentration part 29 Representative color determination part MS Pre-processing part HS Main processing unit DR Dynamic range L, M, N Number of divisions KE Cuboid BKE Block TT Vertex

Claims (11)

For the pixels included in the input image area, obtain the dynamic range for the gradation of each color component in the color space,
By dividing the dynamic range of each color component into division numbers of L, M, or N (L, M, and N are integers of 1 or more), the rectangular parallelepiped corresponding to the dynamic range of each color component is converted into a small rectangular parallelepiped block. Split and
Performing pre-processing to reduce the number of colors in the region of the image by replacing the gradation of each color component of the pixels included in each of the blocks with the gradation of the nearest vertex of the block;
For the pre-processed pixels, this processing for reducing the number of colors in the region of the image is performed by grouping together pixels having a short distance in the color space into a set and determining a representative color of each set. ,
An image compression processing method. The L, M, and N are different division numbers according to the characteristics of the color components of the pixels included in the image area.
The image compression processing method according to claim 1. When any one of the color components has a frequency component, the number of divisions for the color component is larger than a color component that does not have a frequency component.
The image compression processing method according to claim 1. Make the number of divisions for color components with a large dynamic range larger than those with a small dynamic range.
The image compression processing method according to claim 1. For each color component, finely divide in the partial area where the gradation of the color component of the pixel is high, and divide it roughly in the partial area where the density is low.
The image compression processing method according to claim 1. When the number of colors of pixels included in the input image area is smaller than a set threshold value, the preprocessing is not performed and the pixels included in the input image area Do this process,
The image compression processing method according to claim 1. DR calculation means for obtaining a dynamic range for the gradation of each color component in the color space for the pixels included in the input image area;
By dividing the dynamic range of each color component into division numbers of L, M, or N (L, M, and N are integers of 1 or more), the rectangular parallelepiped corresponding to the dynamic range of each color component is converted into a small rectangular parallelepiped block. Block dividing means for dividing;
Pixel replacement processing means for performing processing for reducing the number of colors in the region of the image by replacing the gradation of each color component of the pixels included in each of the blocks with the gradation of the nearest vertex of the block;
A representative that reduces the number of colors in the region of the image by grouping together pixels having a short distance in the color space into a set and determining a representative color of each set for the pixels that have been processed by the pixel replacement processing means Color determining means;
An image compression processing apparatus comprising: Division number determining means for determining L, M, and N, which are the number of divisions, according to the characteristics of each color component of a pixel included in the region of the image;
The image compression processing apparatus according to claim 7. A comparison means for comparing the number of colors of pixels included in the input image area with a set threshold value;
When the number of colors is smaller than the threshold value, the processing by the representative color determination unit is performed on the pixels included in the input image area without performing the processing by the pixel replacement processing unit. Compression control means to control;
The image compression processing device according to claim 7 or 8.   An image forming apparatus comprising the image compression processing apparatus according to claim 7. For the pixels included in the input image area, obtain the dynamic range for the gradation of each color component in the color space,
By dividing the dynamic range of each color component into division numbers of L, M, or N (L, M, and N are integers of 1 or more), the rectangular parallelepiped corresponding to the dynamic range of each color component is converted into a small rectangular parallelepiped block. Split and
Reducing the number of colors in the area of the image by replacing the gradation of each color component of the pixels contained in each of the blocks with the gradation of the nearest vertex of the block;
An image processing method.

Images