Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
  • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
  • H04N1/46—Colour picture communication systems
  • H04N1/56—Processing of colour picture signals
  • H04N1/60—Colour correction or control
  • H04N1/6058—Reduction of colour to a range of reproducible colours, e.g. to ink- reproducible colour gamut
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
  • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
  • H04N1/46—Colour picture communication systems
  • H04N1/64—Systems for the transmission or the storage of the colour picture signal; Details therefor, e.g. coding or decoding means therefor
  • H04N1/644—Systems for the transmission or the storage of the colour picture signal; Details therefor, e.g. coding or decoding means therefor using a reduced set of representative colours, e.g. each representing a particular range in a colour space

Definitions

• the present invention relates to a
• predetermined block predetermined block, and that of performing image processing and restoration processing for the image compressed for each block.
• peripheral but also image processing apparatuses such as a digital camera and a facsimile apparatus compress color image data, which allows for reduction in cost and a gain in speed.
• image processing apparatuses such as a digital camera and a facsimile apparatus compress color image data, which allows for reduction in cost and a gain in speed.
• a JPEG scheme that uses discrete cosine transformation and a scheme that uses wavelet transformation are widely employed as color still image compression schemes.
• Encoding schemes of these kinds generally encode an image for each predetermined block
• variable-length coding for each block is adopted, so random access for referring to a small area is
• compressed data becomes necessary to refer to and convert the pixel data having undergone this image compression. In other words, image processing cannot be performed for compressed data intact, so its
• decoding processing inevitably becomes necessary. This makes it necessary to process all pixels in high- resolution data for each pixel, thus prolonging the processing time.
• a known run length encoding scheme of storing pixel data and their run lengths and a technique (for example, Japanese Patent Laid-Open No. 10-257488) of compressing pixel data by detecting an edge for each block and storing two colors of this edge have been disclosed.
• decoding processing inevitably becomes necessary upon image processing for compressed data, so the processing time considerably prolongs, depending on the number of pixels to be processed.
• high- resolution data is not always necessary depending on the purpose of use of the image data, and therefore must be reduced to low-resolution data. That is, to generate low-resolution data from high-resolution data having undergone compression coding by, for example, JPEG, decoding processing and special reduction
• the present invention provides an apparatus and method which can minimize image deterioration due to factors associated with color-reduction and perform high-efficiency compression when image data is
• an image processing apparatus characterized by comprising: first
• specification means for dividing first image data into blocks each with a size of 2 ⁇ 2 pixels, sequentially determining the divided blocks as first processing objects, and comparing color data of respective pixels in the block which is the first processing object, thereby specifying a first pattern flag indicating a placement pattern of the color data included in the block;
• first extraction means for extracting, as first color data, color data corresponding to a pixel at a predefined position in the block with the size of 2 ⁇ 2 pixels, and, if the number of colors included in the block which is the first processing object is one of 2 to 4, further extracting second color data to fourth color data corresponding to the placement pattern defined in the specified, first pattern flag;
• first output means for outputting the first pattern flag of each block specified by the first specification means, the first color data in each block extracted by the first extraction means, and the second color data to fourth color data in each block extracted by the first extraction means;
• second color-reduction means for dividing second image data formed from the first color data output from the first output means into blocks each with a
• second extraction means for extracting, as first color data, color data corresponding to a pixel at a predefined position in the block with the size of 2 ⁇ 2 pixels after the color-reduction processing is performed by the second color-reduction means, and, if the number of colors included in the block which is the second processing object is one of 2 to 4, further extracting second color data to fourth color data corresponding to the placement pattern defined in the specified, second pattern flag; and second output means for outputting the second pattern flag of each block specified by the second specification means, the first color data in each block extracted by the second extraction means, and the second color data to fourth color data in each block extracted by the second extraction means.
• an image processing apparatus which performs compression processing for image data upon determining blocks each including a predetermined number of pixels as blocks of interest to be processed, characterized by comprising: color- reduction means for performing color-reduction
• specification means for comparing color data of respective pixels in the block of interest, thereby specifying a placement pattern of the color data included in the block of interest; extraction means for extracting, as first color data, color data corresponding to a pixel at a predefined position in the block, and further
• control means for controlling to execute color-reduction processing by the color- reduction means, for second image data formed from the first color data output from the output means, upon determining blocks each including a predetermined number of pixels as blocks of interest to be processed, thereby further controlling to re-execute, for the block of interest having undergone the color-reduction processing by the color-reduction means, processing by specification means, the extraction means, and the output means, wherein in executing the color-reduction processing for the second image data formed from the first color data output from the output means, upon determining blocks each including a predetermined number of pixels as blocks of interest to be processed, the color-reduction means executes the color-reduction processing while weighting a pixel value of each pixel included in the block of interest to be processed, based on the placement pattern specified by the
• an image processing method characterized by comprising: a first specification step of using first specification means to divide first image data into blocks each with a size of 2 ⁇ 2 pixels, sequentially determine the divided blocks as first processing objects, and compare color data of
• an image processing method of performing compression processing for image data upon determining blocks each including a predetermined number of pixels as blocks of interest to be processed characterized by comprising: a color-reduction step of using color-reduction means to perform color-reduction processing for the block of interest; a specification step of using specification means to compare color data of respective pixels in the block of interest, thereby specifying a placement pattern of the color data included in the block of interest; an extraction step of using extraction means to extract, as first color data, color data corresponding to a pixel at a
• a control step o using control means to control to execute color- reduction processing by the color-reduction means, for second image data formed from the first color data output in the output step, upon determining blocks eac including a predetermined number of pixels as blocks o interest to be processed, thereby further controlling to re-execute, for the block of interest having undergone the color-reduction processing in the color- reduction step, processing in the specification step, the extraction step, and the output step, wherein in the color-reduction step, in executing the color- reduction processing for the second image data formed from the first color data output in the output step, upon determining blocks each including a predetermined number of pixels as blocks of interest to be processed the color-reduction processing is executed while weighting a pixel value of each pixel included in the block of interest to be processed, based on the
• the present invention it is possible to compress image data while minimizing image deterioration due to factors associated with color- reduction and keeping the processing cost low. Using the data compressed by the present invention, the processing cost can be kept low, thus facilitating image processing.
• FIG. 1 is a block diagram illustrating an example of the configuration of an image processing apparatus according to an embodiment
• FIG. 2 is a block diagram illustrating an example of the configuration of a controller shown in Fig. 1;
• FIG. 3 is a view showing possible
• Fig. 4 is a view showing the relationship between the patterns shown in Fig. 3 and pattern flags;
• Fig. 5 is a flowchart showing processing of specifying different pixels and a pattern flag
• Fig. 6 is a view for explaining processing of determining the positions of pixels with different pixel values
• Fig. 7 is a view illustrating an example in which image data is written in a memory space by DMACs
• FIGs. 8A and 8B are flowcharts showing compression processing according to the first
• Figs. 9A, 9B, 9C, 9D, and 9E are views showing transitions of image data obtained when
• compression processing is performed independently in two stages
• Figs. 10A, 10B, IOC, 10D, and 10E are views showing image data obtained when two-stage compression processing is performed based on the first-stage compression processing result;
• FIGs. 11A, 11B, 11C, 11D, and HE are views showing image data obtained when two-stage compression processing is performed based on the first-stage compression processing result;
• Fig. 12 is a view showing the relationship between the input and output when a new pattern flag is added
• FIG. 13 is a flowchart showing compression processing according to the second embodiment
• Fig. 14 is a view illustrating an example in which image data is written in a memory space in the second embodiment
• FIG. 15 is a flowchart showing compression processing according to the third embodiment
• Fig. 16 is a view showing the relationship among a page, tiles, and blocks;
• Fig. 17 is a view showing the structure of packed data
• Fig. 18 is a table illustrating an example of the configuration of a packet management table
• Fig. 19 is a view showing the address of each packet written in a memory space
• Fig. 20 is a flowchart showing multistage compression processing according to the third
• Fig. 21 is a view showing the structure of packed data obtained upon multistage compression processing .
• a digital MFP MultiFunction Peripheral
• functions such as scanning, printing, and copying
• a controller 101 is connected to a scanner 102 serving as an image input device, and a printer 103 serving as an image output device.
• the controller 101 is also connected to a network 104 such as a LAN or a public line (WAN) to input/output image information or device information and perform image rasterization of PDL data.
• a CPU 105 serves as a processor which controls the overall MFP in accordance with a program stored in an HDD storage unit 107.
• a memory 106 serves as a system working memory for
• the HDD storage unit 107 serves as a hard disk drive and stores, for example, system software, a program, and image data.
• a scanner image processing unit 201 receives image data in three colors: R, G, and B (Red, Green, and Blue) read by the scanner 102, and performs image processing such as shading processing and filter processing for the received image data, and a compression unit 202 performs image compression processing for the processed data.
• a DMAC (Direct Memory Access Controller) 203 stores the compressed data in the memory 106 via an image memory bus .
• the color processing unit 212 transfers the compressed data stored in the memory 106 to a color processing unit 212 via the image memory bus.
• the color processing unit 212 converts the transferred data into data in a CMYK (Cyan, Magenta, Yellow, and Black) color space.
• CMYK Cyan, Magenta, Yellow, and Black
• a DMAC 221 reads the compressed data stored in the memory 106, and a rasterization unit 222 rasterizes the read data into raster image data.
• a printing image processing unit 223 inputs the raster CMYK image data, performs its area coverage modulation processing using the dither method or the error diffusion method, and outputs the processed data to the printer 103.
• the DMAC 211 transfers the compressed data stored in the memory 106 to the color processing unit
• the processing unit 212 performs, for example, display gamma adjustment and sheet self-color adjustment for the transferred data, it converts the adjusted data into data in a YCbCr (luminance, blue color difference, and red color difference) color space.
• the DMAC 211 stores the data, processed by the color processing unit 212, again in the memory 106 via the image memory bus.
• a DMAC 231 transfers the compressed data stored in the memory 106 to a rasterization unit 232 via the image memory bus.
• the rasterization unit 232 rasterizes the
• a transmission unit 233 performs JPEG compression processing for the raster YCbCr image data in case of color image
• the disk spool high- compression/expansion unit 242 transfers the compressed data stored in the memory 106 to a disk spool high-compression/expansion unit 242 via the image memory bus.
• the disk spool high- compression/expansion unit 242 has a writing speed on the HDD, which is slower than that on the memory, and therefore performs JPEG compression at a higher
• the compressed data is stored in the HDD storage unit 107 via a disk access controller 243. Also, to transfer the compressed data stored in the HDD storage unit 107 to the memory 106 again, the above-mentioned processing need only be performed in reverse order.
• FIG. 2 does not show a PDL (Page Description Language) interpretation unit
• the CPU 105 which functions as the PDL interpretation unit interprets the PDL data, and outputs a display list obtained as a result of the interpretation to the memory 106.
• a rendering unit 251 renders the display list stored in the memory 106 into raster RGB image data
• a compression unit 252 performs its image compression processing.
• a DMAC 253 stores the compressed data in the memory 106 via the image memory bus.
• the PDL image data can be printed, transmitted to the network, and stored by processing similar to that for the scanned image data.
• the present invention is not limited to such a configuration.
• a common compression unit may be provided instead of individually providing the compression units 202 and 252 as in the case of Fig. 2.
• a block includes four pixels, so a maximum of four colors occupy the block, and
• Fig. 3 shows the number of possible
• the first or second color is assigned to the three pixels other than the upper left pixel, so a total of seven combinations are possible upon excluding a combination when the four pixels have the same color.
• a total of 15 patterns are possible upon counting up the number of patterns in all cases of one to four colors. Considering that flags (identifiers) are assigned to all of these patterns to specify them, 4 bits are necessary as the amount of data of each flag. The relationship between these 15 patterns and the flags is defined as shown in Fig. 4, and these flags will be referred to as "pattern flags" hereinafter.
• an image having, for example, R, G, and B components each with 8 bits and 256 gray levels is used as an input, and a 24-bit image for each pixel in a dot sequential system of 8-bit data is used as data.
• the upper left, upper right, lower left, and lower right pixels of a block including 2 x 2 pixels have coordinate positions 1, 2, 3, and 4, respectively (reference numeral 401 shown in Fig. 4). Since the combinations of two pixels have a total of six sets of coordinate positions 1-2, 1-3, 1-4, 2-3, 2- 4, and 3-4 (reference numeral 402 shown in Fig. 4), six comparisons must be made, and 6 bits are output as a result. As in the comparison results shown in Fig. 4, if all of the four pixels have the same color, Is are output as all comparison results; or if all of the four pixels have different pixel values, 0s are output as all comparison results.
• 15 patterns may occur from color matching among the four pixels, so a 4-bit pattern flag is specified in accordance with the 6-bit comparison result (S503) , as shown in Fig. 4.
• the number of colors that have occurred in the four pixels and color data are extracted (S504).
• the pattern flag is defined such that the color (pixel value) of the upper left pixel becomes the first color
• the pattern flag is 1 to 7, the number of colors is 2, so the color (pixel value) of the upper left pixel is extracted as a first color, and that of a pixel at a position with a second color (second color data) defined in accordance with each pattern flag is further extracted. For example, when the pattern flag is 1, the color of the upper right pixel is extracted as a second color. If the pattern flag is 8 to D, the number of colors is 3, so the color (pixel value) of the upper left pixel is extracted as a first color, and those of pixels at positions with a second color
• second color data and a third color (third color data) which are defined in accordance with each pattern flag are further extracted.
• the pattern flag is 8
• the color (pixel value) of the upper right pixel is extracted as a second color
• that of the lower right pixel is extracted as a third color.
• the pattern flag is E
• the color of the upper left pixel is extracted as a first color
• that of the upper right pixel is extracted as a second color
• that of the lower left pixel is extracted as a third color
• that of the lower right pixel is extracted as a fourth color .
• the number of colors in the block is specified (S505, S507, or S509) based on the pattern flag (or the comparison result) , and the pattern flag and color data corresponding to the specified number of colors are output (S506, S508, S510, or S511) .
• the output data will be described in more detail with reference to Fig. 6.
• step S505 if the pattern flag is 0 (that is, the four pixels have only one color) (YES in step S505), none of the second and subsequent colors exist, so 4 bits of the pattern flag and the pixel value (color data corresponding to 24 bits) of the first color are output (S506) . If the pattern flag is 1 to 7 (that is, the four pixels have two colors) (YES in step S507), the coordinate position of a pixel with the second color is calculated based on the pattern flag. Then, 4 bits of the pattern flag and the pixel values (color data corresponding to 48 bits) corresponding to two colors are output (S508) . If the pattern flag is 8 to D (that is, the four pixels have three colors) (YES in step S509) , 4 bits of the pattern flag and the pixels values (color data corresponding to 72 bits) corresponding to three colors are output
• step S510 If the pattern flag is E (that is, the four pixels have four colors) (NO in step S509) , 4 bits of the pattern flag and the pixel values (color data corresponding to 96 bits) corresponding to four colors are output (S511) . In other words, color data that have not previously occurred are output from each block in turn from the first color when the pixels in the block are scanned in ascending order of coordinate position (the upper left, upper right, lower left, and lower right pixels have coordinate positions 1, 2, 3, and 4, respectively) .
• the amount of output data can be reduced with relatively simple processing by outputting a 4-bit pattern flag and pixel values equal in number to that of colors existing in the input data.
• the compression ratio of the amount of output data is relatively high.
• the number of colors in the block can be specified by referring to the pattern flag. Performing such processing for all image blocks allows data compression of the entire image field.
• a threshold according to which pixels are regarded as having the same color may be held in, for example, the memory 106 in advance or be obtained by accepting designation by the user via the network 104.
• a user interface intended for this operation is provided to implement a designation means.
• color-reduction to one color is realized by calculating, for example, the average pixel value of the four pixels. The difference between the pixel value obtained by color-reduction and the input pixel values of the four pixels is calculated
• the error magnitude is determined (S804). For example, the sum of the absolute values of the
• step S804 that the difference is small. That is, if it is determined in step S804 that the difference is small, it is determined that color- reduction to one color is to be performed, and the pattern flag is specified as 0 (S805) .
• the pixel value of one color obtained by color-reduction, and the specified pattern flag are output (S806) .
• step S804 If it is determined in step S804 that the error is large, color-reduction processing to two colors is performed subsequently (S807).
• two pixels A and B with a largest difference in R, G, and B values among the four pixels, for example, are extracted, and the two remaining pixels are clustered in accordance with whether they are similar to pixel A or B.
• Color-reduction to two colors is performed by obtaining the average of the pixel values in each cluster.
• the difference between the pixel values obtained by color-reduction to two colors and the input pixel values of the four pixels is calculated (S808).
• the error magnitude is determined based on the
• step S809 If it is determined in step S809 that the difference is small, a pattern flag is specified in accordance with the positions of the pixel values upon color-reduction to two colors, based on a correspondence as shown in Fig. 4 (S810) .
• the pixel values of two colors obtained by color-reduction, and the specified pattern flag are output (S811) .
• step S809 If it is again determined in step S809 that the error is large, color-reduction processing to three colors is performed subsequently (S812) .
• step S814 If it is determined in step S814 that the difference is small, a pattern flag is specified in accordance with the positions of the pixel values upon color-reduction to three colors, based on a correspondence as shown in Fig. 4 (S815) .
• the pixel values of three colors obtained by color-reduction, and the specified pattern flag are output (S816) .
• step S814 If it is again determined in step S814 that the error is large, it is determined that the block of interest may pose a problem in terms of image quality upon color-color reduction, so the pattern flag is specified as E (S817). The pixel values of all of the four pixels, and the specified pattern flag are output (S818) .
• steps are sequentially compared with each other, thereby deciding the number of colors to which color-reduction is performed.
• pattern flags of respective blocks are collectively stored in a pattern flag storage unit (a memory area for storing a pattern flag) on the memory.
• first color data (data of the first color) in respective blocks are collectively stored in a first color storage unit (a memory area for storing first color data in each block) on the memory.
• second to fourth color data in respective blocks are collectively stored in a second/third/fourth color storage unit (a memory area for storing second to fourth color data in each block) on the memory.
• Fig. 7 is a view illustrating an example in which image data is written in the memory space by the DMACs. Note that when an image having a size of M ⁇ N pixels and R, G, and B components each with 8 bits is input to the compression unit, the pattern flag storage unit which stores pattern flag data has a data size of (M x N x 4/8) bytes. Also, the first color storage unit which stores first color data has a data size of (M x N x 24/8) bytes. Moreover, the
• second/third/fourth color storage unit which stores second, third, and fourth color data (that is, color data other than the first color data) has a data size which differs depending on the type of raster image to be processed. This is because the number of blocks having the second, third, and fourth colors differs between individual images.
• color data of each pixel is stored while the number of its bits is maintained intact. That is, the color (pixel value) of each pixel can be specified even without decoding and restoring compressed data to a raster image, unlike compressed data in, for example, the JPEG format.
• color processing which is completed by one pixel input/output such as color conversion which uses an LUT, gamma correction processing, and color space conversion processing which uses a matrix operation
• the stored data can be directly processed, as shown in Fig. 7, without
• pixel data at memory addresses subsequent to the first color writing start address on the memory 106 are read via the DMAC 211, processed for each pixel, and rewritten on the memory 106.
• the capacity of the memory 106 can be saved by rewriting the processed data at the same location on the memory 106.
• image data is divided into a pattern flag, a first color, and other colors, and discretely stored in the memory.
• pixel values (color data) obtained when a raster image is divided into blocks each including 2 > ⁇ 2 pixels and a pixel at the coordinate position corresponding to the upper left in each block is sampled are continuously present on the memory in the first color storage unit.
• the MFP has functions of, for example, preview display of accumulated PDL image data and scanned image data and the above-mentioned network transmission. For example, even if the print
• thumbnail data when thumbnail data must be obtained, a half-size raster image can be easily obtained by collectively extracting only the first color data stored in the first color storage unit.
• 600-dpi raster image is converted into data as shown in Fig. 7 and accumulated will be described. If 300 dpi is designated as the resolution of the image to be transmitted, the data stored in the first color storage unit need only be directly extracted and transmitted. However, if a resolution higher than 300 dpi sampled in the first color storage unit, such as 400 dpi, is designated, the following processing is performed.
• the compressed data is rasterized once, the rasterized data is scaled using known scaling processing, and the scaled image is transmitted. If a transmission resolution lower than 300 dpi is designated, scaling processing to the
• designated resolution is performed using only the data in the first color storage unit. In this way, data reading can be performed while switching the resolution in accordance with the required image size.
• multistage compression using the half-size image as an input.
• a certain level of redundancy still remains in a thumbnail image of such a high-resolution image, so a high compression ratio is still expected to be obtained even upon multistage compression processing. For example, when compression processing is performed in two stages, the amount of data is expected to
• the original input image data is defined as first image data
• image data having undergone compression processing once is defined as second image data
• a block in the first image data is defined as a block which is a first processing object
• that in the second image data is defined as a block which is a second processing object.
• compression processing for the first image data includes a first color-reduction step, first specification step, first extraction step, and first output step, and that for the second image data
• Figs. 9A, 9B, 9C, 9D, and 9E the pixel value of each pixel is expressed by 0 to 255 in Figs. 9A, 9B, 9C, 9D, and 9E.
• the average or error is calculated using the pixel value in an RGB color space.
• color information (first to fourth color data) in each block are obtained by the first-stage compression processing.
• color-reduction processing may be performed in the first-stage
• the pattern flag of two colors, the pixel value of the lower right pixel, and the pixel value (Fig. 9D) of the upper left pixel, are output as the second-stage
• each of the upper left and lower right blocks shown in Fig. 9A has only one color.
• the values of the upper left and lower right pixels shown in Fig. 9B obtained by sampling the upper left pixels in the input blocks are expanded based on the pattern flag, the expansion result is copied and propagated to four pixels, as shown in Fig. 9A.
• Fig. 10A shows 4 x 4 pixels which are used as an input and are identical to those shown in Fig. 9A
• Fig. 10B shows an image which is obtained by the first-stage compression and is identical to that shown in Fig. 9B.
• the color-reduction value is calculated in consideration of the range over which each pixel exerts an influence in the first-stage compression .
• the upper left pixel influences four pixels
• the upper right pixel influences one pixel
• the lower left pixel influences one pixel
• the lower right pixel influences one pixel
• pattern flag "0" is assigned to the upper left and lower right pixels (all pixels have the same color) . That fact can also be determined because pattern flag "4" is assigned to the upper right and lower left pixels. In this manner, it can be
• the pixel value (Fig. 10D) of the upper left pixel is finally output as a first color, together with a pattern flag (pattern flag "6" indicating two colors) , and a second color (the pixel value of the lower right pixel), as in the case of Fig. 9C.
• the data shown in Fig. 10D is expanded into that as shown in Fig. IOC first, and the data shown in Fig. IOC is further expanded into that as shown in Fig. 10E.
• the sum of errors with respect to the original data in each block is 36 in the upper left block, 23 in the upper right block, 13 in the lower left block, and 0 in the lower right block, so none of them exceed the allowable value.
• the error with respect to the input is 72 in total, so the error is smaller than that of the data shown in Figs. 9A, 9B, 9C, 9D, and 9E.
• FIG. 11A An original image shown in Fig. 11A is identical to that in the above-mentioned example, and a first color-reduction processing result shown in Fig. 11B is also identical to that shown in Fig. 10B.
• the second-stage color-reduction processing is
• threshold is set smaller than 45 (for example, to 40) , color-reduction to three colors alone is tolerated.
• the error can be reduced by adding data on one pixel to the 4 4 pixels in terms of data.
• the second-stage compression processing uses an image with a resolution half that of an image used in the first-stage compression processing, so the tolerance for image deterioration is lower in the second-stage compression processing than in the first- stage compression processing. In other words, spread of deterioration over a wide range can be prevented by limiting the tolerance for deterioration to a larger degree as the stage level increases. [0073] In this manner, by performing color- reduction processing in consideration of the range over which each pixel has an influence on neighboring pixels before the second- and subsequent-stage compression, image deterioration can be minimized although the amount of data may often increase upon this processing.
• the rasterization units 222 and 232 paired with the compression units 202 and 252, respectively, will be described next.
• the rasterization units 222 and 232 perform processing of restoring a pattern flag and color data, as mentioned above, to raster image data.
• the DMACs 221 and 231 read out data from the three addresses, and transfer them to the rasterization units 222 and 232, respectively.
• rasterization units 222 and 232 require neither a pattern flag nor second, third, and fourth color data, as described above, so only the first color-writing start address is designated on the DMACs 221 and 231. Thus, only the first color data is read from the memory 106 to form an image. With this processing, the memory bus bandwidth can be saved.
• the pattern indicating the arrangement of color data and the pattern flag are associated with each other as shown in Fig. 4 in the first embodiment, the present invention is not limited to this.
• the pattern indicating the arrangement of color data and the pattern flag may be associated with each other and defined in advance such that the pixel value of the lower right pixel in a block of 2 ⁇ 2 pixels represents a first color.
• a size of 2 ⁇ 2 pixels is used as the block size in the first embodiment, the present invention is not limited to this.
• compression has been described by taking R, G, and B components each with 8 bits as an example of image data, data represented by a CMYK color space, grayscale data, or data having pixel values other than 8 bits may be adopted.
• weighting is done based on a predetermined criterion in consideration of the influence of the higher-stage compression (first-stage compression) at this time, as described earlier) . Also, although only two stages of compression are set in the first embodiment, the present invention is not limited to this.
• the user can select an appropriate one of a plurality of resolutions
• the processing speed, the data size, and the image quality can be optimized by switching and optimizing the number of compression stages in accordance with various resolutions.
• the image size differs by four times depending on whether a document created using word processing software is printed as a 600- or 1,200-dpi image.
• both the 600- and 1,200-dpi image data have the same actual contents, so the redundancy is higher in the latter than in the former.
• three-stage compression is performed to compress it into a 300-dpi image, and if a 600-dpi image is input, only two-stage compression is performed.
• 1,200-dpi data normally has a high redundancy, as described earlier, so no large difference occurs between 600- and 1,200-dpi data as long as they are identical PDL data.
• the data size and the performance can be optimized by switching the number of compression stages in accordance with the input resolution such that, for example, only two-stage compression is performed for 600-dpi data and only one- stage compression is performed for 300-dpi scanned data for transmission.
• the pattern flag is defined such that the upper left pixel has the first color, as described with reference to Fig. 4, in this embodiment, the present invention is not limited to this.
• the pattern flag and the placement pattern may be
• a pixel for example, the lower right pixel
• An image expected to be highly compressed by the compression method according to the above- described embodiment has a large number of regions in which the adjacent pixel values in the image represent the same color or have differences between them, which are so small that they can be regarded as representing the same color after color-reduction.
• image regions it is often the case that not only the adjacent pixel values in a single block of 2 x 2 pixels but also those in adjacent blocks each including 2 x 2 pixels represent the same color or have differences between them, which are so small that they can be regarded as representing the same color.
• a flag indicating that the current block is identical to the preceding block is set without outputting the pixel values (color data) of the current block.
• a predetermined threshold is held in advance or designated by the user.
• Fig. 12 is a view showing the number of bits when a flag (to be referred to as a repeat flag hereinafter) indicating that the same pattern as in the preceding block is obtained is added as a new pattern flag.
• a flag to be referred to as a repeat flag hereinafter
• "F" indicating coincidence with the previous block (no pixel values are newly output) is added as a new pattern flag.
• a block having pattern flag F can be represented by 4 bits of the pattern flag, as shown in Fig. 12. That is, because a block having pattern flag F can be rasterized by referring to the pixel values (color data) and pattern flag in the preceding block, no pixel values (color data) need to be stored, thus improving the compression ratio.
• raster image data is input (S1301) .
• a block of 2 * 2 pixels is extracted from the input image data (S1302) . If the extracted block is the first block, no block has been processed in the past (YES in step S1303) , so compression processing is performed (S1307), as in the first embodiment. On the other hand, if the extracted block is not the first block (NO in step S1303) , processing of calculating the difference from the block having been processed immediately before it and stored is performed (S1304). This processing can be done by obtaining, for example, the sum of the absolute values of the differences between the pixel values at
• step S1305 If the difference is smaller than the threshold, it is determined in step S1305 that the same data as in the preceding block stored immediately before the current block can be used (YES in step S1305) . Pattern flag F indicating that the current block is identical to the preceding block is specified (S1306) . In this case, no pixel values (color data) are output. That is, if it is determined in step S1305 that the difference between the block which is the current processing object and that- which has become a processing object before it is small, pattern flag F (repeat flag) indicating repetition of the preceding block is output for the block which is the current processing object.
• step S1305 If it is determined in step S1305 that the threshold is not smaller than the threshold (NO in step S1305) , it is determined that the data on the preceding block cannot be reused due to a large difference from the preceding block, and the same compression
• preceding block (S1309) .
• the data on the preceding block is updated in steps S1308 and S1309 after compression processing is performed in step S1307, the present invention is not limited to this.
• the data on the preceding block may be updated using the pixel values in the block which is the
• a pattern flag and color data obtained as a result of the compression processing in step S1307 are output for a block having a difference from the storage block, which is not smaller than the threshold, and only pattern flag F is output for a block having a difference from the storage block, which is smaller than the threshold (S1310) .
• step S1311 compression processing for the entire image is completed.
• the number of bits of the pattern flag remains 4 bits, thus making it possible to more reliably improve the
• the color processing unit 212 which is shown in Fig. 2 and has been
• the first color is not continuous with the second, third, and fourth colors, so not only the address of the first color but also the second/third/fourth color writing start
• pattern flag data is read, simultaneously with reading of first color data. In the process of reading, if pattern flag F indicating repetition of the
• first color data and pattern flag F need only be read, and none of second, third, and fourth color data are
• coincidence with a wider area can be defined by, for example, adopting coincidence with a block positioned on the upper side of the block of interest or expanding the number of bits of the pattern flag.
• a pattern flag and pixel data are restored to raster image data.
• second/third/fourth color writing start address of the compressed data arranged on a memory 106 as shown in Figs. 11A, 11B, 11C, 11D, and HE are designated on DMACs 221 and 231.
• the output of the preceding block is directly used as that of the block of interest if it is specified that the block of interest is identical to the preceding block by referring to the pattern flag of the block of interest, and rasterization can be done using the above-mentioned method if a flag other than that indicating to that effect is obtained. At this time, one block must be buffered because it may be used for the next block.
• rasterization processing data on a block, which coincides with that on an adjacent block, cannot be rasterized only by reading first color data from the memory 106 to form an image, as described earlier.
• a half-size image is restored based on the first color data while referring to the pattern flag.
• rasterization can be done by repeatedly outputting the preceding color data among the first color data. This makes it possible to save the memory bus bandwidth.
• the redundancy between blocks can be reduced, thus further enhancing the compression ratio. This makes it possible to obtain the same effect as that obtained by executing the color processing which is completed by one pixel input/output, described in the above-described embodiment.
• the random access performance is improved by changing or modifying the compression processing for raster image data in the above-described embodiments.
• an MFP will be taken as an example of an image processing apparatus in the third embodiment, as in the first embodiment.
• the compression processing used is variable-length compression
• a compression method which focuses attention on further improving the random access performance compared to the methods described in the. above-described- embodiments will be described in detail with reference to Fig. 15.
• raster image data is input for each page shown in Fig. 16 (S1501) .
• One page background pixel is set for each page of the image data (S1502) .
• This pixel data is used as the initial storage block in compression processing, and typically adopts white (255 for an image with 8 bits for each of R, G, and B colors, and 0 for a CMYK image) .
• Image data which is an input to a compression unit 202 is divided for each predetermined size (S1503) .
• the predetermined size corresponds to 32 ⁇ 32 pixels.
• tile in order to discriminate it from the block of 2 ⁇ 2 pixels described previously.
• One tile includes 16 ⁇ 16 blocks each including 2 * 2 pixels.
• Fig. 16 is a view showing the relationship among a page, tiles each including 32 ⁇ 32 pixels, and blocks each including 2 ⁇ 2 pixels.
• Header information with a predetermined fixed length is assigned to each tile (S1504).
• the header describes pieces of information such as the page ID, the tile coordinate position, the color space, the number of bits of pixel data, the data size of a tile, the presence/absence of attribute information, and the compression flag.
• the page ID describes an ID number which is uniquely assigned to each page.
• the tile coordinate position is coordinate position information indicating the position of the tile on a raster image for each page. In this case, the tile coordinate position is represented by a two-dimensional coordinate position defined by X- and Y-coordinates .
• the color space is information indicating an identifier used to determine whether the tile is an RGB image, a CMYK image, or a grayscale image.
• the number of bits of pixel data is information indicating the bit length per pixel in the tile.
• the data size is information indicating the sizes of first color data and second, third, and fourth color data of the tile in units of bytes.
• the compression flag is
• the initial storage block is defined as a storage block, and a block filled with the
• background pixels is used as this storage block.
• pixel data of the first block of 2 ⁇ 2 pixels need not be output for every tile, thereby making it possible to reduce the corresponding amount of data. This is because a normal document image has white, so the amount of data is most likely to be reduced when background pixels are set to have white, as described above .
• the size of the tile data having undergone the compression processing is calculated.
• calculated size is the sum total of the pattern flag size, the first color data size, and the second, third, and fourth color data size. Because pattern flags are always assigned to the tile data, there is no guarantee that the tile data is compressed more highly than the original image data. Hence, if the size of the tile data exceeds that of the original tile data, it may be as well to output the original image data in terms of achieving a higher compression ratio in total.
• the data size after compression and the original data size are compared with each other (S1506) . If the data size after compression exceeds the original data size (YES in step S1506) , the compression flag value of the header remains the same. On the other hand, if the data size after compression does not exceed the original data size (NO in step S1506) , the flag value is incremented by one (S1508).
• step S1508 the compressed tile data is packed into one data, together with the header information of the tile
• step S1509 if it is determined as a result of the comparison that the data size after compression exceeds the original data size (YES in step S1506) , the original tile data is packed into one data, together with the header information of the tile
• Fig. 17 shows the structure of the packed data.
• the unit of data including the above-mentioned header will be referred to as a "packet" hereinafter.
• Packet To generate such a packet, after the data size is determined upon completion of compression processing for each tile, data is packed while the distance between the first color storage unit and the
• step S1507 or S1509 the data is output onto the memory via the DMAC (S1510) .
• the coordinate position and size of this packet are listed to generate a packet management table (S1511) .
• Fig. 18 illustrates an example of this packet management table.
• Data on the third packet can be obtained by reading the data at the third packet address .
• an image can be partially processed. For example, to extract and process a partial region in an image, data on a packet
• Multistage compression can also be
• tile data of 32 x 32 pixels is input (S2001) .
• the background color is set (S2002), as described earlier.
• First color data is extracted from the packet (S2003) .
• the initial first-stage compression is performed for all of 32 ⁇ 32 pixels.
• Compression processing is performed for the first color image, as described earlier.
• compressed data is compared with that of the original data. If it is determined as a result of the
• step S2005 the compression flag in the packet header remains the same, and raw data (the intact first color data in this case) is packed into a packet (S2006) .
• the packet is output (S2010) .
• step S2007 the compressed data, that is, the first color, the pattern flag, and the second, third, and fourth color data are packed into a packet (S2008).
• S2008 the amounts of first color data and second, third, and fourth color data which are generated upon compression are described in the header for each stage level. If the amount of first color data is equal to or larger than 2 x 2 pixels (YES in step S2009) , the process returns to step S2003, in which compression processing is performed for the first color data again.
• step S2010 If the amount of first color data is smaller than 2 x 2 pixels (NO in step S2010), that is, if only one pixel remains as data to be processed, data to be compressed no longer remains, so the packet is output (S2010) .
• the coordinate position and size of the packet are recorded in the packet management table (S2011), and the
• rasterization processing is performed using information described in each header because the header is assigned to each packet. First, if the compression flag indicates non-compression, data other than the header is output; otherwise, rasterization processing is performed. In the rasterization
• the pattern flag storage position, the first color data storage position, and the second, third, and fourth color data storage position are obtained based on the header, and the data is
• the position of the pattern flag can be obtained by offsetting because it is that of a header with a fixed length. If first color data has a tile size of 32 ⁇ 32 pixels, its pattern flag has a fixed size of 32 ⁇ 32 bits, so the first color data is obtained by multiplication by an offset from the pattern flag position. Lastly, because the first color data has an indeterminate length, second, third, and fourth color data are obtained by multiplication from an offset from the first color data position by
• Rasterization processing is performed in multiple stages.
• aspects of the present invention can also be realized by a computer of a system or apparatus (or devices such as a CPU or MPU) that reads out and executes a program recorded on a memory device to perform the functions of the above-described embodiment (s) , and by a method, the steps of which are performed by a computer of a system or apparatus by, for example, reading out and executing a program recorded on a memory device to perform the functions of the above-described embodiment ( s ) .
• the program is provided to the computer for example via a network or from a recording medium of various types serving as the memory device (for example, computer- readable medium) .

Landscapes

• Engineering & Computer Science (AREA)
• Multimedia (AREA)
• Signal Processing (AREA)
• Compression Of Band Width Or Redundancy In Fax (AREA)
• Image Processing (AREA)
• Facsimile Image Signal Circuits (AREA)
• Color Image Communication Systems (AREA)
• Compression Or Coding Systems Of Tv Signals (AREA)

Priority Applications (1)

Applications Claiming Priority (2)

Publications (1)

Family

ID=44762393

Family Applications (1)

Country Status (3)

Cited By (1)

Families Citing this family (17)

Citations (2)

Family Cites Families (19)

• 2010
  • 2010-04-08 JP JP2010089855A patent/JP5393574B2/ja not_active Expired - Fee Related
• 2011
  • 2011-03-09 WO PCT/JP2011/056133 patent/WO2011125433A1/en not_active Ceased
  • 2011-03-09 US US13/255,238 patent/US8553301B2/en not_active Expired - Fee Related

Patent Citations (2)

Also Published As

Similar Documents

Legal Events