JP4719924B2 - Image processing apparatus and image processing method - Google Patents

Description

  The present invention relates to an image processing apparatus and an image processing method for printing an image in monochrome or color, such as a printer, a digital multi-function peripheral, and a color facsimile apparatus, and in particular, processing of image data in which the density of each pixel is expressed in multiple values. The present invention relates to an image processing apparatus and an image processing method.

  With the development of printing and printing technology, high-quality printing can be easily performed even in the field of printing. Along with this, the number of constituent bits of image data constituting each pixel tends to increase. These image data are usually stored once in the internal memory of the image processing apparatus and then used for printing.

  For example, taking a copying machine as an example, image data read by the scanner is once stored in the internal memory of the copying machine main body. Then, printing is performed using the image data read from the internal memory. In the case of a digital multi-function peripheral, image data externally input from a personal computer or the like can be read into an internal memory instead of reading a document with a scanner. Then, an image is printed using the image data read from the internal memory.

  In such an image processing apparatus, when printing is performed at a high speed, recording is performed on both sides of a sheet, or various image editing is possible, the apparatus accumulates a number of pieces of image data in an internal memory. Often there is a need. Moreover, as described above, the amount of image data constituting one image has become enormous with the improvement in image quality. For this reason, if image data is stored in the internal memory without performing any reduction processing, the price of the apparatus must be significantly increased in order to secure the necessary memory capacity. Therefore, in general, in such a pixel processing apparatus, image data is stored in an internal memory using an image compression technique, and the compressed image data is expanded at the time of printing.

  There are various compression techniques for image data to be stored in the internal memory. For example, standardized encoding techniques such as the MMR (Modified Modified Read) method for binary image data and the JBIG (Joint Bi-level Image experts Group) method exist. There are also various standardized or independently developed multilevel image encoding techniques such as the JPEG (Joint Photographic Experts Group) method for multilevel image data such as color images and gray images.

  The latter multi-valued image data has a much larger amount of image data before compression than binary image data. For this reason, even if it compresses using various encoding systems, such as a JPEG system, the reduction effect of the data amount by compression cannot be expected so much. In general, image quality and compression efficiency are in a trade-off relationship, and if the image quality is maintained at a high quality, the compression performance is limited, and the image data cannot be largely compressed. In particular, image data captured from a personal computer often requires high-quality image quality, such as a portrait photograph taken with a digital camera or a landscape image acquired from a website. Thus, an image processing apparatus that handles multi-value image data such as a color image inevitably requires a large capacity memory.

  In view of this, it has been proposed to divide a multi-valued image into a plurality of weighted bit planes, each of which is represented as a binary image, and to obtain a correlation between them to improve the compression rate of the image data (for example, patents). Reference 1).

  FIG. 12 shows a main part of the configuration of the image processing apparatus according to the first proposal. The image processing apparatus 500 includes a bit plane separator 502 that receives multi-value image data 501 and separates it into a plurality of bit planes. The bit plane separator 502 uses the separation table 503 to separate the multivalued image data 501 into a plurality of bit planes 504. The separated bit plane 504 is input to the pre-compression processor 505, and a composite plane 506 is created as a single bit plane in which bit data of the same position constituting the bit plane is arranged in the vicinity. The synthesis plane 506 is input to the data compressor 507 and image compression is performed.

  FIG. 13 shows a separation table of the bit plane separator according to the first proposal. In this example, multivalued image data 501 is separated into first to third bit plane pixel values. The first to third bit plane pixel values are equal to those obtained by separating the values from “0” to “7” expressed in decimal numbers into each bit (each digit) constituting the binary number.

  FIG. 14 shows the processing contents of the pre-compression processor in principle. The first bit plane is a bit plane composed of a bit string of “0” or “1”, which is arranged for each line in order from the first line by scanning an image composed of the first bit plane pixel values shown in FIG. It is. Similarly, the second bit plane is a bit composed of “0” or “1” which is arranged for each line in order from the first line by scanning the image formed by the pixel values of the second bit plane shown in FIG. The third bit plane is composed of “0” or “1” that is arranged for each line in order from the first line by scanning the image formed by the third bit plane pixel values shown in FIG. It is a bit plane.

  The pre-compression processor 505 creates one composite plane 506 from these first to third bit planes. At this time, first, the bit string of the first line of the first bit plane is arranged in a line as a first line, and the bit string of the first line of the second bit plane is arranged in a line as a second line below. . Below that, the bit line of the first line of the third bit plane is arranged in a line as the third line. In the fourth line to the sixth line, the bit strings of the second lines of the first to third bit planes are arranged one line at a time. The same applies hereinafter.

  The data compressor 507 compresses the single composite plane 506 created in this way using an existing compression technique. At this time, by using the correlation information of the lines adjacent to each other, the compression efficiency of the image data is increased.

  On the other hand, in connection with the present invention, there is also a technique of performing processing for reducing gradation bits of multi-valued image data before compressing a bit plane (see, for example, Patent Document 2). In the second proposal, the first multi-gradation image can be displayed by the dither method, and the first conversion means for converting the original multi-gradation image into simple multi-gradation number data that can be displayed by truncating the lower bits. A second conversion means for converting into dither multi-gradation number data. The gradation number data converted by the first and second conversion means is converted into gray code data by the third conversion means, and this gray code data is converted into the highest bit plane of the simple multi-gradation data. It is decided to sequentially encode from bit plane to lowest bit plane in units of bit planes.

FIG. 15 shows a table for converting the number of gradations according to the second proposal into gray code data. This gray code data is a code in which the Hamming distance between adjacent codes is “1”. In an image where the density changes slowly like a photograph, the correlation between adjacent pixels in the bit plane using the gray code is strong.
JP 2004-312773 A (paragraphs 0010, 0016 to 0018, FIGS. 5 and 8) JP-A-08-009369 (paragraphs 0004 and 0006, FIGS. 1 and 4)

  Multi-value image data has a larger amount of image data before compression than binary image data. Even when compression is performed using various encoding methods such as the JPEG method, there is a trade-off between image quality and compression efficiency. When image quality is given priority, compression performance is greatly limited, and image data compression is performed. The efficiency cannot be increased.

  For this reason, first and second proposals have been made. However, in the first proposal, the bit plane separator 502 converts multi-value image data into binary data, and performs pre-processing for converting each digit value into a bit plane. Therefore, even if bit lines of the same line after preprocessing are made adjacent to each other, they are originally uncorrelated because of bits at different positions of binary numbers (digits). Therefore, the magnitude of the compression effect of the synthesis plane 506 by the data compressor 507 cannot be expected.

In the second proposal, the number of bits is reduced by truncating the original multi-gradation image, or the gradation number data converted to dither multi-gradation number data by the dither method is converted to gray code data. Yes. Since the image data is compressed while maintaining the gradation as described above, only the compression effect within the frame of the existing compression technique can be expected.

  SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide an image processing apparatus and an image processing method that improve the compression efficiency of multi-value image data for printing by creating intermediate data effective for compression.

  In the present invention, (a) preprocessing means for inputting image data representing the density of each pixel in multi-values and reducing the number of constituent bits representing the density in pixel units with a predetermined logic to a predetermined value of 2 or more; (B) bit string conversion means for converting a bit string representing the density of each pixel whose number of constituent bits has been reduced by the pre-processing means into a bit string previously associated with the appearance frequency of each density; and (c) this bit string. Image processing includes: a bit plane expansion unit that expands the bit string after being converted by the conversion unit into a bit plane for each bit; and (d) an image compression unit that compresses the bit plane expanded by the bit plane expansion unit. Install in the device.

  That is, in the present invention, preprocessing is performed to reduce the number of constituent bits of image data in which the density of each pixel is expressed in multiple values to a predetermined value of 2 or more, and the bias of data representing the density in the multivalued data after this preprocessing is performed. Is used to convert the bit string representing each density into a bit string associated with the appearance frequency of these densities. Then, encoding is performed after developing the bit sequence for each bit constituting the bit string subjected to such conversion processing. As a result, the compression efficiency can be increased by converting the bit string in consideration of the compression rate when expanding to the bit plane.

  In the present invention, (a) a preprocessing step of inputting image data representing the density of each pixel in multi-values and reducing the number of constituent bits representing the density in pixel units with a predetermined logic to a predetermined value of 2 or more. (B) a bit string conversion step for converting a bit string representing the density of each pixel whose number of constituent bits has been reduced in the preprocessing step into a bit string previously associated with the appearance frequency of each density; A bit plane expansion step of expanding the bit string converted by the bit string conversion step into a bit plane for each bit; (d) an image compression step of compressing the bit plane expanded in the bit plane expansion step; E) a storage step for storing the data compressed in this image compression step in the data storage memory; and (f) storage in this storage step. An image decompression step for reading out and decompressing the data, and (g) inputting a bit plane obtained by decompression in the image decompression step, performing reverse conversion to the bit string conversion step, And a printing data output step to be included in the image processing method.

  That is, in the present invention, preprocessing is performed to reduce the number of constituent bits of image data in which the density of each pixel is expressed in multiple values to a predetermined value of 2 or more, and the bias of data representing the density in the multivalued data after this preprocessing is performed. Is used to convert a bit string representing each density after pre-processing into a bit string associated with the frequency of appearance of density. Then, the encoded data is developed after being developed into bit planes for each bit constituting the bit string subjected to such conversion processing, so that the compression rate is increased and the encoded data is stored in the data storage memory. . When printing, the data stored in the data storage memory is read and decompressed, and the resulting bit plane is converted in reverse to the bit string conversion step, and the resulting multivalued image data is printed. Output as multi-value data.

  As described above, according to the present invention, the number of constituent bits of multi-value data representing image density is reduced to a predetermined value of 2 or more by preprocessing, and the deviation of the number of pixels for each gradation at this time is used. Therefore, the bit string is converted so as to be effective for compression, and is expanded into a bit plane and encoded. Thereby, multi-valued image data can be efficiently compressed by binary coding, and the memory amount, circuit scale, etc. can be reduced. In addition, the processing speed of image data can be improved. Also, at the time of decompression, multi-value data can be obtained quickly and easily by performing inverse conversion corresponding to bit allocation conversion at the time of compression after binary decoding, and for example, printing processing can be performed in real time.

  Hereinafter, the present invention will be described in detail with reference to examples.

  FIG. 1 shows the configuration of an image compression / decompression apparatus according to an embodiment of the present invention. The image compression / decompression apparatus 100 according to the present exemplary embodiment includes a screen processing unit 102 that inputs multi-value image data 101. The multivalued image data 101 may be output data obtained by reading a document (not shown) by a scanner (not shown), or may be output data of a computer (not shown). When the image compression / decompression apparatus 100 includes a connector connected to various external memories, output data of an image selected from the memory attached to the connector may be used.

  The screen processing unit 102 performs screen processing as preprocessing for reducing the constituent bits of the input multi-value image data 101. A screen pattern memory 103 is connected to the screen processing unit 102, and a screen pattern 104 necessary for screen processing is read out. Details of the screen processing will be described later.

  The multi-value data 105 converted by the screen processing unit 104 is sent to the table conversion processing unit 106. A first table storage memory 107 is connected to the table conversion processing unit 106, and each value of the multi-value data 105 is converted using the first table data 108. Details of this conversion will be described later.

  The multi-value data 109 after the table conversion processing output from the table conversion processing unit 106 is input to the bit plane conversion unit 111 and developed into one bit plane. The expanded bit plane 112 is supplied to the binary compression unit 113. The binary compression unit 113 compresses the bit plane 112 and stores the compressed code data 114 in the data storage memory 115.

  The image compression / decompression apparatus 100 according to the present embodiment constitutes a main part of a digital multi-function peripheral that also functions as a copier, a scanner, or a printer, for example. For this reason, the data storage memory 115 holds various image data after compression in a compressed state until later use.

  The data storage memory 115 is connected to a binary decompression unit 116 in addition to the binary compression unit 113. The binary decompression unit 116 decompresses the code data 117 read from the data storage memory 115 and restores it to bit plane data. The restored bit plane data 118 is sent to the multi-value data restoration unit 119 and restored to multi-value data. The multi-value data 121 output from the multi-value data restoration unit 119 is sent to the table inverse conversion processing unit 122. A second table storage memory 123 is connected to the table inverse conversion processing unit 122. The second table storage memory 123 stores a conversion table for performing a completely opposite conversion to the first table storage memory 107. The table reverse conversion processing unit 122 uses the second table data 124 to perform table conversion opposite to the processing performed by the table conversion processing unit 106. Then, the multi-value data 125 after the original screen is restored and output. The multi-value data 125 is sent to, for example, a printer (not shown) and printed.

  Next, the state of the image compression / decompression process of the image compression / decompression apparatus 100 according to the present embodiment will be described in detail. The multivalued image data 101 input to the screen processing unit 102 is a monochrome gray multivalued image or color image data. In order to input such multi-valued image data 101 to the image compression / decompression apparatus 100 of this embodiment, various methods are possible. For example, multi-value image data 101 generated by a personal computer (not shown) can be input using a communication module such as a LAN (Local Area Network) or a facsimile. At this time, image information obtained in a state of being printed on the recording paper may be read by a scanner (not shown).

  In addition to the LAN, it is also possible to input to the image compression / decompression apparatus 100 from a personal computer through a standard interface such as USB (Universal Serial Bus) or Centronics as a parallel port for a printer or a dedicated interface. . Further, there is a method of inputting data from the personal computer to the image compression / decompression apparatus 100 using a PCI (Peripheral Component Interconnect) bus or other unique bus. Furthermore, a printer description language sent from a personal computer is interpreted and decompressed into a bitmap image by a device such as a CPU (Central Processing Unit) on the terminal side as multi-valued image data 101 to compress and decompress the image. It may be input to the apparatus 100.

  The type of multi-value image data 101 input to the image compression / decompression apparatus 100 according to the present embodiment is not particularly limited. However, it is assumed that the multi-value image data 101 used in this embodiment is color space data for printing. That is, the monochrome gray multi-value image data needs to be density data, not lightness data, and in the case of color image data, it needs to be data in the CMYK color space. Therefore, for example, when using image data expressed in a YUV color space for digital video, a CPU or other conversion means (not shown) provided on the personal computer side or the terminal side is used to change to the CMYK color space. It is necessary to convert colors. Of course, it is free to provide a color space conversion unit that converts the color space of the multi-valued image data 101 as necessary before the screen processing unit 102 in the image compression / decompression apparatus 100 of the present embodiment.

  The screen processing unit 102 uses the screen pattern 104 stored in the screen pattern memory 103 for the input multi-value image data 101, and calculates the number of constituent bits for each pixel of the image data representing the multi-value data of density. Perform pre-processing to reduce. In this embodiment, this is called screen processing. When CMYK color image data is input as the multi-value image data 101, the screen processing unit 102 uses the screen patterns 104 corresponding to the C, M, Y, and K colors stored in the screen pattern memory 103, respectively. Screen processing is performed for each color. In this embodiment, a case will be described in which 256 gray-scale monochrome gray multilevel image data is input as the multilevel image data 101.

FIG. 2 is a diagram for explaining the principle operation of the screen processing unit of this embodiment. The multivalued image data 101 represents, for example, the density of each pixel in 256 levels (8 bits). This multi-valued image data 101 uses screen patterns 131 each having a plurality of pixel lengths in the x direction (main scanning direction) and the y direction (sub scanning direction) stored in the screen pattern memory 103 shown in FIG. Each pixel is converted into multilevel data 105 of 8 levels (3 bits). Here, the screen pattern 131 is set with threshold values TH ₁₁ , TH ₁₂ ,... TH ₄₄ for performing multi-level conversion for each pixel P _{11 to} P ₄₄ . It is configured as a dither pattern.

  The screen processing may use a general screen processing technique that reduces the number of constituent bits for each pixel. In the pseudo halftone processing that attempts to express the gradation of an image with multiple values in this way, when looking at the density of each pixel microscopically, the density itself is directly different from the image data 101 before the screen processing. It is not the data that expresses it automatically. For example, in the multi-value dither processing, the overall density of the 256 gradation 4 × 4 pixel frame portion of the range constituting the screen pattern 131 shown in FIG. 2 is set to the same range of 8 gradation 4 × 4 pixel. This is because the set is expressed as accurately as possible.

  FIG. 3 shows a general state in which image data of a certain image area having image data proportional to the gradation density for each pixel is developed on a bit plane. Here, it is assumed that each pixel of the original image area 141 represented by 256 gradations is converted to 8 gradations (3 bits) according to the density. The least significant bit after conversion is the first bit, and the most significant bit is the third bit.

  In the original image area 141 represented by 256 gradations, it is assumed that the letters “A” are drawn in a relatively dark color on a white background. In the case of an image such as a light and dark comparatively clear sentence, the density is determined by deleting the lower bits of the 8 bits without using a method of faithfully reproducing the halftone as in the previous dither process. Is simply changed from 256 gradation expression to 8 gradation expression, the bit planes 143 to 145 from the first bit to the third bit have substantially the same image pattern.

  FIG. 4 shows how the image data of the image area described above is developed on the bit plane when the gradation bits are reduced. Assume that preprocessing for converting the image data of the original image area 141 expressed in 256 gradations into 8 gradations is performed by the screen processing unit 102 (FIG. 1) of this embodiment. Then, for example, in the character portion (dark portion) “a” in dither processing, the portion of the threshold value that has various values for each pixel is a pixel on the darker side than these threshold values. Is determined. Further, in the ground color portion close to the white background, it is determined that the pixel is brighter than these threshold values in the portion of most of these threshold values. As a result, the bit plane 151 relating to the first bit of the image data represented by 8 gradations has almost the same image pattern as the original image area 141 represented by 256 gradations.

  On the other hand, the second bit and the third bit of the image data represented by 8 gradations are pixel portions constituting a part of the character because the threshold value takes various values for each pixel. However, depending on the threshold value, it is not determined that the pixel is on the dark side. As a result, although depending on the density of the character part “A”, the character pattern does not appear on the bit plane 152 of the second bit or the bit plane 153 of the third bit, and instead of the bits “0” and “1” A mixed pattern will appear.

  Of course, such a pattern in which the bits of “0” and “1” are mixed is not preferable for efficient compression. Therefore, in the present embodiment, after the screen processing is performed by the screen processing unit 102 shown in FIG. 1, the noise that is scattered in this way in the table conversion processing unit 106 before the bit plane conversion unit 111 develops each bit plane. The occurrence of this is minimized. The first table storage memory 107 stores first table data 108 for converting the multi-value data 105 after the screen processing.

  FIG. 5 shows the relationship between an example of the first table data and the second table data. The first table storage memory 107 shown in FIG. 1 stores several types of first table data corresponding to the multi-value image data 101 input to the image compression / decompression apparatus 100. In FIG. 5, one of the items selected by the user is shown as first table data 108. The second table data 124 is paired with the first table data 108. That is, the first table data 108 is conversion data for performing conversion processing of the multi-value data 105 developed on each bit plane by the bit plane conversion unit 111 so that the compression is optimally performed by the binary compression unit 113. The table data 124 of 2 is conversion data for restoring the multi-value data 125 after the original screen processing.

  Next, how the converted data is created will be described. This is also a problem of the user selecting which one of the first table data 108 is prepared in the first table storage memory 107 as in this embodiment.

  FIG. 6 shows a histogram of the multi-value image data representing the character “A” shown in FIG. 4 in the present embodiment. In this figure, the horizontal axis represents the image density in 256 levels from “0” to “255”, and the vertical axis represents the number of corresponding pixels. Since the black-colored character “A” is drawn in the white-colored ground color portion, there are peaks in the number of pixels at the two densities “0” and “255”. Further, the values are dispersed to some extent in other gradations, and the number of pixels of about 10% of the number of pixels having the highest density “255” exists in some other gradations.

  FIG. 7 shows a histogram of multivalued data after the gradation bits of the image data are reduced by screen processing. The horizontal axis represents eight levels of gradation from density “0” to density “7”, and the vertical axis represents the number of pixels of each density. In this case, most of the pixels are concentrated on the gradations of density “0” and density “7”. The number of pixels of other gradations appears only about 0.5% with respect to the number of pixels having the highest density “7”. Therefore, multivalued data such as “000 (binary number)” or “001 (binary number)” is expressed as “111 (binary number)” as the multivalued data with the density “7” having the highest appearance rate. By expressing as, it is possible to effectively compress each bit with a bias when it is developed on the bit plane.

  As can be seen by comparing FIG. 6 and FIG. 7, the multi-value data 105 (FIG. 6) after the screen processing is compared with the multi-value image data 101 (FIG. 6) before the screen processing input to the image compression / decompression apparatus 100. The appearance rate is concentrated in some of all gradations. As a result, for example, the density “6” in FIG. 7 has the lowest appearance rate, and the corresponding pixel is zero. By preferentially arranging a bit string having a high compression rate such as “111 (binary number)” at such a low appearance rate density, for example, “1” in the second bit 152 or the third bit 153 shown in FIG. The number of occurrences of "" can be reduced. In this example, three bit planes are assigned by assigning a bit string that is more advantageous for compression as the ratio of appearance of each density (gradation) in the multi-valued image data after the screen processing shown in FIG. 7 is higher. The pattern after conversion into can be effectively compressed.

  The first table data 108 shown in FIG. 5 has an improved compression rate with respect to a character mode such as when characters are printed on paper. A plurality of modes are set according to the type of image data as in the photo mode mainly for halftones, and the first table data 108 is prepared for each of these modes, and the user selects as appropriate. Making it possible is effective in improving the compression ratio.

  FIG. 8 shows three bit planes as a result of table conversion of image data representing the character “A”. Compared with the bit planes 152 and 153 relating to the second and third bits shown in FIG. 4, the number of “1” bits appearing as noise is reduced, and the compression efficiency is improved.

  By the way, the bit plane conversion unit 111 expands the three bit planes 161 to 163 per image obtained in this way into one bit plane and compresses it by the binary compression unit 113 shown in FIG. I am going to do it.

  FIG. 9 shows a method of developing three bit planes into one bit plane. In this embodiment, as shown in this figure, the bit planes 161 to 163 for X × Y pixels constituting the multi-value data 109 are arranged side by side as a bit plane 171 for 3X × Y pixels, and this is binary compressed. The unit 113 compresses the data. The arrangement direction is not limited to the same direction as this, and may be arranged to be, for example, X × 3Y.

  As a result, the binary compression unit 113 uses only a couple of line memories each having a length corresponding to 3X pixels or Y pixels to perform compression using the continuous state of the same pixels in one line, or the same between adjacent lines. Compression using the difference state is possible. These are, for example, standard binary entropy coding such as MH (Modified Huffman), MR (modified READ: relative element address designate), MMR, and JBIG as binary coding methods in facsimile communication, or original entropy coding. in use.

  By the way, the three bit planes 161 to 163 can be separately compressed and stored without being combined into one bit plane 171 as shown in FIG. However, if it does in this way, it is necessary to unify and manage these three bit planes 161-163 as one set of data, and the data for that will become unnecessary. Further, since the three bit planes 161 to 163 are to be processed as a unit, there is no merit of managing them individually.

  Further, the three bit planes 161 to 163 can be arranged in the vertical direction (vertical direction) without arranging these planes into one horizontal plane as shown in FIG. However, in this case, sequential processing for each of the bit planes 161 to 163 is necessary, and real-time compression processing cannot be performed. Further, when arranged in the vertical direction, the bit planes 161 to 163 have little correlation, so the merit of compression is not great.

  As shown in FIG. 9, by combining the data into one bit plane 171, the memory required for image processing is only required for at least one line. Further, for example, the number of pixels in the main scanning direction is set for each bit plane during compression. By setting it to 3 times 161 to 163, real-time compression processing becomes possible.

  The binary compression unit 113 of the present embodiment can use a normal binary encoding circuit, has a small circuit scale, and can be configured at low cost. Also, when the binary compression unit 113 is configured by software, the processing is relatively light, so that high-speed processing is possible. In this embodiment, the binary compression unit 113 is operated at a processing speed three times as high as the circuit processing up to the bit plane conversion unit 111, so that the process until the compressed code data 114 is written to the data storage memory 115 is performed. It is possible to carry out in real time without delay.

  The code data 114 stored in the data storage memory 115 in this way is read when a printer (not shown) needs to print. The same applies to the case where the print data is transferred via a network (not shown) such as a LAN (Local Area Network) or a communication cable. The code data 117 read from the data storage memory 115 is decompressed by the binary decompression unit 116 into multi-value bit plane data 121. Naturally, the binary decoding method at the time of expansion is a decoding method corresponding to the binary compression unit 113. The expanded bit plane data 118 is returned to the multi-value data 121 by the multi-value data restoration unit 119. This process is the reverse of the process in the bit plane conversion unit 111.

  The first-stage multi-value data restored by the multi-value data restoration unit 119 corresponds to the bit plane 171 of 3X × Y pixels shown in FIG. Therefore, the multi-value data 121 is obtained by separating the data into three pieces of bit plane data corresponding to X × Y pixels. The batch of multi-value data 121 is input to the table inverse conversion processing unit 122. The table inverse conversion processing unit 122 performs table conversion with reference to the second table data 124 sent from the second table storage memory 123.

  As described with reference to FIG. 5, the first table data 108 and the second table data 124 have a pair relationship. Therefore, when a plurality of types of first table data 108 are prepared in the first table storage memory 107, the same number of second table data 124 is prepared in the second table storage memory 123. ing. For example, when the character mode data is read as the first table data 108, the character mode data is similarly read as the second table data 124.

  The table reverse conversion processing unit 122 restores the original multi-value data 125 after the screen processing. Here, the reverse conversion of the table conversion processing unit 106 is performed. FIG. 5 shows the state of this conversion. The multi-value data 125 obtained by performing the table reverse conversion process in the table reverse conversion processing unit 122 is output as data after the screen processing. The output data is printed by a printer (not shown) through post-processing for multi-value conversion such as PWM (Pulse Width Modulation) if necessary. As described above, since the image compression / decompression apparatus 100 according to the present embodiment performs image processing when printing an image such as screen processing, it is assumed that data output from the apparatus is ultimately intended for printing. It is said.

  In the embodiment, the case where monochrome gray multilevel image data is printed has been described as an example. The present invention can be similarly applied to the case of printing color image data or multi-color multi-value image data.

  As described above, according to the image compression / decompression apparatus 100 of the present embodiment, an image is obtained by performing conversion for changing the bit allocation using the bias of the value for the multivalued data after the screen processing. Improves data compression efficiency. In addition, by arranging each bit plane of multilevel data in the horizontal direction to form one bit plane, it is possible to operate a small amount of memory of about one line and a simple binary encoding circuit at high speed. .

  As a result, multi-valued image data such as a monochrome gray image and a color image can be efficiently compressed and decompressed with high compression performance in real time using a relatively simple binary coding method. In the present embodiment, since the screen processing is performed on the multi-value image data, the amount of data before compression can be reduced, and a large amount of multi-value image data is accumulated with a relatively small memory capacity. be able to. Therefore, the cost of the image compression / decompression apparatus 100 can be reduced. In addition, when the compressed image data is expanded, the reverse process of the compression process is performed, so that the image can be restored and output without delay even with respect to the speed of the printer engine.

  <Deformability of invention>

  FIG. 10 shows the configuration of an image compression / decompression apparatus according to a modification of the present invention. In the image compression / decompression apparatus 100A of this modification, the same parts as those of the image compression / decompression apparatus 100 shown in FIG. 1 are denoted by the same reference numerals, and description thereof will be omitted as appropriate.

  In the image compression / decompression apparatus 100A according to this modification, the data analysis unit 201 analyzes the characteristics of the multi-value image data 101 input thereto based on the multi-value data 105 output from the screen processing unit 102, and One table data 108A and second table data 124A corresponding thereto are created. Of these, the first table data 108A is stored in the first table storage memory 107A.

  The second table data 124A is supplied to the data storage memory 115 and stored in correspondence with the code data 114A compressed by the binary compression unit 113. This is because when the plurality of images are stored as the code data 114A at different times in the data storage memory 115, the second table data 124A differs depending on which image of the code data 117A is read. This is because the corresponding second table data 124A is read out.

  By the way, in the image compression / decompression apparatus 100A of this modification, the data analysis unit 201 performs analysis based on the multi-value data 105 output from the screen processing unit 102. Therefore, a temporary memory 202 is prepared for adjusting the time delay required for this analysis and the time timing of the multi-value data 105A after screen processing supplied to the table conversion processing unit 106A.

  The table conversion processing unit 106A converts the multi-value data 105A after the screen processing using the first table data 108A corresponding to the image data supplied from the first table storage memory 107A, and converts the multi-value data 105A. Data 109A is output. The multi-value data 109A is converted into a bit plane by the bit plane conversion unit 111 as described in the previous embodiment, and the developed bit plane 112A is compressed by the binary compression unit 113 to be encoded data 114A. Is stored in the data storage memory 115A.

  When the code data 117A of the image to be printed is read from the data storage memory 115A, the second table data 124A is supplied to the second table storage memory 123A at this timing. Therefore, the table reverse conversion processing unit 122A uses the second table data 124A output at this time to perform table conversion opposite to the processing performed by the table conversion processing unit 106A. Then, the original multi-value data 125A after the screen processing is restored and output. The multi-value data 125A is sent to, for example, a printer (not shown) and printed.

  FIG. 11 shows an outline of the analysis processing of the data analysis unit. The image compression / decompression apparatus 100A of this modification shown in FIG. 10 includes a CPU (Central Processing Unit) and a storage medium that stores a control program executed by the CPU (not shown). The CPU executes a control program so as to perform predetermined control such as analysis control of the data analysis unit 201. Control of the data analysis unit 201 when the multi-value image data 101 is newly input to the image compression / decompression apparatus 100A will be described with reference to FIG.

  When the multi-value data 105 after screen processing is input for one page or a predetermined amount of data (for example, a band unit such as 256 lines or 1024 lines or a half page), the data analysis unit 201 appears for each gradation. Frequency statistics are taken (step S301). For example, when screen processing is performed in which 256 levels of gradation are reduced to 8 levels, the frequency of appearance is counted for each of the 8 levels of gradation as described with reference to FIG. . While this processing is being performed, the multi-value data 105 after the screen processing is sequentially stored in the temporary memory 202 until processing in step S304 described later is performed.

  When the statistical processing in step S301 is completed, the data analysis unit 201 allocates bit strings that can be suppressed from appearing in each bit plane in order from the most frequently occurring gradation, and associates them with a conversion table (first table data). 108A) and a conversion table (second table data 124A) used for restoration is created (step S302). These are output as first table data 108A and second table data 124A (step S303). The first table data 108A is sent to the first table storage memory 107A. The second table data 124A is sent to the data storage memory 115A.

  When the above processing is completed, the data analysis unit 201 instructs the temporary memory 202 to output the multilevel data 105A in order from the multilevel data 105A temporarily stored in the table conversion processing unit 106A (step S304). As a result, the table conversion processing unit 106A starts table conversion using the first table data 108A corresponding to the input multi-value data 105A.

  The table reverse conversion processing unit 122A performs reverse conversion on the multi-value data 121A sent from the multi-value data restoration unit 119 using the second table data 124A, but detailed description thereof will be omitted.

  Regarding the creation of the first table data 108A in the data analysis unit 201, for example, when multi-value data is 3 bits, the most frequently occurring data is “000 (binary number)”, and the next most frequently appearing frequency. A conversion table such as “001 (binary number)” for the next one, “010 (binary number)” for the next highest one, and “000 (binary number)” for the one with the lowest appearance frequency is created. Here, the number of bits changes according to the number of bits after the screen. In the conversion table, the conversion from “0” to “000 (binary number)” is fixed, and the appearance frequency of other data is fixed. Accordingly, rules such as “001 (binary number)”, “010 (binary number)”, “100 (binary number)”,..., “111 (binary number)” may be used.

  According to the image compression / decompression apparatus 100A of such a modification, the conversion table can be dynamically generated by analyzing the characteristics of the input image in accordance with the input multi-valued image data 101. For this reason, it is possible to flexibly cope with any image type input, and to improve the compression performance of image data stably without relying on the user's mode selection skill as compared with the previous embodiment. be able to. Instead, since an image type analysis process is included, a temporary storage memory is required during compression. The amount of memory depends on the range of data to be analyzed. Further, since only the analyzed result is used at the time of expansion, it can be processed in real time as in the previous embodiment.

  In this modification, the first table data 108A and the second table data 124A for the conversion table generated by the data analysis unit 201 are managed separately from the code data and stored in the memory. The management method is not limited to this. For example, it is possible to insert the code data itself as metadata. As an example, an inverse conversion table may be arranged at the head of the code data, and then the code data may be added and stored in the memory. There is no problem whether the inverse transform data is raw or compressed. On the decompression side, the reverse conversion table portion is extracted from the head, set in the second table storage memory 123A, and the remaining code data is decompressed by the binary decompression unit 116. This eliminates the trouble of managing the code data separately.

  In the embodiment and the modification, the image compression / decompression apparatus 100, 100A incorporates a circuit portion for compressing image data and a circuit portion for decompression, but they may be stored in separate devices. Further, compression and decompression may be performed by hardware or may be realized by software.

  Further, in the modification of the present invention, the processing is automated by creating the first table data and the second table data corresponding to the multi-value image data input to the apparatus, but the present invention is not limited to this. . That is, as in the embodiment, a plurality of types of first table data and second table data corresponding to various types of multi-value image data are created in advance, and the characteristics of the multi-value image data to be processed are measured. Thus, the device side may determine which type of table data is suitable and perform the selection operation. In this case, information indicating which type of table data is used is supplied to the first table storage memory and the second table storage memory.

1 is a block diagram illustrating a configuration of an image compression / decompression apparatus according to an embodiment of the present invention. It is explanatory drawing which showed the principle operation | movement of the screen process part of a present Example. It is explanatory drawing of the general technique showing a mode that the image data of a fixed image area | region was expand | deployed to a bit plane. It is explanatory drawing showing a mode that the gradation bit was reduced and the image data of a predetermined image area | region was expand | deployed to the bit plane in a present Example. It is explanatory drawing showing these relationships about an example of 1st table data and 2nd table data in a present Example. FIG. 5 is a characteristic diagram showing a histogram of multi-valued image data representing the character “A” shown in FIG. 4 in the present embodiment. It is a characteristic view showing a histogram of multi-value data after screen processing in the present embodiment. It is explanatory drawing showing three bit planes as a result of carrying out table conversion of the image data in a present Example. It is explanatory drawing which showed the method of expand | deploying 3 bit planes in 1 bit plane in a present Example. It is a block diagram showing the structure of the image compression / decompression apparatus in the modification of this invention. It is a flowchart showing the outline | summary of the analysis process of the data analysis part in a modification. It is a block diagram showing the principal part of the structure of the image processing apparatus by a 1st proposal. It is explanatory drawing showing the separation table of the bit plane separator by a 1st proposal. It is explanatory drawing which represented in principle the processing content of the compression preprocessor by a 1st proposal. It is explanatory drawing showing the table which converts the gradation number by a 2nd proposal into gray code data.

Explanation of symbols

100, 100A Image compression / decompression apparatus 101 Multi-valued image data (input side) 102 Screen processing unit 103 Screen pattern memory 106, 106A Table conversion processing unit 107, 107A First table storage memory 108, 108A First table data 111 Bit plane conversion unit 113 Binary compression unit 115, 115A Data storage memory 116 Binary decompression unit 119 Multi-value data restoration unit 122, 122A Table inverse conversion processing unit 123, 123A Second table storage memory 124, 124A Second table Data 125, 125A Multi-level data (output side) 161-163, 171 Bit plane

Claims (6)

Preprocessing means for inputting image data representing the density of each pixel in multi-values and reducing the number of constituent bits representing the density in units of pixels with a predetermined logic to a predetermined value of 2 or more;
A bit string conversion means for converting a bit string representing the density of each pixel whose number of constituent bits has been reduced by the preprocessing means into a bit string previously associated with the appearance frequency of each density;
Bit plane expansion means for expanding the bit string after being converted by the bit string conversion means into a bit plane for each bit;
An image processing apparatus comprising: an image compression unit that compresses the bit plane expanded by the bit plane expansion unit. The image processing apparatus according to claim 1, wherein the bit plane expanding unit expands each bit plane constituting the bit string in a single plane . The image processing apparatus according to claim 1, wherein the image data is color image data, and the preprocessing unit performs preprocessing for each constituent color . The bit string conversion means includes a frequency calculation means for determining the frequency of each density of pixels whose number of constituent bits has been reduced by the preprocessing means, and the bit plane expansion means in descending order of the frequency calculated by the frequency calculation means. The image processing apparatus according to claim 1 , further comprising: a bit string allocating unit that allocates a bit string that can be suppressed from appearing when expanding to a bit plane . 5. The image processing apparatus according to claim 4, wherein the bit string conversion means calculates a bit string conversion table according to the type of image data comprising pixels whose number of constituent bits has been reduced by the preprocessing means. A preprocessing step of inputting image data representing the density of each pixel in a multi-value, and reducing the number of constituent bits representing the density in units of pixels with a predetermined logic to a predetermined value of 2 or more;
A bit string conversion step for converting a bit string representing the density of each pixel whose number of constituent bits has been reduced in this preprocessing step into a bit string previously associated with the appearance frequency of each density;
A bit plane expansion step of expanding the bit string after being converted by this bit string conversion step into a bit plane for each bit;
An image compression step for compressing the bit plane expanded in the bit plane expansion step;
An accumulation step for accumulating the data compressed in this image compression step in a data accumulation memory;
An image decompression step for reading and decompressing the data accumulated in this accumulation step;
A print data output step for inputting the bit plane obtained by the image expansion step and performing a reverse conversion to the bit string conversion step to obtain multi-value data for printing.
An image processing method comprising:

Images