JP3864681B2 - Image data encoding apparatus, image data encoding method, recording medium, and image forming apparatus - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image data encoding apparatus, an image data encoding method, a recording medium, and an image forming apparatus that encode multi-value image data in which a pixel value of one pixel is represented by a plurality of bits.
[0002]
[Prior art]
In recent years, in the field of image forming apparatuses, for example, image data for an image that has already been formed is stored in a memory, and when there is an instruction from a user, the image data is taken out of the memory and then formed again. A device having a so-called memory recall function has been put into practical use.
[0003]
When this memory recall function is used, when image formation is performed again using image data stored in the memory, an original is read again by an image reader or the like, or a personal computer (hereinafter referred to as “PC”). Therefore, it is generally considered that the user-friendliness improves as the number of images that can be stored in the memory increases.
[0004]
However, if the memory capacity is increased in order to increase the number of images that can be stored, there arises a problem that the manufacturing cost of the image forming apparatus increases. This problem occurs when, for example, an attempt is made to increase the number of pages that can be sorted in the case of using a so-called electronic sort function in which image data is once stored in a memory and then sorted to output a plurality of copies. This is particularly noticeable in a full-color image forming apparatus that needs to store image data for each reproduction color of yellow, cyan, magenta, and black as well as a monochrome image.
[0005]
In order to suppress an increase in manufacturing cost associated with an increase in the capacity of such a memory, various methods for encoding image data have been studied in order to compress and store the image data in the memory. Hereinafter, as an example of a conventional image data encoding method, block truncation compression (hereinafter referred to as “BTC method”) in which multi-value image data in which a pixel value of one pixel is expressed by a plurality of bits is fixed-length encoded. Will be described. This method is described as a prior art in Japanese Patent Laid-Open No. 10-271299.
[0006]
FIG. 9 is a diagram for explaining the BTC method. The figure shows an example in which 256-level multi-value image data in which the pixel value of one pixel is represented by 8 bits is fixed-length encoded, as shown in FIG. The figure shows a case where encoding is performed by sequentially cutting out a block 902 for every 4 vertical pixels * 4 horizontal pixels from the multi-valued image data 901.
[0007]
Specifically, first, from the pixel values of the respective pixels D0 to Df (see FIG. 5B) included in the block 902 of 4 vertical pixels × 4 horizontal pixels (16 pixels in total) cut out from the image data 901. The maximum gradation level (QMAX) and the minimum gradation level (QMIN) are obtained. This process simply obtains the largest pixel value and the smallest pixel value from the 16 pixels.
[0008]
Next, average data (LA) and gradation dynamic range (LD) are obtained. Here, LA is the sum of QMAX and QMIN divided by 2, and LD is the subtraction of QMIN from QMAX, each represented by 8 bits. The LA and LD are hereinafter referred to as “gradation characteristic data” in the sense of data representing the gradation characteristics of the pixels in the block.
[0009]
After obtaining the gradation characteristic data, the values of LMAX and LMIN are further obtained as quantization references. Here, LMAX and LMIN are expressed by the following (formula 1) and (formula 2), and are values serving as threshold values when obtaining quantized data described later.
LMAX = (3 * QMAX + QMIN) / 4 (Equation 1)
LMIN = (QMAX + 3 * QMIN) / 4 (Expression 2)
Using the values obtained as described above, the quantization process of 16 pixels included in the block 902 is performed. Here, the case where each pixel is quantized to 2 bits will be described. When the pixel value exceeds LMAX, the quantization bit is set to “11”, and when the pixel value exceeds LA and is equal to or less than LMAX, The quantization bit is set to “10”.
[0010]
Further, when the pixel value exceeds LMIN and is equal to or less than LA, the quantization bit is set to “01”, and when the pixel value is equal to or less than LMIN, the quantization bit is set to “00”. By the above processing, compressed data (hereinafter referred to as “BTC data”) 903 as shown in FIG. 9C is obtained. Since BTC data is obtained by compressing 16 pixels of image data in which one pixel is represented by 8 bits into 48 bits, the data amount is compressed to 3/8. This BTC method is an encoding method that utilizes the fact that the pixel values of pixels in the same block are almost equal, and is a so-called irreversible compression with little deterioration in image quality, easy rotation processing, and simple This is an image compression technique that is considered to be particularly suitable for application to an image forming apparatus.
[0011]
In the conventional method described in Japanese Patent Laid-Open No. 10-271299, after obtaining BTC data corresponding to all image data by the BTC method, the obtained BTC data is directly used as the original image data for each block. They are arranged vertically and horizontally while maintaining the positional relationship in 901, and are encoded using a technique such as arithmetic coding defined in JBIG (Joint Bilevel Image Group), for example.
[0012]
[Problems to be solved by the invention]
However, the conventional method has a limit in improving the compression rate. The present invention has been made in view of such a problem, and opens up a new possibility of improving the compression rate more than before, and thus can increase the amount of images that can be stored in a recording medium of a certain capacity. It is an object of the present invention to provide an image data encoding device, an image data encoding method, a recording medium, and an image forming apparatus that have a possibility.
[0013]
[Means for Solving the Problems]
  In order to achieve the above object, an image data encoding device according to the present invention is an image data encoding device that encodes multi-value image data in which a pixel value of one pixel is represented by a plurality of bits. Read-out means for reading out image data for each predetermined number of pixels, and from the pixel values of the plurality of pixels read out by the read-out means, fixed-length gradation characteristic data representing the gradation characteristics of the plurality of pixels, The pixel value of each of the plurality of pixels is set to a fixed length by a predetermined threshold value.nFixed-length encoding means for generating quantized data quantized into bits, and data generated by the fixed-length encoding means for the gradation characteristic data and the quantized data corresponding to each pixelEach mth bit (1 ≦ m ≦ n)Consisting of a bit ofn setsDividing means for dividing into bit data and divided by the dividing meansThe gradation characteristic data and the n sets of bit data are each developed on a two-dimensional plane.And entropy encoding means for performing entropy encoding.
[0014]
  In the image data encoding method according to the present invention, the fixed-length encoded data obtained from the original multi-valued image data is composed of gradation characteristic data and quantized data for each pixel.nIdentical in bit sequence of bitsM thConsist of bitsn setsDivided into bit data,The gradation characteristic data and the n sets of bit data are each developed on a two-dimensional plane.Entropy encoding.Each of the n sets of bit data of the gradation characteristic data and the quantized data divided in this way has lower randomness than the data in which the gradation characteristic data before the division and the quantized data are continuous, and the entropy. The improvement of the compression rate in encoding can be expected.
[0015]
Therefore, for example, as a method of entropy encoding, based on the probability value of the value of the target bit corresponding to the state of the value of a plurality of reference bits existing at a predetermined position with respect to the target bit to be encoded In the case of performing arithmetic coding with prediction, for example, in the case of performing arithmetic coding specified in JBIG, the possibility of improving the compression rate of gradation characteristic data compared to the conventional method is shown. Even when other entropy coding methods are used, the compression rate can be improved particularly for bit data composed of the most significant bits. This is because the bit data composed of the most significant bits is likely to have the same value for adjacent pixels as in the original image data. The predetermined threshold value can be determined based on the gradation characteristic data.
[0016]
Here, the fixed-length encoding means includes gradation difference determining means for determining the magnitude of the gradation difference of the pixel values of the read-out pixels from the gradation characteristic data, and the gradation When the difference determining unit determines that the gradation difference is smaller than a predetermined size, a predetermined bit string may be generated as quantized data corresponding to the plurality of pixels. According to this configuration, for example, the quantized data in the case where there is almost no gradation difference among the pixel values in the block can be a bit string of all 0 (or all 1), for example. In addition to the case where the above arithmetic coding is performed, the compression rate can be improved not only when the entropy coding method such as MH, MR, or MMR is used.
[0017]
The fixed-length encoding unit further includes an average value acquisition unit that acquires an average value of pixel values of the plurality of read-out pixels from the gradation characteristic data, and the gradation difference determination unit Therefore, when it is determined that the gradation difference is smaller than a predetermined size, and the value acquired by the average value acquisition unit is within a predetermined range, it further corresponds to the plurality of pixels. It is also possible to generate a predetermined bit string as gradation characteristic data to be performed. Specifically, when the gradation difference between the plurality of pixels is smaller than a predetermined size and the average value of the pixel values of the plurality of pixels is close to the maximum value of the pixel values, the image is a solid black image. If tone characteristic data representing that is generated or if the average value of the pixel values is close to the minimum value of the pixel values, it is assumed that the image is a solid white image, It is possible to generate characteristic data. Furthermore, in this case, the number of quantization bits of the quantized data may be reduced. Specifically, quantized data quantized with a plurality of bits can be converted into quantized data with one bit quantized. This is because it is considered to be substantially close to a binary image, and even if the number of bits of quantized data is reduced, it is considered that there is little influence on the image quality.
[0018]
The image data encoding apparatus may further include code conversion means for performing code conversion prior to entropy encoding on at least part of the data divided by the dividing means. As described above, in the method of the present invention, the data to be entropy-encoded is very weakly related to the original image data. Specifically, for example, binary data is obtained every predetermined number of bits. The compression ratio can be improved by converting from gray code to gray code.
[0019]
Further, the readout means reads out pixels for each block comprising the same number of pixels in the vertical direction and the horizontal direction from the multi-valued image data, and the entropy encoding means reads the data to be entropy encoded. Each of which includes storage means for storing in a form expanded on a two-dimensional plane so that the number of bits in the vertical direction and the number of bits in the horizontal direction for each block are equal, and entropy coding for each line of the storage means It is convenient to perform image rotation processing (especially 90 ° rotation). This is because, for example, it is possible to easily perform rotation processing in a state where data that has been fixed-length encoded by the BTC method is divided.
[0020]
The gradation characteristic data includes an average value of pixel values of a plurality of pixels read by the reading unit and a value of a gradation dynamic range, and is read by the reading unit. It may be composed of a maximum value and a minimum value of pixel values of a plurality of pixels. Here, the dividing unit may further divide an average value and a gradation dynamic range value of the pixel values of the plurality of pixels, or a maximum value and a minimum value of the pixel values of the plurality of pixels. .
[0021]
In addition, the image data encoding device further obtains a compression rate of the encoding by the entropy encoding unit for the bit data corresponding to the most significant bit among the quantized data divided by the dividing unit. An decoding unit for comparing the compression rate acquired by the acquisition unit, the compression rate acquired by the compression rate acquisition unit, and a comparison result by the comparison unit; Determining means for determining whether other bit data of the divided quantized data is necessary. Based on the compression rate of the most significant bit, it is possible to determine whether the image is close to a binary image such as a character image or whether the image includes many intermediate gradation pixels such as a photographic image. is there.
[0022]
Specifically, the entropy encoding unit can prevent the encoding of the other bit data when the determination unit determines that the other bit data is not necessary. When the image data encoding device determines that the other bit data is not required by the encoded data storage unit that stores the data encoded by the entropy encoding unit and the determination unit. Can also include encoded data deleting means for deleting encoded data corresponding to the other bit data stored in the encoded data storage means.
[0023]
  The image data encoding method according to the present invention is an image data encoding method for encoding multi-value image data in which a pixel value of one pixel is represented by a plurality of bits, and the multi-value image data is A readout step for reading out a predetermined number of pixels, and a fixed-length gradation characteristic data representing gradation characteristics of the plurality of pixels from the read pixel values of the plurality of pixels, and pixel values of each of the plurality of pixels A fixed length with a predetermined thresholdnA fixed-length encoding step for generating quantized data quantized into bits, and the generated data for the gradation characteristic data and quantized data corresponding to each pixel;Each mth bit (1 ≦ m ≦ n)Consisting of a bit ofn setsDividing step into bit data and dividedThe gradation characteristic data and the n sets of bit data are each developed on a two-dimensional plane.An entropy encoding step for entropy encoding.
[0024]
  Furthermore, the first recording medium according to the present invention is a computer-readable recording in which a program for realizing an image data encoding process for encoding multi-value image data in which a pixel value of one pixel is represented by a plurality of bits is recorded. The medium is a program recorded on the recording medium, or the program recorded on the recording medium is operated on a computer together with other general-purpose programs. Read processing for reading out multi-valued image data for each predetermined number of pixels, fixed-length gradation characteristic data representing gradation characteristics of the plurality of pixels from the read pixel values of the plurality of pixels, and the plurality Each pixel value of thenFixed-length encoding processing that generates quantized data quantized into bits, and the generated data is converted into the gradation characteristic data and the quantized data corresponding to each pixel.Each mth bit (1 ≦ m ≦ n)Consisting of a bit ofn setsDividing process into bit data and dividedThe gradation characteristic data and the n sets of bit data are each developed on a two-dimensional plane.It is a program for causing a computer to execute processing including entropy encoding processing for entropy encoding.
[0025]
According to the above recording medium, the image data encoding method according to the present invention can be used not only with an image forming apparatus but also with a general-purpose information processing apparatus such as a PC. Note that the recording medium may record all the programs necessary for realizing the image data encoding method of the present invention, or the recording medium records only the minimum necessary programs. After being installed in the computer, the method of the present invention is realized by using a function such as an operating system or a general-purpose program such as another software product and operating together with the other program. Also good.
[0026]
  Furthermore, the second recording medium according to the present invention is a recording medium that records data obtained by encoding multi-value image data in which a pixel value of one pixel is represented by a plurality of bits, and the encoded data is: Multi-valued image data before encoding is read for each predetermined number of pixels, fixed-length gradation characteristic data representing gradation characteristics of the plurality of pixels from the pixel values of the plurality of pixels read, and the plurality The pixel value of each of the pixels of a fixed length by a predetermined thresholdnAfter generating quantized data quantized into bits, the gradation characteristic data and the quantized data corresponding to each pixelEach mth bit (1 ≦ m ≦ n)Consisting of a bit ofn setsDivided into bit data and dividedThe gradation characteristic data and the n sets of bit data are each developed on a two-dimensional plane.It is characterized by being data encoded by entropy encoding.
[0027]
As described above, the recording medium storing the data encoded by the image data encoding method according to the present invention has a possibility of storing more images even when the recording medium having the same capacity is used. For example, an embodiment in which the recording medium is sold and image formation is performed in an image forming apparatus of the present invention described later is conceivable. Here, the recording medium includes a disk type medium such as a CD-ROM and a DVD-ROM, and a memory card such as Smartmedia (registered trademark).
[0028]
An image forming apparatus according to the present invention includes data encoded by the image data encoding apparatus according to the present invention, data encoded by the image data encoding method according to the present invention, and the first recording medium according to the present invention. Decoding means for decoding at least one of data encoded in a computer in which the program recorded on the computer is operable, data recorded on the second recording medium according to the present invention, and the decoding And image forming means for forming an image using the image data decoded by the converting means. Thereby, image formation can be performed using the image data encoded by the image data encoding method according to the present invention. In the image forming apparatus of the present invention, it is also possible to perform image formation by decoding encoded data transmitted via a wired or wireless network.
[0029]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of an image data encoding apparatus and the like according to the present invention will be described with reference to the drawings, taking as an example a case where the image data encoding apparatus is applied to an image forming apparatus.
(Embodiment 1)
(1) Overall configuration of image forming apparatus
FIG. 1 is a schematic cross-sectional view showing an overall configuration of a digital copying machine (hereinafter simply referred to as “copying machine”) 1 as an example of an image forming apparatus.
[0030]
The copying machine 1 of the present embodiment includes a scanning system 10 that reads a document and converts it into an image signal, an image signal processing unit 20 that processes an image signal sent from the scanning system 10, and image data that is input from the image signal processing unit 20. Or, multi-value image data transmitted from an external device such as a PC via a network is encoded and stored in a memory, or output from the memory unit 30 and the memory unit 30 for decoding the encoded data. Print processing unit 40 for driving the semiconductor laser 62 based on the image data to be processed, an external interface unit 50 for sending image data transmitted from an external device via a network to the memory unit unit 30, and a laser beam from the semiconductor laser 62 An optical system 60 that guides the photosensitive drum 71 to an exposure position on the photosensitive drum 71, develops the electrostatic latent image formed by the exposure, transfers it to a recording paper, And image forming system 70 to form an image, and the like document transport unit 500 that performs reversal of sides as required to convey the document.
[0031]
The external interface unit 50 sends image data from an external device to the memory unit unit 30 and also performs control related to transmission and reception of control signals and image data with the external device. The image data is sent from the memory unit 30 to the print processing unit 40, and the semiconductor laser 62 is driven based on this image data. Laser light generated from the semiconductor laser 62 is deflected by the optical system 60 and guided to the image forming system 70.
[0032]
In the image forming system 70, an electrostatic latent image is formed on the surface of the photosensitive drum 71 by irradiating the photosensitive drum 71 that is uniformly charged by the charging charger 72 and rotating in the direction of arrow A with the laser beam. The electrostatic latent image is visualized by the supply of toner from the developing device 73. The visualized toner image is transferred to the recording paper conveyed from the paper feed cassettes 80 a and 80 b by the transfer charger 74, and further, the recording paper is conveyed to the position of the fixing roller 84, and the toner is conveyed by the fixing roller 84. By fixing the image on the recording paper, the image is finally formed on the recording paper.
[0033]
(2) Configuration of image data encoding device
Next, an embodiment of an image data encoding device according to the present invention will be described. In the present embodiment, the image data encoding device of the present invention is provided in the memory unit 30. FIG. 2 is a functional block diagram showing the configuration of the image data encoding apparatus according to the present embodiment. In this embodiment, an example applied to a copying machine will be described, but it can also be applied to various other purposes.
[0034]
The image data encoding apparatus according to the present embodiment includes an image data acquisition unit 101 that acquires image data, a BTC compression processing unit 102 that obtains BTC data by performing fixed-length encoding on the acquired image data using the BTC method, and an obtained BTC. The BTC division processing unit 103 that divides the data and stores it in the buffer memory 104 in a form expanded on a two-dimensional plane, and stipulated in JBIG for the divided BTC data (hereinafter also referred to as “divided BTC data”). The buffer memory 104 for temporarily storing the divided BTC data and the buffer memory 104 in order to perform compression or expansion using a method of arithmetic coding (hereinafter referred to as “JBIG compression”) with prediction using a template that has been used. JBIG performs compression and decompression processing based on the JBIG compression technique on the divided BTC data stored in Reduced expansion processing unit 105, coded data compressed by JBIG compression expansion processing unit 105 and a code memory 106 to be stored. In the code memory 106, encoded data is stored in association with, for example, an identifier indicating the current image data, and the stored encoded data is subjected to expansion processing by the JBIG compression / decompression processing unit 105 upon decoding. The data is stored in the buffer memory 104, and is output via the image data output unit 109 after the BTC data restoration processing by the BTC restoration processing unit 107 and the decoding processing by the BTC decoding processing unit 108.
[0035]
The image data acquisition unit 101 receives image data transmitted from the image signal processing unit 20, image data transmitted from an external device such as a PC, and the like. In the present embodiment, it is assumed that 256-tone monochrome image data in which the pixel value of one pixel is represented by 8 bits is input.
The BTC compression processing unit 102 generates BTC data from the input image data. Since the BTC data generation processing has already been described, the description thereof is omitted here.
[0036]
The BTC division processing unit 103 divides the BTC data into gradation characteristic data and each bit data generated from the quantized data, and stores the divided BTC data in the buffer memory 104 in a format to be developed on a two-dimensional plane. To do. The image data encoding apparatus according to the present embodiment opens the possibility of realizing a higher compression rate in the JBIG compression, for example, by performing this BTC data division processing. The division processing of BTC data and the development processing of the divided BTC data on a two-dimensional plane (hereinafter also referred to as “two-dimensional development”) will be described in detail later.
[0037]
The data stored in the buffer memory 104 is subjected to JBIG compression processing in the JBIG compression / decompression processing unit 105 and stored in the code memory 106. The JBIG compression / decompression processing unit 105 used in the present embodiment is configured around a so-called QM coder, and includes a line memory for two lines or three lines therein. Prediction using a so-called template described later is performed by this line memory.
[0038]
The encoded data stored in the code memory 106 is decoded by the JBIG compression / decompression processing unit 105 and expanded on the buffer memory 104 before being decoded by the BTC restoration processing unit 107. The processing opposite to the BTC data division processing (BTC data restoration processing) by the processing unit 103 is performed, and the BTC decoding processing unit 108 performs processing opposite to the encoding processing by the BTC compression processing unit 102. The image data is output via the image data output unit 108. The BTC decoding processing unit 108 performs decoding from the quantized data as described above. Since the quantized data cannot reproduce the same image as the original image, the decoded image data Is strictly different from the original image (lossy compression). The BTC data restoration process and decoding process will be described later.
[0039]
(3) Contents of image data encoding process
Next, the contents of the image data encoding process in the present embodiment will be described. Note that the image data encoding operation of the present embodiment is controlled by a control unit (not shown) that is configured around a CPU.
FIG. 3 is a flowchart showing the processing content of the image encoding processing. As shown in the figure, in the image encoding processing of the present embodiment, when image data is acquired, BTC data is generated as described above (S101), and the BTC data is divided and two-dimensionally expanded to be buffered. Store in the memory 104 (S102). Here, BTC data division processing and two-dimensional development will be described in detail.
[0040]
FIG. 4 is a diagram for explaining the BTC data division processing and the two-dimensional development of the divided BTC data according to the present embodiment. As shown in the figure, in the BTC data division and two-dimensional expansion processing, when the BTC data 201 is stored in the buffer memory 104, the gradation characteristic data (8 bits per block * 2 = 16 bits), the upper bits (16 bits per block) of the quantized data (2 bits for each pixel), and the lower bits (16 bits per block) of the quantized data, Each of the 16 bits is two-dimensionally expanded into vertical 4 bits * horizontal 4 bits and individually stored in the buffer memories 104a to 104c. This processing is performed for all blocks cut out from the image data, and sequentially stored in the buffer memories 104a to 104c. In addition, regarding the arrangement of each block in the buffer memories 104a to 104c, in this embodiment, the arrangement is sequentially arranged in the vertical and horizontal directions while maintaining the positional relationship on the original image data. However, this arrangement can be changed. It is.
[0041]
Returning to the flowchart of FIG. 3, when the above-described BTC data division and two-dimensional expansion processing are completed for a predetermined area (S103: Yes), the divided BTC data divided and stored in the buffer memories 104a to 104c. Is subjected to JBIG compression processing for each line (S104). Here, the processing for the predetermined area is determined to be finished when the JBIG compression processing for each line becomes possible. Specifically, for example, when arrangement of each block of 4 bits in the vertical direction and 4 bits in the horizontal direction is completed for all the horizontal areas of the buffer memories 104a to 104c (4 columns of data are arranged in the vertical direction) It can be determined that the processing for the predetermined area has been completed.
[0042]
In the present embodiment, in the subsequent JBIG compression processing (S104), the JBIG compression processing is performed in parallel for each of the data stored in the buffer memories 104a to 104c. To that end, specifically, the JBIG compressor may be provided in parallel for the number of divided BTC data (three in the present embodiment), or may be operated in parallel. A small number may be provided to perform time-sharing processing.
[0043]
As described above, when the encoding process for all areas is completed (S105: Yes), the data encoded by the JBIG compression process is associated with, for example, an identifier indicating original image data, for example, in the code memory 106. And the image data encoding process ends. The encoded data can be easily decoded by performing a process reverse to the encoding process described above. That is, as shown in FIG. 5, first, the data encoded in the divided state is decompressed by the JBIG compression / decompression processing unit 102 to restore the divided BTC data 104a to 104c.
[0044]
Since the restored divided BTC data 104a to 104c are stored in the buffer memory 104, the BTC restoration processing unit 107 rearranges the data to restore the BTC data 201. Since a known method can be used for the decoding process of the BTC data 201 by the BTC decoding processing unit 108, detailed description thereof is omitted.
[0045]
Next, the reason why a high compression rate can be expected in the JBIG compression performed after that by performing the above-described BTC data division and two-dimensional expansion processing will be described.
FIG. 6 is a diagram showing an example of a so-called template used in the JBIG compression process. FIG. 4A is a template used when referring to bit data extending over three lines, and FIG. 4B is a template when referring to data of two lines.
[0046]
In each template, “?” Represents a bit to be encoded (hereinafter referred to as “target bit”), and “X” represents a bit referred to in predicting the target bit (hereinafter referred to as “reference bit”). It is said.) Although “A” is a reference bit, it is a bit that can be moved within a certain range, and a template in which A is moved from the initial position is referred to as an adaptive template (AT).
[0047]
As described above, in the arithmetic coding process involving the prediction using the template, prediction is performed based on the probability value of the value of the bit of interest based on the reference bit (including the pixel indicated by A) 10-bit state (Markov state). Therefore, it is expected that the higher the ratio of the probability of the value of the target bit derived from the Markov state of the reference bit biased to either 0 or 1, the higher the compression rate will be. On the other hand, the two-dimensionally developed gradation characteristic data, in particular, is a bit string that includes only the gradation characteristic data and has significantly different characteristics from the BTC data that was the target of JBIG compression in the conventional method. . That is, the bit string appearing on the two-dimensional plane does not include a random part due to a normal natural image, and the probability value of the value of the bit of interest corresponding to the Markov state of the reference bit is also compared with the conventional BTC data. It is thought that a large bias may occur, and therefore it can be expected that a very good compression rate is realized as compared with the conventional method.
[0048]
On the other hand, bit data obtained from quantized data, particularly data including only the upper bits, is likely to have the same value for adjacent pixels as in the original image data. An improvement in compression rate can be expected as compared with the case where encoding is performed in a state where upper bits and lower bits are mixed. As described above, it is considered that the compression ratio of the entire image can be greatly improved by dividing the BTC data.
[0049]
However, even considering the principle of the JBIG compression process, it is not guaranteed that the compression rate is dramatically improved for all images as compared with the conventional art. For example, it seems that there is a possibility that the compression rate may differ depending on the type of image such as characters and photographs and colors, but depending on the type of the image, a considerable improvement in the compression rate may be realized. it is conceivable that.
[0050]
Further, the gist of the present invention lies in the division of BTC data. Therefore, as described above, by performing the JBIG compression processing on the divided BTC data, there is a possibility that a particularly high compression ratio can be realized. However, the present invention is not limited to the above encoding method. For example, instead of prediction using a template, so-called predictive encoding (TP) may be performed as appropriate, or when encoding divided BTC data, Other entropy encoding methods such as general MH, MR, and MMR may be used. It is also possible to improve the compression rate using the AT.
[0051]
Further, in the above-described embodiment, the case where two-dimensional expansion is performed on all of the three parts of the divided BTC data (gradation characteristic data, upper bits and lower bits of the quantized data) has been described. However, the divided BTC data For example, only the gradation characteristic data may be developed two-dimensionally, and the other parts may be encoded without being two-dimensionally developed. As already mentioned, since the compression ratio can change depending on the properties of the image, it is not limited to the method of the above embodiment, and various methods can be used for specific encoding. It is thought to be.
[0052]
(Modification)
As described above, the image data encoding method according to the present invention has been described based on the embodiments. However, the content of the present invention is not limited to the specific examples described in detail above and the modified examples considered as appropriate. Further, for example, the following modifications can be considered.
[0053]
(1) In the above embodiment, the case where 256-level multi-value image data in which the pixel value of one pixel is represented by 8 bits has been described has been described. However, when the pixel value of one pixel is 8 bits, Of course, it is not limited, and the quantization processing in the BTC method is not limited to the case where the pixel value of each pixel is quantized to 2 bits, and is easy even when the number of bits of the quantized data is increased. It is possible to deal with. Further, the present invention can be applied not only to a monochrome image but also to a full color image. This is because, for example, encoding may be performed for each reproduction color of cyan, magenta, yellow, and black.
[0054]
(2) In the above embodiment, the values of LA and LD are used as the gradation characteristic data. However, the present invention is not limited to this, and any other information can be used as long as it is information that can determine the gradation characteristics of the pixels in the block. It is also possible to use data. Specifically, the maximum value (QMAX) and the minimum value (QMIN) among the pixel values of the pixels in the block may be held. Further, it is possible to improve the compression rate by determining the type of image (for example, character, photograph, etc.) from the pixel value of the multi-value image data and changing the content of the gradation characteristic data according to the determination result. Is also possible.
[0055]
(3) Some image data includes a white solid portion, a black solid portion, or an intermediate density solid portion. In these solid areas, the pixel values of the pixels included in the cut-out block may all be the same or almost the same. In this case, even if the pixel value quantization processing in the BTC method is performed, all pixels Or, the quantization bits of most of the pixels are the same (hereinafter, such a state is referred to as “a state where there is no gradation difference in the block”).
[0056]
On the other hand, whether or not there is a gradation difference can be determined from the gradation characteristic data. Specifically, when the value representing LD is smaller than a predetermined threshold value (when QMAX and QMIN are held as gradation characteristic data, the difference between the two is smaller than the predetermined threshold value), It can be determined that there is no gradation difference.
[0057]
If it is determined by the above determination that there is no gradation difference, the same amount is set as the quantized data, thereby reducing the amount of computation by the BTC compression processing unit 102 required for generating the BTC data. It is also possible to improve the compression rate in entropy coding (including JBIG compression) by setting all quantized data to 0 (or all 1). That is, by performing such processing, the compression rate can be improved not only when performing JPIG compression but also when using run-length encoding such as MH and MR.
[0058]
FIG. 7 is a flowchart showing an example of the content of the data rounding process based on such a gradation difference discrimination. This data rounding process can be executed in step S101 in the flowchart of FIG. That is, in the example of FIG. 7, it is first determined whether or not the acquired value of the gradation dynamic range LD is smaller than a predetermined threshold value (LD_TH) (S201).
[0059]
If the LD value is equal to or greater than the predetermined threshold value (S201: No), the data rounding causes degradation of the image quality, so that subsequent data rounding processing is not performed, but the LD value is smaller than the processing threshold value. In this case (S201: Yes), it is determined whether or not the value of LA is smaller than the threshold value (LAW_TH) for white solid image discrimination (S202). Since the value of LA indicates an average value of the pixel values in the block, when this value is smaller than the white solid discrimination threshold (S202: Yes), the block is a white solid image. And BTC data indicating that fact is generated (S203). Specifically, the data can be rounded by setting all the values of LA and LD to 0 and making the quantized data all the same for all pixels (all the bits of the quantized data are set to 0 or 1).
[0060]
On the other hand, if the LA value is equal to or greater than the threshold value for white solid determination (S202: No), it is next determined whether the LA value is larger than the threshold value for black solid determination (S204). If the value of LA is larger than the black solid determination threshold (S204: Yes), it is determined that the block indicates a black solid image portion, and BTC data indicating that is generated (S205). . Specifically, the maximum pixel value is set as the LA value, and 0 is set as the LD value. Then, it is possible to round the data by making the quantized data the same for all the pixels (all the bits of the quantized data may be 0 or 1).
[0061]
If the LA value is equal to or less than the threshold value for black solid determination in step S204 (S204: No), it can be determined that it is a solid portion of intermediate density, and therefore data rounding processing for intermediate density is performed (S206). Specifically, the data is rounded by keeping the value of LA as it is, setting the values of LD as 0, and making the quantized data the same for all pixels. All the bits of the quantized data may be set to 0 or 1, as in the case of white solid or black solid.
[0062]
By performing the data rounding process as described above, the amount of calculation can be reduced and the compression rate can be improved. Each of the threshold values (LD_TH, LAW_TH, LAB_TH) can be set in advance in a register, for example. Further, in the case of a binary image other than white solid or black solid, the number of bits of the quantized data may be reduced to 1 bit. This is because in the case of a binary image, it is considered that there is almost no influence on the image quality of the decoded image.
[0063]
(4) Also, the number of quantization bits of the quantized data is initially quantized with a plurality of bits and entropy-coded when a bit plane composed of any one bit of the quantized data is entropy-coded. May be reduced. This is because it can be determined from the compression rate of the bit plane whether the image is close to a binary image such as a character image or includes many intermediate gradation pixels such as a photographic image.
[0064]
An example of specific processing contents when the number of bits of quantized data is reduced based on the compression rate will be described below. FIG. 8 is a flowchart showing the contents of the image data encoding process including the quantization data reduction process. Note that steps S301 to S305 in the figure are the same as the content of the image data encoding process described in the above embodiment, and thus the description thereof is omitted.
[0065]
In the example shown in the figure, in step S306, the compression rate for the bit plane composed of the most significant data is calculated. In the example of the figure, the most significant bit plane is used, but other bit planes may be used. This is because, regardless of which bit plane is used, there is a difference in compression rate between a case close to a binary image and an image including many intermediate gradation pixels.
[0066]
After calculating the compression rate, the calculated compression rate is compared with a predetermined threshold (S307). When the compression rate is larger than the predetermined threshold (S307: Yes), it is determined as a binary image, and the encoded data for the bit plane other than the most significant data stored in the code memory 106 is deleted. (S308). On the other hand, when the compression rate is equal to or lower than the predetermined threshold (S307: No), the image data encoding process is terminated as it is. Since it is difficult to determine an absolute value as the threshold value, it is necessary to separately determine in consideration of various conditions such as an entropy encoding method. However, when JBIG compression is performed, the average compression ratio for character originals (eight types of CCITT standard test charts) is about 20, and the compression ratio for pseudo-halftones (SCID dither / error diffusion) is about 1.25. Since it is set to about 7.17, when performing JBIG compression, it is conceivable to set the threshold to about 10-15.
[0067]
In the above example, all the bit planes are once encoded, and then unnecessary encoded data is deleted according to the compression rate. Entropy coding may be performed on a plane, and it may be determined whether to perform entropy coding on another bit plane based on the compression rate.
[0068]
(5) In the above-described embodiment, the two-dimensional expansion processing is performed on the gradation characteristic data so that LA and LD are combined to be vertical 4 bits * horizontal 4 bits. For the bit data, the two-dimensional expansion was performed so as to be vertical 4 bits * horizontal 4 bits. It is convenient to make the number of vertical bits equal to the number of horizontal bits in this manner in order to perform image rotation processing (especially 90 degree rotation) in the image forming apparatus. The number of bits in the vertical and horizontal directions is not particularly limited. Further, LA and LD (or QMAX and QMIN) may be further divided to perform two-dimensional expansion and encoding.
[0069]
Furthermore, the way of arranging the bits when two-dimensionally developing in the vertical and horizontal directions is not limited to the method of arranging the bits in order so that the most significant bit is arranged at the upper left and the least significant bit is arranged at the lower right as in the above embodiment. Considering the reason why the compression ratio can be expected to be improved by the method of the present invention, the rearrangement of bits is arbitrary as long as decoding is possible. Further, the compression rate can be further improved, for example, by performing code conversion from normal binary data corresponding to pixel values to gray codes every 4 bits. Further, for the same reason, the arrangement order of pixel units can be arbitrarily changed.
[0070]
(6) In the above embodiment, the case where the image data encoding apparatus according to the present invention is applied to the accumulation of image data in the image forming apparatus has been described. For example, it is also realized by operating software installed in an information processing apparatus such as a PC via various recording media such as a CD-ROM and a DVD-ROM, or via a wired or wireless network. As a distribution mode of the software, all software necessary for the image data encoding processing according to the present invention may be stored in the various recording media, or various operating systems installed in the computer in advance. A case where a function of a general-purpose program is used may be considered.
[0071]
On the other hand, in view of recent usage of image forming apparatuses, when image data encoding processing according to the present invention is performed using an information processing apparatus such as a PC, the encoded data is stored in a recording medium. The image data can be decrypted and formed using an image forming apparatus installed at the store. Such a configuration is also conceivable when the code memory of the image forming apparatus described in the above embodiment is configured using a detachable recording medium. As the recording medium used, for example, various recording media such as various media such as CD-R and floppy disk, and various memory cards such as SmartMedia (registered trademark) and CompactFlash can be considered and encoded. It goes without saying that an embodiment in which data is distributed via a wired or wireless network is also possible.
[0072]
(7) In addition to the image forming apparatus and the information processing apparatus, the method of the present invention may be applied to, for example, data encoding before image data transmission in a facsimile apparatus, or may be applied to various other apparatuses. Can do.
[0073]
【The invention's effect】
As described above, the image data encoding device according to the present invention is an image data encoding device that encodes multi-value image data in which a pixel value of one pixel is represented by a plurality of bits, and is a multi-value image. The data is read for each predetermined number of pixels, and from the read pixel values of the plurality of pixels, the fixed-length gradation characteristic data representing the gradation characteristics of the plurality of pixels and the pixel values of each of the plurality of pixels are obtained. Quantized data quantized into a plurality of fixed-length bits according to a predetermined threshold is generated, and the generated data is converted into the same order in the bit sequence of the gradation characteristic data and the quantized data corresponding to each pixel. Since each bit data is divided into bit data and at least part of the divided data is entropy-encoded, compared with the conventional image data compression method It can be said that opens up the possibility to further improve the compression ratio. The same effect can be obtained by the image data encoding method according to the present invention.
[0074]
Further, since the software necessary for realizing the image data encoding method according to the present invention is recorded on the first recording medium of the present invention, the software is, for example, a PC or other general-purpose software. When installed in a simple information processing apparatus or the like, the method of the present invention can be used in various apparatuses. In addition, data encoded by the image data encoding method according to the present invention is recorded on the second recording medium of the present invention. For example, even when a recording medium having the same capacity is used, more data is recorded. It opens up the possibility of storing images.
[0075]
Furthermore, since the image forming apparatus of the present invention has image forming means for decoding the data encoded by the image data encoding method according to the present invention and forming an image, the image forming apparatus is encoded by the method of the present invention. The processed data can be used for image formation. This opens up the possibility of storing a larger amount of image data with a small-capacity image memory, particularly when storing image data in the image forming apparatus.
[Brief description of the drawings]
FIG. 1 is a schematic cross-sectional view showing an overall configuration of a digital copying machine 1 as an example of an image forming apparatus as an application example of the present invention.
FIG. 2 is a functional block diagram showing a configuration of an image data encoding device according to an embodiment of the present invention.
FIG. 3 is a flowchart showing the processing contents of image data encoding processing in the embodiment of the present invention.
FIG. 4 is a diagram for explaining BTC data division processing and two-dimensional development of divided BTC data;
FIG. 5 is a diagram for explaining BTC data restoration processing;
FIG. 6 is a diagram illustrating an example of a template used in JBIG compression processing.
FIG. 7 is a flowchart showing an example of the contents of data rounding processing;
FIG. 8 is a flowchart showing the contents of image data encoding processing including quantization data reduction processing;
FIG. 9 is a diagram for explaining the BTC method.
[Explanation of symbols]
30 Memory unit
50 External interface section
101 Image data acquisition unit
102 BTC compression processing unit
103 BTC division processing unit
104 Buffer memory
105 JBIG compression / decompression processor
106 Code memory
107 BTC restoration processing unit
108 BTC decoding processing unit
109 Image data output unit
201 BTC data

Claims (24)

An image data encoding device for encoding multi-value image data in which a pixel value of one pixel is represented by a plurality of bits,
Reading means for reading out the multi-value image data for each predetermined number of pixels;
From the pixel values of the plurality of pixels read by the reading means, fixed-length gradation characteristic data representing the gradation characteristics of the plurality of pixels and the pixel values of the plurality of pixels are fixed by a predetermined threshold. Fixed-length encoding means for generating quantized data quantized to n bits of
The data generated by the fixed-length encoding means is an n-set bit composed of the gradation characteristic data and the m-th bit (1 ≦ m ≦ n) of the quantized data corresponding to each pixel. Dividing means for dividing into data,
Image data, comprising: entropy encoding means for entropy encoding the gradation characteristic data and the n sets of bit data divided by the dividing means in a form developed on a two-dimensional plane. Encoding device. The predetermined threshold is:
The image data encoding device according to claim 1, wherein the image data encoding device is determined based on the gradation characteristic data. The fixed length encoding means includes
A gradation difference determining means for determining a magnitude of a gradation difference of pixel values of the plurality of read pixels from the gradation characteristic data;
The predetermined bit string is generated as quantized data corresponding to the plurality of pixels when the gradation difference determining unit determines that the gradation difference is smaller than a predetermined size. The image data encoding device according to 1 or 2. The fixed length encoding means further includes:
Average value acquisition means for acquiring an average value of pixel values of the plurality of read-out pixels from the gradation characteristic data;
By the gradation difference determination means, even if the gradation difference is judged to be smaller than the predetermined size, when the value obtained by the previous SL average value acquisition unit is within a predetermined range, further The image data encoding apparatus according to claim 3, wherein a predetermined bit string is generated as gradation characteristic data corresponding to the plurality of pixels. The fixed length encoding means further includes:
When it is determined by the gradation difference determination means that the gradation difference is smaller than a predetermined magnitude, and the value acquired by the average value acquisition means is within a predetermined range, the plurality of The image data encoding apparatus according to claim 4, wherein a predetermined bit string is generated as gradation characteristic data corresponding to each pixel and the number of quantization bits of the quantization data is reduced. The entropy encoding means includes
Performing arithmetic coding with prediction based on a probability value of a value of the target bit corresponding to a state of a plurality of reference bit values existing at a predetermined position with respect to the target bit to be encoded. The image data encoding device according to claim 1. The image data encoding device further includes:
The image data encoding according to any one of claims 1 to 6, further comprising code conversion means for performing code conversion prior to entropy encoding on at least a part of the data divided by the dividing means. apparatus. The code conversion means includes
8. The image data encoding apparatus according to claim 7, wherein code conversion from binary data to gray code is performed for each predetermined number of bits. The reading means includes
From multi-valued image data, pixels are read out for each block composed of the same number of pixels in the vertical and horizontal directions,
The entropy encoding means includes
Storage means for storing data to be subjected to entropy encoding in a form expanded on a two-dimensional plane so that the number of bits in the vertical direction and the number of bits in the horizontal direction for each block are equal,
The image data encoding apparatus according to any one of claims 1 to 8, wherein entropy encoding is performed for each line of the storage unit. The gradation characteristic data is
The image data encoding device according to claim 1, comprising an average value of pixel values of a plurality of pixels read by the reading unit and a value of a gradation dynamic range. The gradation characteristic data is
The image data encoding device according to any one of claims 1 to 9, comprising a maximum value and a minimum value of a plurality of pixel values read by the reading unit. The dividing means includes
The average value of the pixel values of the plurality of pixels and the value of the gradation dynamic range, or the maximum value and the minimum value of the pixel values of the plurality of pixels are further divided. Image data encoding device. The reading means includes
A pixel is read out from the multi-value image data in which one pixel is represented by N bits for each block of (p * q) pixels of vertical p pixels * horizontal q pixels,
The fixed length encoding means includes
From the (p * q) pixels read out by the reading means, the gradation characteristic data consisting of (2 * N) bits and the pixel values of the (p * q) pixels are set to n bits ( n <N) and quantized quantized data are generated,
The dividing means includes
The data generated by the fixed-length encoding means is obtained from the gradation characteristic data and (p * q) bits of the m-th bit (1 ≦ m ≦ n) of the quantized data corresponding to each pixel. Divided into n sets of bit data,
The entropy encoding means includes
The image data encoding device according to any one of claims 1 to 11, wherein the gradation characteristic data and the n sets of bit data are entropy-encoded in a format developed on a two-dimensional plane. . The reading means includes
From multi-valued image data, pixels are read out for each block consisting of 16 pixels of 4 vertical pixels x 4 horizontal pixels,
The fixed length encoding means includes
From the 16 pixels read by the reading means, the gradation characteristic data consisting of 16 bits and quantized data obtained by quantizing the pixel values of the 16 pixels into 2 bits with a predetermined threshold value are generated,
The dividing means includes
The data generated by the fixed-length encoding means includes the gradation characteristic data, the first bit data composed of the upper 16 bits of the quantized data corresponding to each pixel, and the quantum corresponding to each pixel. Divided into the second bit data consisting of the lower 16 bits of the digitized data,
The entropy encoding means includes
The gradation characteristic data, the first bit data, and the second bit data are two-dimensional planes so that the number of bits in the vertical direction and the number of bits in the horizontal direction for each block are equal. The image data encoding device according to any one of claims 1 to 11, wherein entropy encoding is performed in a form expanded above. The image data encoding device further includes:
Compression rate acquisition means for acquiring a compression rate of encoding by the entropy encoding means for bit data corresponding to the most significant bit among the quantized data divided by the dividing means,
Comparison means for comparing the compression rate acquired by the compression rate acquisition means with a predetermined threshold;
Determination means for determining whether other bit data of the divided quantized data is necessary for decoding the encoded multi-valued image data based on the result of comparison by the comparison means. The image data encoding device according to claim 1. The entropy encoding means includes
16. The image data encoding device according to claim 15, wherein when the determination unit determines that the other bit data is not necessary, the other bit data is not encoded. The image data encoding device includes:
Encoded data storage means for storing data encoded by the entropy encoding means;
Encoded data for deleting the encoded data corresponding to the other bit data stored in the encoded data storage unit when the determining unit determines that the other bit data is not necessary The image data encoding device according to claim 15, further comprising: a deleting unit. The entropy encoding means includes
A format in which the gradation characteristic data and the n sets of bit data divided by the dividing unit are expanded on a two-dimensional plane so that the number of bits in the vertical direction and the number of bits in the horizontal direction are equal to each other. Entropy encoding with
  The image data encoding apparatus according to claim 1. The entropy encoding means includes
The gradation characteristic data and the n sets of bit data divided by the dividing unit are entropy-encoded by the same encoding method.
The image data encoding apparatus according to claim 1 . The entropy encoding means includes
  JBIG (Joint Bilevel Image Group) Entropy coding
  19. The image data encoding device according to claim 1 or 18, wherein the image data encoding device is an image data encoding device. An image data encoding method for encoding multi-value image data in which a pixel value of one pixel is represented by a plurality of bits,
A readout step of reading out the multi-value image data for each predetermined number of pixels;
Based on the read pixel values of the plurality of pixels, the fixed-length gradation characteristic data representing the gradation characteristics of the plurality of pixels and the pixel values of the plurality of pixels are converted to n bits of a fixed length by a predetermined threshold value. A fixed-length encoding step for generating quantized quantized data;
A dividing step of dividing the generated data into the gradation characteristic data and n sets of bit data including the m-th bit (1 ≦ m ≦ n) of the quantized data corresponding to each pixel. When,
An entropy encoding step for entropy encoding the divided gradation characteristic data and the n sets of bit data in a form expanded on a two-dimensional plane, respectively . A computer-readable recording medium storing a program for realizing image data encoding processing for encoding multi-value image data in which a pixel value of one pixel is represented by a plurality of bits,
When the program recorded on the recording medium operates on the computer, or the program recorded on the recording medium operates on the computer together with other general-purpose programs,
A readout process for reading out multi-valued image data for each predetermined number of pixels;
Based on the read pixel values of the plurality of pixels, the fixed-length gradation characteristic data representing the gradation characteristics of the plurality of pixels and the pixel values of the plurality of pixels are converted to n bits of a fixed length by a predetermined threshold value. A fixed-length encoding process for generating quantized quantized data;
Division processing for dividing the generated data into the gradation characteristic data and n sets of bit data including the m-th bit (1 ≦ m ≦ n) of the quantized data corresponding to each pixel When,
A program for causing a computer to realize processing including entropy encoding processing for entropy encoding the divided gradation characteristic data and the n sets of bit data in a format developed on a two-dimensional plane. A computer-readable recording medium. A recording medium on which data obtained by encoding multi-value image data in which a pixel value of one pixel is represented by a plurality of bits is recorded,
The encoded data is
Multi-valued image data before encoding is read for each predetermined number of pixels, fixed-length gradation characteristic data representing gradation characteristics of the plurality of pixels from the pixel values of the plurality of pixels read, and the plurality And generating the quantized data obtained by quantizing the pixel value of each pixel to a fixed length of n bits with a predetermined threshold, and then generating the gradation characteristic data and each mth bit of the quantized data corresponding to each pixel It is divided into n sets of bit data composed of the first (1 ≦ m ≦ n) bits, and the divided gradation characteristic data and the n sets of bit data are respectively developed on a two-dimensional plane. A recording medium encoded by entropy encoding. The data encoded by the image data encoding device according to any one of claims 1 to 20, the data encoded by the image data encoding method according to claim 21 , and the computer-readable data according to claim 22 24. Decoding means for decoding at least one of data encoded in a computer in which a program recorded in a recording medium is operable, and data recorded in the recording medium according to claim 23 ,
An image forming apparatus comprising: an image forming unit that forms an image using the image data decoded by the decoding unit.

Images