JP4971881B2 - Image encoding apparatus, image decoding apparatus, and control method thereof - Google Patents

Description

  The present invention relates to an image encoding / decoding technique.

Conventionally, a compression technique using orthogonal transformation is generally used to execute an encoding process on a divided block image. Encoding using orthogonal transform is an effective compression method for natural images, but on the other hand, for images that include characters and line drawings, mosquito noise is caused by the loss of high-frequency components of the image. It was known that block distortion occurred and good results could not be obtained. In order to solve such problems, an image whose resolution is more important than gradation, such as an image with characters and line drawings, tries to maintain the resolution by approximating the pixels in the block image with two colors. Two-color processing has been proposed (for example, Patent Document 1). In Patent Document 1, an image in a block is regarded as an image composed of two colors, data representing the two colors and identification data for identifying which color each pixel corresponds to are generated. A technique is disclosed in which the data is packed into one and output as encoded data.
Japanese Patent Laid-Open No. 05-056282

  By the way, in the case of a character line drawing, the resolution is maintained by dividing the image into two colors, the background color and the color indicated by the character line drawing, and generating identification information indicating the two color data and each pixel. The encoded data can be generated. Assuming that one pixel is 8 bits and encoding is performed in block units of m × n pixels, the encoded data has a data amount of 16 bits representing two color data + m × n bits representing identification information It becomes.

  The present invention intends to provide a technique for further reducing the amount of encoded data for the entire image by encoding one block of encoded data to a smaller number of bits.

  In addition to the above-described effects, another invention provides the size of one block of encoded data as an integer multiple of 8 bits, that is, a 1-byte integer convenient for access by an electronic device or a computer program. A technique for generating double encoded data is also proposed.

In order to solve the above problems, for example, an image encoding device of the present invention has the following configuration. That is,
An image encoding device for encoding multi-value image data in which one pixel is represented by a plurality of bits,
An input means for inputting image data in units of blocks composed of a plurality of pixels;
An average value AVE of the pixel values in the block of interest is calculated, a pixel group having a value equal to or less than the calculated average value AVE is a first group, and a pixel group having a value larger than the average value AVE is a second group. Extraction means for extracting an average value of the pixel values of the first group and the second group as representative pixel values C0 and C1 of the block of interest ;
Identification information generating means for generating binary identification information for indicating which of the representative pixel values C0 and C1 each pixel in the block of interest belongs to;
Generating means for generating data of C0 ′ and C1 ′ excluding the least significant bit of one of the preset pixel values of the representative pixel values C0 and C1 and the least significant bit of each of the representative pixel values C0 and C1; ,
An output unit that generates and outputs encoded data of the block of interest by combining the least significant bit of the one pixel value, the C0 ′, C1 ′, and the identification information;
The output means determines the order of the C0 ′ and C1 ′ to be arranged in the encoded data according to the value of the least significant bit of the other representative pixel value of the representative pixel values C0 and C1, and C0 ′ and C1 ′. Or C1 ′ or C0 ′.

  According to the present invention, it is possible to further reduce the amount of encoded data for the entire image by encoding the encoded data of one block to a smaller number of bits.

  According to another aspect of the invention, in addition to the above-described effects, the size of one block of encoded data is an integer multiple of 8 bits, that is, one byte convenient for access by an electronic device or a computer program. It becomes possible to generate encoded data of an integer multiple of.

  Hereinafter, embodiments according to the present invention will be described in detail with reference to the accompanying drawings.

  In the embodiment, an example will be described in which a block composed of 8m × 8n pixels (m and n are integers of 1 or more) in multi-valued image data in which one pixel is represented by a plurality of bits is encoded. In the following description, it is assumed that one pixel is a gray scale of 8 bits (256 gradations). In the embodiment, encoded data for one block is generated as fixed-length L bits (L = 16 + 8m × 8n). Note that a color image has a plurality of color components. In the case of a color image, the above-described fixed length encoded data may be generated for each component.

  FIG. 1 is a block diagram of an image encoding apparatus according to the embodiment. This apparatus includes a blocking unit 101, a color extracting unit 102, a two-color encoding unit 103, a multi-level encoding unit 104, and a selector 105.

  The blocking unit 101 inputs partial image data (hereinafter referred to as block data) composed of 8m × 8n pixels from the image data to be encoded. The generation source of the image data to be encoded is described as being an image scanner, but it may be a storage device storing uncompressed image data, and the type thereof is not limited. In order to simplify the description, the following description will be made assuming that the block size is 8 × 8 pixels (m = n = 1). Accordingly, the fixed length of the encoded data of one block in the following description is 16 + 8 × 8 = 80 bits. Although details will be made clear from the following description, the first 1 bit of fixed length 80 bits is data indicating the type of encoding. Therefore, the remaining 79 bits are a capacity for storing one block of encoded data.

  The two-color extraction unit 102 extracts two representative colors (or representative values) C0 and C1 from the input block data. Specifically, the two-color extraction unit 102 calculates an average value AVE of the values of each pixel in the block. Then, the pixel group (first group) having a pixel value equal to or less than the calculated average value AVE and the pixel group (second group) having a pixel value larger than the average value are classified. Then, the two-color extraction unit 102 calculates an average value of each pixel belonging to the first group and determines it as the representative color C0. Similarly, the two-color extraction unit 102 calculates the average value of each pixel belonging to the second group and determines it as the representative color C1. Then, the two-color extraction unit 102 outputs the determined representative colors C0 and C1 and the average value AVE to the two-color encoding unit 103.

  Further, the two-color extraction unit 102 compares the difference D (= C1−C0) between the representative colors C0 and C1 with a preset threshold Th (Th> 1), and the comparison result is a determination signal (1 bit). Are output to the two-color encoding unit 103, the multi-level encoding unit 104, and the selector unit 105. Here, the determination signal is “1” when D> Th, and the determination signal is “0” when D ≦ Th.

  The two-color encoding unit 103 performs encoding processing when the determination signal is “1”. On the other hand, the multi-level encoding unit 104 performs encoding processing when the determination signal is “0”. The selector 105 selects and outputs the encoded data generated by the two-color encoding unit 103 when the determination signal is “1”, and multilevel code when the determination signal is “0”. The encoded data generated by the encoding unit 104 is selected and output. Since selector 105 selects one of the encoded data in accordance with the determination signal, the two encoding units may perform encoding processing regardless of the determination signal.

  The multi-level encoding unit 104 uses a known irreversible JPEG encoding. That is, multilevel encoding section 104 performs DCT transformation, quantization, and entropy encoding (encoding is performed in the order of a DC component and an AC component).

  In the embodiment, an example has been described in which encoded data for one block is set to a fixed length of “80” bits. In this embodiment, since there are two types of encoding, it is necessary to indicate which type of encoded data the encoded data is. Therefore, the multi-level encoding unit 104 adds encoding type information (1 bit) indicating that the generated encoded data is generated by the multi-level encoding unit 104, followed by a 79-bit code. Output data.

  Here, it should be noted that the selector 105 selects the encoded data generated by the multi-level encoding unit 104 when the determination signal is “0”. That is, this is a case where the difference D between the representative colors C0 and C1 of each pixel in the block of interest is equal to or smaller than a threshold value (D ≦ Th). D ≦ Th indicates that a high-frequency component such as a character / line drawing is not included or small in the target block. Therefore, each conversion coefficient of the alternating current (AC) component by DCT conversion becomes a small value, and it means that most DCT conversion coefficients are quantized to “0” by the quantization processing. Therefore, it can be said that the amount of encoded data from the multilevel encoding unit 104 is sufficiently small.

  The actual amount of encoded data varies according to the threshold value Th. However, if the threshold value Th is set to a certain value or less, when encoded data from the multi-level encoding unit 104 is selected, the data amount of the encoded data may be within a fixed length. Is possible. It should be noted that the number of bits of the encoded data is not known unless the encoding process is actually performed. Therefore, when the number of bits of the generated encoded data is less than 79 bits, the multi-level encoding unit 104 adds dummy bits subsequent to the encoded data.

  If the encoded data from the multi-level encoding unit 104 exceeds 79 bits, the multi-level encoding unit 104 outputs the first 79 bits as the encoding result. Data after 79 bits is discarded, but such a situation is rare in the first place, and even if it is, it is data of a high frequency component, so the influence on the image quality is small.

  Next, the two-color encoding unit 103 in the embodiment will be described. FIG. 2 is a block configuration diagram of the two-color encoding unit 103. The two-color encoding unit 103 includes a comparison unit 201 and a pack unit 202.

  The comparison unit 201 performs raster scan on the block data from the blocking unit 101 and inputs it in units of pixels. Then, the comparison unit 201 compares the input target pixel value with the average value AVE from the two-color extraction unit 102. When the target pixel value is equal to or less than the average value, the comparison unit 201 sets the target pixel value to the average value. When the value is larger than the value, a binary signal of “1” is output to the pack unit 202. In short, the comparison unit 201 generates identification information (binary) for identifying whether each pixel belongs to the first group or the second group in the two-color extraction unit 102 shown above. Function as. Hereinafter, the data generated by the comparison unit 201 is referred to as identification information. In the case of the embodiment, since one block is 8 × 8 pixels, the identification information is 8 × 8 = 64 bits in total.

  The pack unit 202 combines the representative colors C0 and C1 and the identification information and outputs them. However, since the representative colors C0 and C1 are both 8 bits, if they are simply combined, 8 × 2 + 64 = 80 bits, which cannot be contained in 79 bits excluding 1 bit indicating the coding type. .

Therefore, the pack unit 202 in this embodiment is configured to represent the two colors C0 and C1 with 15 bits as described below. Note that, as already described, there is a relationship of C0 <C1. Further, C0 ′ represents a value represented by bits 1 to 7 excluding the least significant bit (bit 0) of the representative color C0, and C1 represents a value represented by bits 1 to 7 excluding the least significant bit (bit 0) of the representative color C1. ' C0 ′ (same for C1 ′) can be easily generated by shifting C0 by 1 bit in the lower direction and then using the bits 0 to 6. That is, assuming that “+” represents bit combination, the data arrangement is as follows.
When bit 0 of representative color C1 is “0”:
Bit 0 + C0 '+ C1' of representative color C0
When bit 0 of representative color C1 is “1”:
Bit 0 + C1 '+ C0' of representative color C0

  FIGS. 3A and 3B show the data structure of encoded data of the two-color encoding unit 103 including encoding type information (1 bit) and identification information (64 bits). Although details of the decoding process will be described later, 8-bit data of each of the representative colors C0 and C1 can be restored from this data. If the first value of two representative colors represented by 7 bits in the encoded data is data V0 and the next data V1, it is assumed that V0 <V1. In this case, the representative colors in the encoded data are in the order of C0 and C1, and are uniquely determined to be the case of FIG. That is, the original 8-bit representative color C1 is restored by shifting the 7-bit representative color C1 to the left (upper) by 1 bit and setting (adding) “0” to the least significant bit (LSB). it can. The representative color C0 can be restored by simply shifting left by one bit and storing the second bit (bit 0 of C0) of the encoded data in the LSB.

  When V0> V1, the representative colors represented by 7 bits in the encoded data are in the order of C1 and C0, and the case of FIG. 3B is uniquely determined. That is, the original 8-bit representative color C1 can be restored by shifting the 7-bit representative color C1 to the left (upper) by 1 bit and setting (adding) 1 to the LSB. The representative color C0 can be restored by simply shifting left by one bit and storing the second bit (bit 0 of C0) of the encoded data in the LSB.

  As described above, two representative colors can be restored according to the size comparison of the data V0 and V1 and the arrangement of the sizes.

  Note that V0 = V1 only when the representative color C0 is an even value (when the least significant bit is 0) and the representative color C1 is larger than C0 by “1” (when the least significant bit is 1). Become. In other words, V0 = V1 can be said only when the representative color C0 is an even value and the representative color C1 is larger by “1” than C0. Since this situation is only when the second bit (bit 0 of C0) of the encoded data is “0” and V0 = V1, V0 (or V1) is shifted by 1 bit in the upper direction, A device in which 0 is set in the least significant bit can be specified as C0. C1 can be obtained by adding “1” to C0.

  In the embodiment, it is assumed that the data structure of the encoded data generated by the two-color encoding unit 103 is one of FIGS. However, this does not limit the present invention. For example, the order of identification information (64 bits), {C0 ', C1'}, and the least significant bit of one of the representative colors may be used. In short, the data structure may be arranged in common on the encoding device side and the decoding device side.

  As described above, the two-color encoding unit 103 according to the present embodiment packs 16 bits representing two representative colors and 64 bits representing identification information, not 16 + 64 bits but encoded data one bit less than that. Can be generated. Since the resolution of recent image scanners and digital cameras is becoming higher, the number of blocks included in the resolution is inevitably increasing. Therefore, reducing one bit per block means that the amount of encoded data of the entire image is greatly reduced.

  Further, if the encoding apparatus is configured with the configuration excluding the multi-level encoding unit 104 in the embodiment, it is not necessary that one pixel is 8 bits, one block is 8 m × 8 n, and one block is a plurality of pixels. That's fine. Of course, it can be easily understood that the encoding type information is not necessary.

  Further, as in the above embodiment, when the two-color encoding unit 103 and the multi-level encoding unit 104 are mixed and image data of 8 bits per pixel is encoded in units of 8 × 8 pixels, one block It becomes possible to generate encoded data having a fixed length of 80 bits. In addition, since one block of data is represented by an integer multiple of 8, that is, an integer multiple of bytes, it is promised that the access efficiency will be high.

  In the above embodiment, a monochrome multi-value image is taken as an example, but it may be applied to a color image. In the case where a color image, for example, three components of R, G, and B are represented by 8 bits each, the processing of the above embodiment may be performed for each component.

  In the embodiment, the encoding type information is described as being generated by each encoding unit. However, the selector 105 may add the encoding type information based on information from the two-color extraction unit. I do not care.

  Next, the image decoding apparatus in the embodiment will be described. FIG. 4 is a block configuration diagram of the image decoding apparatus according to the embodiment. This apparatus includes an encoding type determination unit 301, a color decoding unit 302, a multi-level decoding unit 303, and a selector unit 304.

  The encoding type determination unit 301 determines whether the first bit (encoding type information) of the 80-bit encoded data is “0” or “1”. In other words, the encoding type determination unit 301 determines whether the target encoded data is encoded data generated by the two-color encoding unit 103 or the multi-level encoding unit 104. The encoding type determination unit 301 outputs the determination result as a control signal to the two-color decoding unit 302, the multi-level decoding unit 303, and the selector unit 304.

  The two-color decoding unit 302 assumes that the target encoded data is two-color encoded data when the control signal from the encoding type determination unit 301 indicates that the two-color encoding unit 103 has encoded the control signal. Decrypt.

  The two-color decoding unit 302 determines whether the input encoded data is one of FIGS. 3A and 3B. That is, the values of two representative colors represented by 7 bits are compared to determine which type is shown in FIGS. 3A and 3B, and the 8-bit representative colors C0 and C1 are restored. The restoration algorithm is as described above. Then, it is assumed that the subsequent 64-bit identification information is arranged in the raster scan order, and the representative color C0 is output as the pixel value for each pixel position where the identification information is “0”. Then, the representative color C1 is output as the pixel value for the pixel position whose identification information is "1". By performing this for 64 bits, image data of 8 × 8 pixels is restored.

  On the other hand, when the multilevel decoding unit 303 indicates that the control signal from the encoding type determination unit 301 is encoded by the multilevel encoding unit 104, the encoded data of interest is the multilevel encoded data. JPEG decoding as a thing. The decoding process is performed in the order of entropy decoding, inverse quantization, and inverse DCT transform. At this time, if the EOB (End Of Block) is detected as a result of entropy decoding of 79 bits excluding the encoding type information in order, the data after the EOB is ignored as a dummy. If EOB cannot be detected by the entropy decoding process, the decoding process is continued assuming that the insufficient number of DCT coefficients after quantization is “0”.

  The selector unit 304 selects 8 × 8 pixel data from the two-color decoding unit 302 when the control signal from the coding type determination unit 301 indicates that the two-color coding unit 103 has encoded the control signal. And output. On the other hand, when the control signal from the encoding type determination unit 301 indicates that the multilevel encoding unit 104 has encoded, the 8 × 8 pixel data from the multilevel decoding unit 303 is selected and output. To do. By repeating this, image restoration (decoding) is performed.

As described above, the two-color decoding unit of this embodiment inputs 15 + 64 bits as a unit,
From there, two representative colors C0 and C1 are determined, and the encoded data of one block of 8 × 8 pixels can be restored.

  Note that in the case of a decoding apparatus that includes a two-color decoding unit alone, the encoded data of one block may be 79 bits. That is, in general terms, when one pixel is represented by M bits and the encoded data of an m × n pixel block, the data amount (data length) of the encoded data of one block is 2M−1 + m × n bits. It becomes.

  Further, according to the present embodiment, it is possible to decode encoded data in which multi-level encoded data and 2-color encoded data generated by the encoding apparatus described above are mixed.

  If the data length of the encoded data of one block is an integer multiple of 8, the size of one block may be 8n × 8m pixels. Also, let L be the fixed code length when the size of one block is 8n × 8m. The multi-level encoding unit 104 performs JPEG encoding m × n times for 8 × 8 pixels, assuming that there are n × m sub-blocks of 8 × 8 pixels in the block. At this time, the JPEG encoded data of one subblock has a fixed length of L / (m × n) −1, and the JPEG encoded data of other subblocks has a fixed length of L / (m × n). You can do it.

  In the embodiment, bit 0 of color C0 is stored after the encoding type information, but bit 0 of color C1 may be stored. In this case, the order of the colors C0 and C1 may be determined according to the value of bit 0 of the color C0.

<Description of modification>
An example in which the same processing as in the above embodiment is realized by a computer program will be described below. The computer program may be, for example, a personal computer, and its hardware resources may be constituted by a CPU, ROM, RAM, HDD, display device (display control unit), and image input device (image scanner), and the description thereof is unnecessary. Will. Therefore, hereinafter, a processing procedure by an application program for encoding and decoding will be described with reference to the drawings.

  First, an encoding process procedure in an application program executed by the CPU will be described with reference to the flowchart of FIG. The application program is stored in a storage medium such as a hard disk device, and is loaded into the RAM and executed under the control of the OS (Operating System). The generation source of the image data to be encoded is not particularly limited, but will be described as an image scanner, and the encoding result will be described as being stored as a file on the hard disk.

  First, in step S1, one block (8 × 8 pixels) of image data to be encoded is input. Next, in step S2, an average value AVE of the pixels in the input block is calculated. Then, the average value of the pixel group having a value equal to or less than the average value AVE is calculated as the representative color C0, and the average value of the pixel group having a value exceeding the average value AVE is calculated as the representative color C1. At this time, if all the pixels in the block have the same pixel value, C0 = C1.

  In step S3, it is determined whether “C1-C0” is greater than a threshold value Th. If it is determined that “C1-C0” is equal to or smaller than the threshold Th, that is, if it is determined that the high frequency component is small in the block of interest, the process proceeds to step S4. In step S4, multi-level encoding (JPEG encoding) is performed, 1-bit data indicating that multi-level encoding has been performed is added to the head, and subsequent 79-bit encoded data is generated. Thereafter, in step S5, the generated encoded data with a fixed length of 80 bits (multi-level encoded data) is output.

  On the other hand, in step S3, when it is determined that “C1-C0” is greater than the threshold value Th, it can be determined that the target block includes a character line drawing. Then, it progresses to step S6 and a two-color encoding process is performed. That is, 1-bit data indicating that the two-color encoding is performed, bit 0 of color C0, and upper 7 bits of colors C0 and C1 are generated in the order corresponding to the value of bit 0 of color C1. Then, the average value AVE calculated in the previous step S2 is compared with each pixel value in the block, and identification information indicating which of the two groups each pixel belongs to is generated. And these data are put together and the data of Fig.3 (a) or FIG.3 (b) is constructed | assembled.

  Thereafter, in step S7, the constructed data is output.

  When the process of step S5 or step S7 is completed, the process proceeds to step S8, and it is determined whether or not the encoding process for all blocks has been completed. If it is determined that the process is incomplete, the processes after step S1 are repeated.

  Thus, when it is determined in step S8 that the encoding process for all blocks has been completed, the encoding process is terminated.

  Next, the decoding process in the application program executed by the CPU will be described. Here, an example of decoding and displaying a compression-encoded data file (assuming that the user has already selected via the GUI) stored in the hard disk will be described. Since the decoding result is displayed, the output destination of the decoded image data is a video memory included in the display control unit.

  First, in step S11, one block of encoded data (fixed length 80 bits) is input from the encoded data file into the RAM. Next, in step S12, it is determined whether the leading 1-bit encoding type information is “0” or “1”. That is, it is determined whether the input encoded data is two-color encoded data or multi-level encoded data.

  If it is determined that the data is multi-level encoded data, the process proceeds to step S13 to perform JPEG decoding. This JPEG decoding process is performed in the order of known entropy decoding, inverse quantization, and inverse DCT transform. However, if EOB can be detected as a result of entropy decoding of 79 bits excluding the encoding type information in order, the data after EOB is ignored as a dummy. If EOB cannot be detected by the entropy decoding process, the decoding process is continued assuming that the insufficient number of DCT coefficients after quantization is “0”. When the 8 × 8 pixel block image data is decoded in this way, the block image data is output in step S14.

  On the other hand, if it is determined that the input encoded data is two-color encoded data, the process proceeds to step S15. In step S15, it is determined whether the input encoded data is one of FIGS. 3A and 3B. That is, the values of two representative colors represented by 7 bits are compared to determine which type is shown in FIGS. 3A and 3B, and the 8-bit representative colors C0 and C1 are restored. Then, it is assumed that the subsequent 64-bit identification information is arranged in the raster scan order, and the representative color C0 is replaced with the pixel value for the pixel position where the individual identification information is “0”. The representative color C1 is replaced with the pixel value for the pixel position whose identification information is “1”. By performing this for 64 bits, image data of 8 × 8 pixels is restored. Thereafter, in step S16, the restored 8 × 8 pixel image data is output.

  When the process of step S14 or step S16 is completed, the process proceeds to step S17, and it is determined whether or not the decoding process for all blocks has been completed. If it is determined that the process is incomplete, the processes after step S11 are repeated.

  As described above, the encoding and decoding processes described above can be realized by a computer program, and the same operational effects can be achieved.

<Second Embodiment>
In the first embodiment and the modification thereof, the criterion for selecting one of the two encoded data is based on whether or not the difference between the representative colors C0 and C1 is equal to or less than the threshold Th. However, this does not limit the present invention.

  For example, an error between the result of decoding the two-color encoded data and the original 8 × 8 pixel image data, and an error between the result of decoding the multi-level encoded data and the original 8 × 8 pixel image data Ask for. Then, the smaller of the two errors may be determined as encoded data to be output. More detailed description is as follows.

Each pixel value X (i, j) (i, j = 0, 1, 2,... 7) of 8 × 8 pixels of the original (before encoding) is defined. Further, each pixel value of 8 × 8 pixels obtained by decoding the two-color encoded data is Y (i, j), and each pixel value of 8 × 8 pixels obtained by decoding the multi-value encoded data is Z (i , J). Then, selector section 105 calculates error D1 and error D2 according to the following equations.
D1 = Σ | X (i, j) −Y (i, j) |
D2 = Σ | X (i, j) −Z (i, j) |
Here, | p | represents the absolute value of the value p, and Σ represents the sum of i, j = 0, 1,. Further, instead of obtaining the absolute value of the difference between the pixels, the sum of squares of the difference between the pixels may be obtained as in the following equation.
D1 = Σ {X (i, j) −Y (i, j)} ^ 2
D2 = Σ {X (i, j) −Z (i, j)} ^ 2
(P ^ 2 indicates the square of P)

  In any case, the selector unit 105 selects two-color encoded data when D1 ≦ D2 is satisfied, and selects multi-level encoded data when D1> D2.

  Further, when all the pixel values in the block of interest are the same, the two-color extraction unit 102 sets one of the two representative colors C0 and C1 as an actual color. The other is dummy data. Then, the median value may be output to the two-color encoding unit 103 as AVE. An example is as follows.

  Now, it is assumed that the values of all the pixels in the block of interest are “255”. In this case, the two-color extraction unit 102 outputs C0 = 255 (existing color), C1 = 0 (dummy data), and the median “128” to the two-color encoding unit 103 as AVE. The two-color encoding unit 103 only determines the identification information of all the pixels as “1” by the configuration of FIG. 3, and no contradiction occurs.

  As described above, according to the second embodiment, in addition to the operational effects of the first embodiment, it is possible to generate encoded data capable of decoding an image that approximates the original image more accurately. Become.

  Note that the decoding device in the second embodiment is the same as that in the first embodiment, and a description thereof will be omitted. Further, it is obvious that the function corresponding to the second embodiment can be realized by a computer program as in the modification of the first embodiment described above.

  Usually, the computer program is stored in a computer-readable storage medium such as a CD-ROM, and is set in a reading device (such as a CD-ROM drive) included in the computer, and can be executed by copying or installing in the system. Therefore, it is obvious that such a computer-readable storage medium is also within the scope of the present invention.

It is a block block diagram of the image coding apparatus in embodiment. It is a block block diagram of the two-color encoding part in FIG. It is a figure which shows the data structure of the encoding data which a 2 color encoding part produces | generates. It is a block block diagram of the image decoding apparatus in embodiment. It is a flowchart which shows an encoding process procedure. It is a flowchart which shows a decoding process procedure.

Claims (11)

An image encoding device for encoding multi-value image data in which one pixel is represented by a plurality of bits,
An input means for inputting image data in units of blocks composed of a plurality of pixels;
An average value AVE of the pixel values in the block of interest is calculated, a pixel group having a value equal to or less than the calculated average value AVE is a first group, and a pixel group having a value larger than the average value AVE is a second group. Extraction means for extracting an average value of the pixel values of the first group and the second group as representative pixel values C0 and C1 of the block of interest ;
Identification information generating means for generating binary identification information for indicating which of the representative pixel values C0 and C1 each pixel in the block of interest belongs to;
Generating means for generating data of C0 ′ and C1 ′ excluding the least significant bit of one of the preset pixel values of the representative pixel values C0 and C1 and the least significant bit of each of the representative pixel values C0 and C1; ,
An output unit that generates and outputs encoded data of the block of interest by combining the least significant bit of the one pixel value, the C0 ′, C1 ′, and the identification information;
The output means determines the order of the C0 ′ and C1 ′ to be arranged in the encoded data according to the value of the least significant bit of the other representative pixel value of the representative pixel values C0 and C1, and C0 ′ and C1 ′. Or an image encoding apparatus that outputs the data as either C1 ′ or C0 ′. When the pixel values of all the pixels in the block of interest are the same, the extraction unit determines the pixel value as one representative pixel value, and sets the other representative pixel value as a dummy different from the one representative pixel value. The image encoding apparatus according to claim 1 , wherein the value of the image is extracted. The block is composed of 8m × 8n (m, n is an integer of 1 or more) pixels, and when the number of bits of encoded data output by the output unit is defined as L-1 bits,
Furthermore,
JPEG encoding is performed m × n times for each 8 × 8 pixel sub-block in the block of interest to generate JPEG encoded data, and the JPEG encoded data of one sub-block is converted to L / (m × n) JPEG for generating L-1 bit JPEG encoded data from the block of interest by setting the JPEG encoded data of the other sub-blocks to a fixed length of L / (m × n) bits. Encoding means;
1 indicating whether the selected encoded data is one of the encoded data output from the output means and the JPEG encoded data generated by the JPEG encoding means. and subsequent to the bits of the encoded type information, the image encoding apparatus according to claim 1 or 2, characterized in that it comprises a selection means for outputting the selected coded data. The selection means includes
The difference between the representative pixel values C1 and C0 is compared with a preset threshold value. If the difference is less than or equal to the threshold value, the encoded data generated by the JPEG encoding means is selected, and 4. The image encoding apparatus according to claim 3 , wherein when the difference is larger than the threshold value, the encoded data output by the output unit is selected. The selection means includes
Calculating an error D1 that is the sum of pixel value differences between each pixel of the image obtained by decoding the encoded data generated by the JPEG encoding means and the image of the target block before encoding;
Calculating an error D2 that is the sum of the pixel value differences between each pixel of the image obtained by decoding the encoded data output from the output means and the image of the block of interest before encoding;
D2 ≦ D1
When the relationship is, select the encoded data output by the output means,
The image encoding device according to claim 3 , wherein when the relationship of D2> D1 is satisfied, the encoded data generated by the JPEG encoding unit is selected. A control method of an image encoding device for encoding multi-value image data in which one pixel is represented by a plurality of bits,
An input step in which the input means inputs image data in units of blocks composed of a plurality of pixels;
The extraction means calculates an average value AVE of the pixel values in the block of interest, a first group of pixels having a value equal to or less than the calculated average value AVE, and a first group of pixels having a value greater than the average value AVE. An extraction step of extracting an average value of the pixel values of the first group and the second group as representative pixel values C0 and C1 of the block of interest when two groups are provided ;
An identification information generating step in which identification information generating means generates binary identification information for indicating which of the representative pixel values C0 and C1 each pixel in the block of interest belongs to;
The generating means obtains the data of C0 ′ and C1 ′ excluding the least significant bit of one representative pixel value set in advance of the representative pixel values C0 and C1 and the least significant bit of each of the representative pixel values C0 and C1. A generation process to generate;
An output unit comprising: an output step of generating and outputting encoded data of the block of interest by combining the least significant bit of the one pixel value, the C0 ′, C1 ′, and the identification information;
In the output step, the order of the C0 ′ and C1 ′ arranged in the encoded data is changed to C0 ′ and C1 ′ according to the value of the least significant bit of the other representative pixel value of the representative pixel values C0 and C1. Or a control method for an image encoding apparatus, wherein the output is performed as either C1 ′ or C0 ′. A computer program for causing a computer to function as each unit according to any one of claims 1 to 5 when the computer reads and executes the computer program. An image decoding device for decoding encoded data generated by the image encoding device according to claim 1,
Input means for inputting encoded data for one block in units;
From the preset position in the input encoded data, 1-bit data and the data V0, V1 excluding the least significant bit of the representative pixel values C0 and C1 of the block and each pixel in the block are Extraction means for extracting identification information indicating which of the representative pixel values C0 and C1 belongs;
By adding the 1-bit data to one least significant bit of the data V0 and V1 determined by comparing the data V0 and V1, one representative pixel value of the representative pixel values C0 and C1 is obtained. Restore,
Representative pixel value restoring means for restoring the other representative pixel value of the representative pixel values C0 and C1 in accordance with the order of arrangement of the values of the data V0 and V1;
An image for restoring an image of a block composed of a plurality of pixels by selecting and outputting one of the representative pixel values C0 and C1 restored by the representative pixel value restoration unit according to the value of the identification information An image decoding apparatus comprising: a restoration unit. A control method for an image decoding device for decoding encoded data generated by the image encoding device according to claim 1, comprising:
An input step in which the input means inputs encoded data for one block in units;
The extraction means removes 1-bit data and the data V0, V1 excluding the least significant bits of the representative pixel values C0 and C1 of the block from the preset position in the input encoded data, An extraction step of extracting identification information indicating which of the representative pixel values C0 and C1 each pixel belongs to;
The restoring means adds the 1-bit data to the least significant bit of one of the data V0 and V1 determined by comparing the data V0 and V1, so that one of the representative pixel values C0 and C1 Restore the representative pixel value,
A representative pixel value restoring step of restoring the other representative pixel value of the representative pixel values C0 and C1 according to the order of arrangement of the values of the data V0 and V1;
The image restoration unit selects and outputs one of the representative pixel values C0 and C1 restored in the representative pixel value restoration step according to the value of the identification information, and thereby outputs a block composed of a plurality of pixels. A method for controlling an image decoding apparatus, comprising: an image restoration step for restoring an image. A computer program that causes the computer to function as each unit according to claim 8 by being read and executed by the computer. A computer-readable storage medium storing the computer program according to claim 7 or 10 .

Images