JP2013255092A - Image data compression coding device, image data compression coding method, and image data compression coding program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image data compression coding device which effectively compresses image data, an image data compression coding method, and an image data compression coding program.SOLUTION: The image data compression coding device includes: threshold determination means for determining a first deviation threshold and a second deviation threshold larger than the first deviation threshold on the basis of the standard deviation in pixel value of each pixel in image data; and coding means which assigns a pixel whose deviation in pixel value is less than the first deviation threshold, a pixel whose deviation is equal to or greater than the first deviation threshold and less than the second deviation threshold, and a pixel whose deviation is equal to or greater than the second deviation threshold to a first division, a second division, and a third division, respectively, for each pixel, and which encodes the pixel value of each of the assigned pixels for each of the first to the third divisions to generate coded data.

Description

  The present invention relates to an image data compression coding apparatus, an image data compression coding method, and an image data compression coding program.

  In recent years, the amount of image processing (higher image quality, enlargement / reduction, object recognition, etc.) per picture has increased with the increase in screen size and resolution. For example, an image display device such as a digital television receives about 30 pieces of image data per minute from a broadcasting station, performs image processing such as image data enlargement processing in accordance with the resolution of the display screen, and displays it on the display screen or the like. Display. However, since the amount of image data after image processing has increased with the increase in resolution of the display screen, it takes time for the transfer processing and the like, and within a predetermined time of image data transmitted from a broadcasting station It may be difficult to process.

  Therefore, in the middle of image processing or when outputting image data after image processing, it is required to reduce the processing speed and bandwidth on the transmission path by compressing the image data. Further, there are advanced standards and techniques of MPEG2 and H.264 regarding pixel data compression processing, but pixel data compression is also required in these compression processing steps as the resolution of video is increased.

  The compression of pixel data is generally performed by variable length coding of the difference value between adjacent pixels using the high degree of correlation with adjacent pixels. However, there is a problem in that compression is hardly effective in a portion where the correlation between adjacent pixels is low. Thus, a technique is disclosed in which image data is divided into a plurality of highly correlated areas, and compression processing is performed for each divided area. Thereby, the difference value between adjacent pixels in the area is reduced, and effective compression is performed (for example, Patent Document 1).

JP-A-7-114499

  However, when the correlation between adjacent pixels is low in the image data, an area with high correlation is not always found. In the case of image data with low correlation between adjacent pixels in this way, even if the image data is divided into a plurality of areas, the difference value between adjacent pixel values becomes large. Thus, effective compression of image data having low correlation between adjacent pixels has been difficult.

  An object of the present invention is to provide an image data compression coding apparatus, an image data compression coding method, and an image data compression coding program that effectively compress image data.

  The first aspect is a threshold value determining means for determining a first deviation threshold value and a second deviation threshold value larger than the first deviation threshold value based on the standard deviation of the pixel value of each pixel in the image data; For each of the pixels, a pixel whose deviation value is less than the first deviation threshold is a first division, and a pixel whose deviation value is greater than or equal to the first deviation threshold and less than the second deviation threshold is the first. 2 divisions, a pixel whose deviation value is equal to or greater than the second deviation threshold is assigned to the third division, and encoded data is generated by encoding the assigned pixel value for each of the first to third divisions. Encoding means.

  According to the first aspect, the image data is effectively compressed.

It is a figure which shows an example of a structure of an image data compression encoding apparatus. It is a figure which shows an example of the image data made into the object of compression encoding. It is a figure explaining the distribution area in the example of this embodiment. It is a figure explaining the deviation threshold value in case a standard deviation is large. It is a figure explaining the deviation threshold value in case a standard deviation is small. It is a figure explaining the detailed structure of a compression encoding part. It is a flowchart figure explaining the allocation process to the distribution area of each pixel. It is a figure which shows the specific example of area A memory, area B memory, and area C memory. It is a flowchart figure explaining the flow of the compression encoding process for every distribution area. It is a figure showing an example of the coding data produced | generated by the compression coding process.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the technical scope of the present invention is not limited to these embodiments, but extends to the matters described in the claims and equivalents thereof.

[Image Data Compression Encoding Device 10]
FIG. 1 is a diagram illustrating an example of the configuration of the image data compression coding apparatus 10. The image data compression encoding apparatus 10 includes, for example, a processor 16, a memory 17, a video input interface 11, an image processing unit 12, a compression encoding unit 13, a decoding unit 14, and a video display unit 15. The processor 16 performs overall control of the image data compression encoding apparatus 10, and the memory 17 stores image data input via the video input interface 11 and image data after image processing.

  In this embodiment, the image data is, for example, RGB image data, YCbCr image data, or the like. However, the present invention is not limited to this example, and other types of image data may be used.

  The image processing unit 12 reads image data input and stored in the memory, and performs image processing such as image quality enhancement, enlargement / reduction, and object recognition. The compression encoding unit 13 performs compression encoding of the read image data, and the decoding unit 14 decodes the compression encoded image data and outputs the decoded image data to the video display unit 15. The video display unit 15 displays the decoded image data on a display screen or the like.

  The image data compression encoding apparatus 10 may perform image processing a plurality of times for image data generated by a handycam camera or the like, for example. In such a case, writing processing and reading processing of image data after image processing to the memory 17 occur multiple times. Even in such a case, the compression-coded image data is written into the memory 17, and the image data read from the memory 17 is decoded and processed.

  As described above, the image data compression encoding apparatus 10 according to the present embodiment compresses and encodes the image data after the image processing and stores it in the memory. As a result, the writing and reading of the image data after the image processing to the memory and the delivery to the video display unit 15 are accelerated.

  Specifically, the image data compression coding apparatus 10 determines a deviation threshold AB and a deviation threshold AC greater than the deviation threshold AB based on the standard deviation of the pixel values of each pixel in the image data. For each pixel, the image data compression encoding apparatus 10 distributes pixels whose deviation value is less than the deviation threshold AB, and distributes pixels whose deviation value is greater than or equal to the deviation threshold AB and less than the deviation threshold AC. A, a pixel having a deviation value equal to or greater than the deviation threshold AC is assigned to the distribution area C, and the pixel value assigned to each distribution area is encoded to generate encoded data.

  That is, the image data compression coding apparatus 10 in the present embodiment assigns each pixel of the image data to three sections (distribution areas A to C) based on the pixel values, and performs compression coding processing for each section. Do. As described above, by assigning each pixel to a section based on the pixel value rather than the pixel arrangement, even if the image data has low correlation between the pixel values of adjacent pixels, the pixels between the pixels in each section The correlation of the values becomes high. Thereby, the difference value of the pixel value between the pixels in each section is reduced and the variation is reduced, so that the image data is effectively compressed and encoded. Details of the processing will be described later.

  Here, an example of image data to be compression-encoded by the image data compression-encoding device 10 in the present embodiment will be described.

[Specific examples of image data]
FIG. 2 is a diagram illustrating an example of image data PI that is an object of compression encoding by the image data compression encoding apparatus 10 according to the present embodiment, and a correspondence relationship PIx between each pixel of the image data and the distribution area. The image data PI in FIG. 6A is pixel data to be compressed and encoded having 16 horizontal pixels and 16 vertical pixels. In this example, the image data PI indicates, for example, R-value image data in the RGB system, and each pixel has a value of 0 to 255.

  FIG. 2B is an example of a correspondence diagram PIx showing to which distribution area each pixel of the image data PI of FIG. In the figure, A is assigned to the distribution area A, B is assigned to the distribution area B, and C is assigned to the distribution area C. Each pixel is assigned to a corresponding distribution area based on a deviation value calculated from the pixel value.

  Next, the distribution area will be described.

[Distribution area]
FIG. 3 is a diagram for explaining the distribution area in the present embodiment. The figure shows an example of a normal distribution diagram L1 of pixel deviation values at a predetermined standard deviation. In the figure, the horizontal axis indicates the deviation value, and the center indicates the deviation value 50. The vertical axis represents the number of pixels. The image data compression coding apparatus 10 according to the present embodiment, as shown in the figure, has three distribution areas (distribution area Aea, distribution area Beb) in which the deviation value of the pixel value is divided by the deviation threshold value AB and the deviation threshold value AC. , Each pixel is assigned to the distribution area Cec). Each distribution area corresponds to a predetermined range of deviation values.

  Specifically, pixels whose deviation value corresponding to the pixel value is less than the deviation threshold AB are assigned to the distribution area Beb. Further, pixels whose deviation value corresponding to the pixel value is greater than or equal to the deviation threshold AB and less than the deviation threshold AC are assigned to the distribution area Aea. Also, pixels whose deviation value corresponding to the pixel value is greater than or equal to the deviation threshold AC are assigned to the distribution area Cec.

[Deviation threshold]
Next, a method for determining the deviation threshold AB and the deviation threshold AC will be described. In the image data compression coding apparatus 10 according to the present embodiment, the deviation threshold AB and the deviation threshold AC are determined based on the standard deviation of the pixel values of the pixels included in the image data.

  Specifically, the deviation threshold AB is determined by subtracting a coefficient value calculated by multiplying the reciprocal of the standard deviation by 100 from 50, and the deviation threshold AC is determined by adding the coefficient value to 50. The reciprocal number is a number that becomes 1 when a certain number is multiplied. That is, the reciprocal of the standard deviation is obtained as a value obtained by dividing 1 by the standard deviation. For this reason, the image data compression coding apparatus 10 calculates the reciprocal by dividing the standard deviation of the pixel value of each pixel of the image data from 1, and calculates the coefficient value by multiplying by 100. Then, the image data compression coding apparatus 10 determines the number obtained by subtracting 50 from the coefficient value as the deviation threshold AB, and the number obtained by adding 50 to the coefficient value as the deviation threshold AC.

  For example, a case where the standard deviation is 10 is taken as an example. When the standard deviation is a value of 10, it is calculated by the arithmetic expression 1 ÷ 10, and the reciprocal of 0.1 is obtained. Then, the coefficient value 10 is obtained by multiplying the reciprocal 0.1 by 100. Thereby, the deviation threshold value AB becomes a value 40 obtained by subtracting the coefficient value 10 from 50. Further, the deviation threshold AC is a value 60 obtained by adding 50 to the coefficient value 10. As a result, for example, pixels having a deviation value less than 40 are allocated to the distribution area Beb. Further, pixels having a deviation value of 40 or more and less than 60 are assigned to the distribution area Aea, and pixels having a deviation value of 60 or more are assigned to the distribution area Cec.

  Here, the difference in the deviation threshold value based on the standard deviation will be described according to a specific example.

[Specific example of deviation threshold: large standard deviation]
FIG. 4 is a diagram for explaining a deviation threshold when the standard deviation of the pixel value of each pixel of the image data is large. The case where the standard deviation is large indicates a case where the variation of the pixel value of each pixel is large. This figure shows an example of a normal distribution diagram L2 of pixel deviation values when the standard deviation is large.

  As shown in FIG. 4, when the standard deviation is large, the pixel deviation values are not distributed in the vicinity of the deviation value 50, but are distributed over a wide deviation value range. When the pixel distribution is dispersed, if the difference value of each pixel is calculated from the average value of the pixel values of all the pixels, the difference value increases and the variation of each difference value increases. For this reason, when compression and variable encoding are performed based on the difference value, compression is not effectively performed.

  In the present embodiment, by using the reciprocal of the standard deviation, when the standard deviation is large, the section of the distribution area Aea centered on the standard deviation 50 is narrowed. That is, when the standard deviation is large, the reciprocal of the standard deviation is a small value, and the coefficient value is also a small value. Therefore, the deviation threshold AB is a large value and the deviation threshold AC is a small value compared to the case where the standard deviation is small, the range of the deviation value corresponding to the distribution area Aea is narrowed, and the distribution thresholds B and C are corresponded. The deviation value range is widened.

  As a result, even if the image data has a large variation in pixel values and the distribution is distributed, the deviation value range of the distribution areas B and C is widened so that the difference value of the pixel values between the pixels assigned to each distribution area is As it becomes smaller, the variation of the difference value becomes smaller. For this reason, the difference value between the average value of the pixel values of the pixels assigned to each distribution area and each pixel becomes small. Thereby, compression and encoding processing is effectively performed based on the difference value, and the compression rate of the image data is improved.

[Specific example of deviation threshold: small standard deviation]
FIG. 5 is a diagram for explaining the deviation threshold when the standard deviation of the pixel value of each pixel of the image data is small. The case where the standard deviation is small indicates a case where there is little variation in the pixel value of each pixel. This figure shows an example of a normal distribution diagram L3 of pixel deviation values when the standard deviation is small.

  As shown in FIG. 5, when the standard deviation is small, pixel deviation values are concentrated and distributed near the deviation value 50. When the distribution of pixel values is concentrated, the pixel group where the distribution is concentrated is assigned to the same distribution area, so that the difference value of the pixel values between the pixels in the distribution area can be kept small.

  For this reason, in this embodiment, by using the reciprocal of the standard deviation, when the standard deviation is small, the section of the distribution area Aea centered on the standard deviation 50 is widened. That is, when the standard deviation is small, the reciprocal of the standard deviation is a large value, and the coefficient value is also a large value. For this reason, as compared with the case where the standard deviation is large, the deviation threshold AB is a small value and the deviation threshold AC is a large value, and the range of the deviation value corresponding to the distribution area Aea is widened.

  Thereby, a pixel group in which the distribution is concentrated is assigned to one distribution area Aea, the difference value of the pixel value between the pixels assigned to the distribution area Aea is reduced, and the variation of the difference value is reduced. For this reason, the difference value between the average value of the pixel values of the pixels assigned to each distribution area and each pixel becomes small. Thereby, compression and encoding processing is effectively performed based on the difference value, and the compression rate of the image data is improved.

  Next, the configuration of the compression encoding unit 13 in the image data compression encoding apparatus 10 according to the present embodiment will be described.

[Details of configuration of code compression unit]
FIG. 6 is a diagram illustrating a detailed configuration of the compression encoding unit 13 in the image data compression encoding apparatus 10 according to the present embodiment. In the figure, the compression encoding unit 13 includes, for example, a deviation value calculating unit 131, a standard deviation reciprocal calculating unit 132, a distribution boundary determining unit 133, a distribution area determining unit 134, a distribution area A encoding unit 135, and a distribution area B code. And a distribution area C encoding unit 137, a result memory 138, and an area offset calculation unit 139. The external memory 103 corresponds to the memory 17 in the image data compression coding apparatus 10 of FIG.

  In FIG. 6, when image data is stored in the memory 17, the buffering unit 102 inputs the image data, and the cumulative addition unit 100 cumulatively adds the pixel values of each pixel in the image data. The standard deviation calculation unit 101 obtains an average value of pixel values from the cumulative addition value, and calculates a standard deviation. The deviation value calculation unit 131 calculates a deviation value of each pixel based on the obtained standard deviation, average value, and pixel value of each pixel. The standard deviation reciprocal computing unit 132 divides 1 by the standard deviation to obtain the reciprocal of the standard deviation. Further, the distribution boundary determination unit 133 subtracts the coefficient value obtained by multiplying the reciprocal of the standard deviation by 100 from the value 50 and adds the deviation threshold value AB, and adds the coefficient value to the value 50 to determine the deviation threshold value AC.

  The distribution area determination unit 134 performs allocation determination processing for each pixel, based on the pixel deviation value, for allocation to the three distribution areas divided by the pixel deviation threshold value AB and the deviation threshold value AC. In this example, each pixel of the image data is assigned to any of the distribution area Aea, distribution area Beb, and distribution area Cec described above with reference to FIG.

  For each pixel of the image data, the distribution area A encoding unit 135 has a pixel value for the pixel assigned to the distribution area A, and a pixel assigned to the area for the pixels assigned to the distribution area B and the distribution area C. Each area intermediate identification code indicating that it is is encoded. Further, the distribution area B encoding unit 136 performs pixel value encoding processing on the pixels assigned to the distribution area B. The same applies to the distribution area C encoding unit 137.

  The result memory 138 stores the encoded data encoded by the distribution area A encoding unit 135, the distribution area B encoding unit 136, and the distribution area C encoding unit 137. The area offset calculation unit 139 calculates the size of the encoded data of each distribution area as an offset value, combines the encoded data as one encoded data based on the offset value, and writes it into the result memory 138.

  Next, the flow of compression encoding processing will be described.

[Pixel division allocation processing in compression encoding processing]
FIG. 7 is a flowchart for explaining the flow of the process of allocating image data to the distribution area of each pixel in the process of the compression encoding unit 13 in the present embodiment. As described above with reference to FIG. 6, when image data is buffered in the memory, a cumulative addition value and a standard deviation of pixel values of each pixel included in the image data are calculated. Then, the standard deviation reciprocal calculation unit 132 of the compression encoding unit 13 calculates the reciprocal of the standard deviation, and the distribution boundary determination unit 133 deviates the distribution area based on the coefficient value obtained by multiplying the reciprocal by 100. AB and deviation threshold AC are determined in advance.

  Then, as shown in the flowchart of FIG. 7, the distribution area determination unit 134 first reads image data from the memory (S11). In general, the pixel values of all the pixels in the image data are read by performing the reading process a plurality of times. Therefore, first, the distribution area determination unit 134 reads the pixel value read in one reading process, and calculates the deviation value of the pixel based on the pixel value and the standard deviation (S12).

  Subsequently, the distribution area determination unit 134 determines whether or not the calculated deviation value is less than the deviation threshold AB for each read pixel (S13). If the deviation value of the pixel is less than the deviation threshold AB (YES in S13), the pixel is assigned to the distribution area B. Therefore, the distribution area determining unit 134 stores the pixel value in the area B memory mb (S15). The area B memory mb is a memory included in the distribution area B encoding unit 136 of the compression encoding unit 13.

  In addition, the distribution area determination unit 134 cumulatively adds the pixel values of the pixels assigned to the distribution area B (S16), and counts the number of pixels assigned to the distribution area B (S17). Then, the distribution area determination unit 134 generates a distribution area B intermediate identification code indicating that the distribution area B has been allocated (S18), and stores it in an address corresponding to the pixel in the area A memory ma (S19). The area A memory ma is a memory included in the distribution area A encoding unit 135 of the compression encoding unit 13. The distribution area B intermediate identification code is a common value.

  Returning to the determination in step S13, on the other hand, if the deviation value of the pixel is greater than or equal to the deviation threshold AB (NO in S13), the distribution area determination unit 134 further determines whether or not the deviation value of the pixel is greater than or equal to the deviation threshold AC. Is determined (S14). If the deviation value of the pixel is equal to or greater than the deviation threshold AC (YES in S14), the pixel is assigned to the distribution area C. Therefore, the distribution area determination unit 134 stores the pixel values in the area C memory mc in the same manner as when pixels are allocated to the area B (S20), and accumulates the pixel values of the pixels allocated to the distribution area C. Addition is performed (S21), and the number of pixels assigned to the distribution area C is counted (S22).

  Then, the distribution area determination unit 134 stores a distribution area C intermediate identification code indicating that the distribution area C has been assigned to the address corresponding to the pixel in the area A memory ma (S24). The distribution area C intermediate identification code is a common value and is different from the distribution area B intermediate identification code.

  Then, returning to the determination in step S14, when the deviation value corresponding to the pixel value of the pixel is greater than or equal to the deviation threshold AB and less than the deviation threshold AC (NO in S14), the pixel is assigned to the distribution area A. Therefore, the distribution area determination unit 134 stores the pixel values in the area A memory ma (S25), and cumulatively adds the pixel values of the pixels assigned to the distribution area A (S26), and is assigned to the distribution area A. The number of pixels is counted (S27).

  Then, the distribution area determination unit 134 determines whether or not the distribution area determination has been completed for all the pixels included in the image data to be compressed and encoded (S28). If all the pixels have not been completed (NO in S28), the distribution area determination unit 134 then reads the next pixel value that continues from the memory 17. Similarly, the distribution area determination unit 134 determines the distribution area to be assigned by determining the deviation value of the pixel value of the read pixel (S13, S14), and performs a processing process according to the distribution area.

  As described above, the distribution area determination unit 134 assigns a distribution area to each pixel included in the image data based on the deviation value corresponding to the pixel value, the deviation threshold value AB, and the deviation threshold value AC. Then, the distribution area determination unit 134 stores the pixel value for the pixels assigned to the distribution area A and the area intermediate identification code for the pixels assigned to the distribution area B and the distribution area C in the order of the scanning pixels. Store in ma. In addition, the distribution area determination unit 134 stores the pixel values of the pixels assigned to the distribution area B in the area B memory mb and the pixel values of the pixels assigned to the distribution area C in the area C memory mc.

[Specific examples of area memory]
FIG. 8 is a diagram illustrating specific examples of the area A memory ma, the area B memory mb, and the area C memory mc generated in the process of the flowchart of FIG. The correspondence relationship diagram PIx with the distribution area of the image data in the figure is the same as the correspondence relationship diagram PIx between the pixel and the distribution area shown in FIG. In this example, each pixel in the image data is processed along the scanning direction indicated by the arrow YA.

  In FIG. 8, for example, the pixel x1 is assigned to the distribution area A based on the standard deviation of the pixel value. The pixel x2 to be subsequently processed is assigned to the distribution area B. In this example, pixel x3 is assigned to distribution area A, pixels x4 and x5 are assigned to distribution area B, and pixel x6 is assigned to distribution area C.

  First, a specific example of the area A memory ma will be described.

[Specific example of area A memory ma]
As described above, the area A memory ma has pixel value information for the pixels assigned to the distribution area A, and the corresponding area B intermediate identification code and area C intermediate identification code for the pixels assigned to the distribution areas B and C. Have In this example, the pixel value is an R value. The R value is 8-bit information and takes a value in the range of 0000000 to 11111111 (binary number). Therefore, in order to enable discrimination between the area intermediate identification code and the pixel value, the information of each pixel in the area A memory ma is expanded from 8 bits to 9 bits. Specifically, for the pixel value, for example, 0 (binary number) is added to the most significant bit. On the other hand, for the area B intermediate identification code and the area C intermediate identification code, 1 is added to the most significant bit.

  Specifically, for the pixels assigned to the distribution area A, the pixel value is held in 0 in the most significant bit and in the subsequent 8 bits. For the pixels assigned to the distribution area B, 0 (binary number) is held in the most significant bit, and the value 00000000 (binary number) is held in the following 8 bits. That is, the area B intermediate identification code indicates 100000000. For the pixels allocated to the distribution area C, 0 (binary number) is held in the most significant bit, and the value 00000001 (binary number) is held in the following 8 bits. That is, the area C intermediate identification code indicates 100000001.

  Thus, it is determined whether the pixel is assigned to the distribution area A, or the pixels are assigned to the distribution area B and the distribution area C, based on the most significant bit. When the most significant bit is 1, it is determined whether the pixel is allocated to the distribution area B or the pixel allocated to the distribution area C based on the subsequent 8-bit value.

  For this reason, in FIG. 8, the area A memory ma is assigned to the distribution area A for the pixel x1, and therefore has 9-bit information “000010101 (pixel value)” at address 0. Addresses 0, 1, and 2 indicate relative addresses of the area A memory ma. Since the subsequent pixel x2 is assigned to the distribution area B, the area A memory ma has the 9-bit area B intermediate identification code “100000000” at the address 1. Similarly, since the subsequent pixel x3 is allocated to the distribution area A, the area A memory ma has 9-bit information “0, 000001001 (pixel value)” at address 2. Since the pixel x6 is assigned to the distribution area C, the area A memory ma has the 9-bit area C intermediate identification code “100000001” at the address corresponding to the pixel x6.

[Specific example of area B memory mb]
Next, the area B memory mb will be described. As described above, the area B memory mb has information only about the pixels allocated to the distribution area B. In the area B memory mb, since only the pixel values of the pixels assigned to the distribution area B are held, the information of each pixel is not expanded and the 8-bit pixel value is held.

  Specifically, the area B memory mb of FIG. 8 has an 8-bit pixel value “00011011” at address 0 for the pixel x2 assigned to the distribution area B. Addresses 0, 1, and 2 indicate relative addresses of the area B memory mb. Further, the area B memory mb has an 8-bit pixel value “00011010” at the address 1 for the pixel x4 to be assigned to the distribution area B next in the pixel scanning order. The same applies to other pixels (such as x5) allocated to the distribution area B.

[Specific example of area C memory mc]
Next, the area C memory mc will be described. Similarly to the area B memory mb, the area C memory mc has information only about the pixels assigned to the distribution area C, so the information of each pixel is not expanded and an 8-bit pixel value is held. Specifically, the area C memory mc of FIG. 7 has an 8-bit pixel value “00000011” at address 0 for the pixel x6 assigned to the distribution area C. Addresses 0, 1, and 2 indicate relative addresses of the area C memory mc. The same applies to other pixels assigned to the distribution area C.

  Next, compression encoding processing of pixel values of pixels assigned to each distribution area by the distribution area determination unit 134 will be described.

[Compression coding process flow]
FIG. 9 is a flowchart for explaining the flow of the compression encoding process for each distribution area among the processes of the compression encoding unit 13 in the present embodiment.

  The distribution area A encoding unit 135 of the compression encoding unit 13 first calculates each distribution for each distribution area based on the cumulative addition values (S16, S21, and S26 in FIG. 7) of the calculated pixel values. An average value of pixel values assigned to the area is calculated (S31). Subsequently, the distribution area A encoding unit 135 stores the calculated average value for each distribution area in the result memory 138 (S32). The calculated average value for each distribution area is used when the compressed image data is decoded.

  Subsequently, when the reading of all the pixels has not been completed (NO in S33), the distribution area A encoding unit 135 reads information for one pixel from the area A memory ma (S34). When the read information indicates the area B intermediate identification code (YES in S35), it is determined that the pixel is assigned to the distribution area B, and the processes of steps S36 to S40 are performed.

[For pixels assigned to distribution area B]
First, the distribution area A encoding unit 135 performs compression encoding processing of the area A memory ma related to the target pixel. Specifically, the distribution area A encoding unit 135 encodes the area B intermediate identification code into a variable-length encoded value, and stores it in the area A area ZA included in the result memory 138 (S36). In the compression encoding process according to the present embodiment, information on each pixel is variable-length encoded according to entropy encoding such as Huffman encoding. Entropy coding is lossless coding that converts a value to be coded into a smaller number of bit strings using statistical properties. Specifically, in Huffman coding, a short coded value is assigned to a value having a high appearance frequency, and a long coded value is assigned to a value having a low appearance frequency, thereby compressing the value.

  Therefore, the distribution area A encoding unit 135 in the present embodiment, for the pixels assigned to the distribution area A, the difference value between the pixel value and the average pixel value of the pixels assigned to the distribution area A, the distribution area For the pixels assigned to B and C, the area B intermediate identification code and the area C intermediate identification code are variable length encoded. Distribution area A encoding section 135 assigns short codewords in descending order of appearance frequency based on the difference value, area B intermediate identification code, and area C intermediate identification code appearance frequency. For this reason, in the case of pixels assigned to the distribution area B, the distribution area A encoding unit 135 encodes the area B intermediate identification code and stores it in an address corresponding to the pixel in the area A region ZA.

  Subsequently, the distribution area B encoding unit 136 performs compression encoding processing of the area B memory mb related to the target pixel. The distribution area B encoding unit 136 reads the pixel value of the target pixel from the area B memory mb (S37), and encodes the pixel value (S38). Specifically, the distribution area B encoding unit 136 calculates a short encoded value in descending order of appearance frequency for the difference value between the pixel value of the target pixel and the average pixel value of the pixels assigned to the distribution area B. Variable length coding is performed by assigning. Subsequently, the distribution area B encoding unit 136 cumulatively adds the sizes of the encoded values allocated and encoded to the distribution area B (S39), and also stores the encoded values in the area B area ZB included in the result memory 138. The data is stored at an address corresponding to the pixel (S40).

[For pixels assigned to distribution area C]
On the other hand, when the read information is not the area B intermediate identification code (NO in S35) and indicates the area C intermediate identification code (YES in S41) in step S35 of the flowchart, the target pixel is assigned to the distribution area C. Indicates a pixel. Therefore, the distribution area A encoding unit 135 encodes the area C intermediate identification code into a variable length encoded value, and stores the encoded value in the address corresponding to the pixel in the area A area ZA (S42).

  Subsequently, the distribution area C encoding unit 137 reads the pixel value of the target pixel from the area C memory mc (S43), and encodes the pixel value (S44). Specifically, the distribution area C encoding unit 137 calculates, for the difference value between the pixel value of the target pixel and the average pixel value of the pixels assigned to the distribution area C, the shorter encoded values in descending order of appearance frequency. Variable length coding is performed by assigning. Subsequently, the distribution area C encoding unit 137 cumulatively adds the sizes of the encoded values allocated and encoded to the distribution area C (S45), and the pixels of the area C area ZC that the encoding result memory 138 has. (S46).

[For pixels assigned to distribution area A]
On the other hand, when the read information is not the area B intermediate identification code or the area C intermediate identification code (NO in S35, NO in S41), it indicates that the target pixel is a pixel assigned to the distribution area A. In this case, the distribution area A encoding unit 135, as described above, for the target pixel read from the area A memory ma, the difference value between the pixel value and the average pixel value of the pixels assigned to the distribution area A, and the area B Based on the appearance frequency of the intermediate identification code and the area C intermediate identification code, variable length coding is performed by assigning short codewords in descending order of appearance frequency (S47). Subsequently, the distribution area A encoding unit 135 cumulatively adds the size of the encoded values allocated and encoded to the distribution area A (S48), and also the area A area in which the result memory 138 encodes the difference value. Store in ZA (S49).

  When the compression encoding process for all the pixels included in the image data to be compression encoded is completed (YES in S33), the distribution area A encoding unit 135 obtains the size of each encoded data in the distribution areas A to C as a result. Store in the memory 138. Thereby, the offset value of the encoded data corresponding to the distribution area B and the distribution area C in the combined encoded data is detected. Then, the compression encoding unit 13 combines the encoded data of each distribution area stored in each area area of the result memory 138 and stores the combined data in the external memory 103.

  In this way, based on the pixel information read from the area A memory ma, it is determined to which distribution area the pixel is assigned, and based on the difference value between the pixel value and the average pixel value in the distribution area. A pixel value compression encoding process is performed to generate encoded data.

  At this time, in the area A area ZA in the encoded data, for all pixels, for the pixels allocated to the distribution area A, the encoded values and distributions of the difference values from the average pixel value of the pixels allocated to the distribution area A The pixels assigned to the areas B and C have an encoded value of each area intermediate identification code. As a result, even when the pixel is divided based on the pixel value of each pixel and is compressed and encoded in units of division, the pixel arrangement information is retained, so that decoding can be performed based on the encoded data.

  As described above, the compression encoding unit 13 in the present embodiment assigns each pixel of the image data to a distribution area divided based on the pixel value of the pixel instead of the pixel arrangement position, and compresses the pixel in the distribution area unit. Encode. Thereby, even in the case of image data having a high frequency and a large difference value with adjacent pixels, pixels having similar pixel values in the image data are assigned to the same distribution area. As a result, the difference value between the pixels in the distribution area is suppressed to be small, so that the size of the variable-length encoded value corresponding to the difference value is kept small, and the compression rate of the image data is improved.

  In the present embodiment, for each distribution area, the compression encoding unit 13 sets the difference value between the pixel value of each pixel assigned to the distribution area and the average value of the pixel values in the distribution area. Based on the compression encoding. However, the compression encoding process of the pixels assigned to the distribution area is not limited to this example. The compression encoding unit 13 compresses the pixel value of each pixel assigned to each distribution area based on the difference value between the pixels before and after the process in the distribution area, that is, the difference value between arbitrary pixels in the distribution area. It may be encoded. Also in this case, since highly correlated pixels are assigned to each distribution area, the difference value can be kept small, and the variation of the difference value is also small, so that the compression rate of the image data is improved.

[Example of encoded data]
FIG. 10 is a diagram illustrating an example of encoded data generated by the compression encoding process of the image data compression encoding apparatus according to the present embodiment. As shown in the figure, the encoded data has an area A area ZA, an area B area ZB, and an area C area ZC.

  As shown in FIG. 10, each area region has the size of the area region and the average value of the pixel values assigned to the distribution area in the first blocks a0, b0, c0 of the region. Then, the next block a1 in the area A area ZA has an encoded value A1. The encoded value A1 is compression-encoded based on the difference value between the pixel value of the pixel x1 (FIG. 8) assigned to the distribution area A at the address 0 of the area A memory ma and the average pixel value in the distribution area A. Value.

  The block a2 next to the area A area ZA has an encoded value A2 obtained by compressing and encoding the area B intermediate identification code 100000000 held at the address 1 of the area A memory ma. Similarly, the block a3 is compressed based on the difference value between the pixel value of the pixel x2 (FIG. 8) assigned to the distribution area A in the address 2 of the area A memory ma and the average pixel value in the distribution area A. It has a coded value A3.

  Further, since the block a5 corresponds to the pixel x5 (FIG. 8) assigned to the distribution area B, the block a5 has an encoded value A2 obtained by compressing and encoding the area B intermediate identification code 100000000 as in the block a2. Further, since the block a6 corresponds to the pixel x6 (FIG. 8) assigned to the distribution area C, the block a6 has an encoded value A5 in which the area C intermediate identification code 100000001 is compression-encoded. The same applies to other blocks in the area A area ZA.

  Next, the area B area ZB will be described. The block b1 of the area B area ZB is based on the difference value between the pixel value of the pixel x2 (FIG. 8) assigned to the distribution area B at the address 0 of the area B memory mb and the average pixel value in the distribution area B. It has a coded value B1 that is compression-coded. Similarly, the subsequent block b2 is also compression-encoded based on the difference value between the pixel value of the pixel assigned to the distribution area B in the address 1 of the area B memory mb and the average pixel value in the distribution area B. Has the encoded value B2. The same applies to other blocks in the area B area ZB.

  Next, the area C area ZC will be described. The block c1 of the area C region ZC is based on the difference value between the pixel value of the pixel x6 (FIG. 8) assigned to the distribution area C at the address 0 of the area C memory mc and the average pixel value in the distribution area C. It has an encoded value C1 that has been compression encoded. The same applies to other blocks in the area C area ZC.

[Decoding of compressed code data]
The encoded data shown in FIG. 10 is decoded as follows. As described above with reference to FIG. 10, the head blocks a0, b0, and c0 of each area of the encoded data hold the size information of each area and the average value of the pixel values assigned to the corresponding distribution area. Is done.

  Therefore, first, the decoding unit 14 determines the size of the area A area ZA and the average pixel value of the pixels assigned to the distribution area A based on the size information of the area A area ZA held at the head of the compression encoded data. Is read. Subsequently, based on the size of the area A area ZA, the decoding unit 14 reads out the area B area ZB size held in the head block of the area B area ZB and the average pixel value of the pixels assigned to the distribution area B. . Similarly, the decoding unit 14 reads out the area C region ZC size held in the top block of the area C region ZC and the average pixel value of the pixels assigned to the distribution area C based on the area B region ZB size. .

  Then, the decoding unit 14 reads and decodes the encoded value A1 of the block a1 in the area A area. When the decoded value is not the area B intermediate identification code or the area C intermediate identification code, the decoding unit 14 determines that the pixel corresponding to the block a1 is a pixel assigned to the distribution area A, and the distribution area A pixel value is generated based on the average pixel value in A and the decoded difference value.

  On the other hand, when the decoded value is the area B intermediate identification code, the decoding unit 14 decodes the encoded value of the corresponding block in the area B region ZB to generate a difference value. Then, the decoding unit generates a pixel value based on the average pixel value and the difference value in the distribution area B. Similarly, when the decoded value is an area C intermediate identification code, the decoding unit 14 generates a difference value by decoding the encoded value held in the area C region ZC, and distributes the distribution area. A pixel value is generated based on the C average pixel value and the difference value.

  In this manner, the decoding unit 14 can decode the image data that has been compression-encoded by the compression-encoding unit 13 in the present embodiment.

  As described above, the image data compression coding apparatus 10 according to the present embodiment has the first deviation threshold (deviation threshold AB) and the first deviation based on the standard deviation of the pixel value of each pixel in the image data. A second deviation threshold value (deviation threshold value AC) larger than the deviation threshold value is determined. In addition, for each pixel, the image data compression encoding device 10 has pixels in which the deviation value of the pixel value is less than the first deviation threshold value in the first division (distribution area B), and the deviation value is greater than or equal to the first deviation threshold value. Pixels less than the second deviation threshold are assigned to the second division (distribution area A), pixels having deviation values equal to or greater than the second deviation threshold are assigned to the third division (distribution area C), and assigned for each of the first to third divisions. The encoded pixel data is generated by encoding the pixel value of each pixel.

  As described above, the image data compression coding apparatus 10 according to the present embodiment assigns each pixel of the image data to the divided distribution area based on the pixel value instead of the pixel position. Thereby, even in the case of image data having a high frequency and low correlation with adjacent pixels, pixels having similar pixel values in the image data are assigned to the same distribution area, so that the difference value between the pixels in the distribution area can be kept small. . For this reason, the size of the variable length encoded value assigned to the difference value is kept small, and the compression rate of the image data is improved. Accordingly, the image data compression coding apparatus 10 can effectively perform compression coding even for image data having a low correlation with adjacent pixels and a high frequency.

  Further, the image data compression coding apparatus 10 according to the present embodiment uses a value obtained by subtracting a calculated value obtained by multiplying the reciprocal of the standard deviation by 100 from the first deviation threshold value (deviation threshold value AB), and the calculated value as the calculated value. A value added to 50 is determined as the second deviation threshold (deviation threshold AC). Thereby, in the case of the first standard deviation having a small standard deviation and a small variation in pixel values, the deviation value range corresponding to the second section (distribution area A) is widened. Further, in the case of the second standard deviation in which the standard deviation is larger than the first standard deviation and the pixel value variation is large, the deviation value range corresponding to the second section (distribution area A) is narrowed.

  As described above, when the standard deviation is large and the variation of the pixel value in the image data is large, the image data compression encoding apparatus 10 according to the present embodiment classifies the distribution area based on the reciprocal of the standard deviation, thereby The deviation value range corresponding to the distribution area A centered on the deviation 50 can be narrowed, and the deviation value range corresponding to the distribution areas B and C can be widened. Thereby, the image data compression encoding apparatus 10 can reduce the difference value of the pixel value between the pixels assigned to each distribution area even when the pixel value varies greatly and is distributed over a wide range. . As a result, the size of the encoded value corresponding to the difference value is kept small, and the compression rate of the image data is improved.

  In addition, when the standard deviation is small and the variation in pixel values is small, the image data compression coding apparatus 10 according to the present embodiment sorts the distribution area based on the reciprocal of the standard deviation, so that the standard deviation 50 is the center. The deviation value range corresponding to the distribution area A can be widened. Thus, the image data compression encoding apparatus 10 assigns a pixel group in which the distribution is concentrated to one distribution area (distribution area A) when the pixel value variation is small and distributed in a narrow range, and is assigned to each distribution area. The difference value of the pixel value between the obtained pixels can be reduced. As a result, the size of the encoded value corresponding to the difference value is kept small, and the compression rate of the image data is improved.

  In addition, the image data compression coding apparatus 10 according to the present embodiment, for each pixel assigned to the first to third sections (distribution areas), averages the pixel values of the assigned pixels for each of the first to third sections. The encoded data is generated by variable-length encoding the difference value between the value and the pixel value. Alternatively, the image data compression coding apparatus 10 according to the present embodiment, for each pixel assigned to the first to third sections (distribution area), the pixel value between the assigned pixels for each of the first to third sections. The encoded data is generated by variable-length encoding the difference value.

  According to the image data compression encoding device 10 in the present embodiment, the difference value of the pixel value between the pixels in each section is suppressed to be small by being assigned to the section based on the pixel value. For this reason, the image data compression coding apparatus 10 performs variable length coding based on the average value of the pixel value and the divided pixel or the difference value of the pixel value between the divided pixels, thereby changing the variable length corresponding to the difference value. The size of encoding can be kept small, and the compression rate of image data can be improved.

  In addition, the image data compression encoding apparatus 10 according to the present embodiment has a code of one of the first to third sections (distribution area A to distribution area C) (distribution area A in the present embodiment). In the conversion processing, in the pixel order of the image data, the difference value is applied to the pixels assigned to one section (distribution area A), and the pixels assigned to the sections other than one section (distribution areas B and C) are Each segment identification value (area intermediate identification code) indicating that it is allocated to a segment is variable-length encoded to generate the encoded data.

  As a result, the image data compression encoding device 10 assigns to a segment based on the pixel value of each pixel, and maintains the arrangement of each pixel in the image data even when compression encoding is performed for each segment. Data can be compression encoded. For this reason, the image data compression encoding apparatus 10 enables decoding of the original image data based on the encoded data that has been compression encoded. Note that one section is preferably a section to which the most pixels are assigned. This is because the size of the entire encoded data can be further reduced by reducing the number of encoded values corresponding to the segment identification values in the encoded data corresponding to one segment.

  Note that the compression encoding process according to the present embodiment may be performed by being stored as a program in a computer-readable recording medium, and the computer reading and executing the program.

  The above embodiment is summarized as follows.

(Appendix 1)
Threshold determination means for determining a first deviation threshold and a second deviation threshold larger than the first deviation threshold based on the standard deviation of the pixel value of each pixel in the image data;
For each of the pixels, a pixel whose deviation value is less than the first deviation threshold is a first division, and a pixel whose deviation value is greater than or equal to the first deviation threshold and less than the second deviation threshold is the first. 2 divisions, a pixel whose deviation value is equal to or greater than the second deviation threshold is assigned to the third division, and encoded data is generated by encoding the assigned pixel value for each of the first to third divisions. An image data compression coding apparatus.

(Appendix 2)
In Appendix 1,
The threshold value determining means determines a value obtained by subtracting a calculated value obtained by multiplying the reciprocal of the standard deviation by 100 from 50 as the first deviation threshold value, and a value obtained by adding the calculated value to 50 as the second deviation threshold value. And
When the standard deviation is the first standard deviation, a deviation value range corresponding to the second section is widened, and when the standard deviation is a second standard deviation larger than the first standard deviation, the second section The image data compression encoding apparatus in which the deviation value range corresponding to the above becomes narrower.

(Appendix 3)
In Appendix 1 or 2,
The encoding means, for each of the pixels assigned to the first to third sections, varies a difference value between a pixel value average value of the assigned pixels and each pixel value for each of the first to third sections. An image data compression encoding apparatus that generates the encoded data by encoding.

(Appendix 4)
In Appendix 1 or 2,
The encoding means performs variable length encoding on the difference value of the pixel value between the allocated pixels for each of the first to third sections for each of the pixels allocated to the first to third sections, and performs the encoding. An image data compression encoding device for generating data.

(Appendix 5)
In Appendix 3 or 4,
In the encoding processing of one of the first to third sections, the encoding unit is configured to apply the difference value for pixels assigned to the one section other than the one section in the pixel order of the image data. An image data compression encoding apparatus that generates the encoded data by variable-length encoding each segment identification value indicating that the pixel is allocated to each segment.

(Appendix 6)
Determining a first deviation threshold and a second deviation threshold greater than the first deviation threshold based on the standard deviation of the pixel values of each pixel in the image data;
For each of the pixels, a pixel whose deviation value is less than the first deviation threshold is a first division, and a pixel whose deviation value is greater than or equal to the first deviation threshold and less than the second deviation threshold is the first. 2 divisions, a pixel whose deviation value is equal to or greater than the second deviation threshold is assigned to the third division, and encoded data is generated by encoding the assigned pixel value for each of the first to third divisions. A method for compressing and encoding image data.

(Appendix 7)
Determining a first deviation threshold and a second deviation threshold greater than the first deviation threshold based on the standard deviation of the pixel values of each pixel in the image data;
For each of the pixels, a pixel whose deviation value is less than the first deviation threshold is a first division, and a pixel whose deviation value is greater than or equal to the first deviation threshold and less than the second deviation threshold is the first. 2 divisions, a pixel whose deviation value is equal to or greater than the second deviation threshold is assigned to the third division, and encoded data is generated by encoding the assigned pixel value for each of the first to third divisions. An image data compression encoding program for causing a computer to execute the processing to perform.

10: Image data compression encoding device, 11: Video input interface, 12: Image processing unit, 13: Compression encoding unit, 14: Decoding unit, 15: Video display unit, 16: Processor, 17: Memory

Claims (7)

Threshold determination means for determining a first deviation threshold and a second deviation threshold larger than the first deviation threshold based on the standard deviation of the pixel value of each pixel in the image data;
For each of the pixels, a pixel whose deviation value is less than the first deviation threshold is a first division, and a pixel whose deviation value is greater than or equal to the first deviation threshold and less than the second deviation threshold is the first. 2 divisions, a pixel whose deviation value is equal to or greater than the second deviation threshold is assigned to the third division, and encoded data is generated by encoding the assigned pixel value for each of the first to third divisions. An image data compression coding apparatus. In claim 1,
The threshold value determining means determines a value obtained by subtracting a calculated value obtained by multiplying the reciprocal of the standard deviation by 100 from 50 as the first deviation threshold value, and a value obtained by adding the calculated value to 50 as the second deviation threshold value. And
When the standard deviation is the first standard deviation, a deviation value range corresponding to the second section is widened, and when the standard deviation is a second standard deviation larger than the first standard deviation, the second section The image data compression encoding apparatus in which the deviation value range corresponding to the above becomes narrower. In claim 1 or 2,
The encoding means, for each of the pixels assigned to the first to third sections, varies a difference value between a pixel value average value of the assigned pixels and each pixel value for each of the first to third sections. An image data compression encoding apparatus that generates the encoded data by encoding. In claim 1 or 2,
The encoding means performs variable length encoding on the difference value of the pixel value between the allocated pixels for each of the first to third sections for each of the pixels allocated to the first to third sections, and performs the encoding. An image data compression encoding device for generating data. In claim 3 or 4,
In the encoding processing of one of the first to third sections, the encoding unit is configured to apply the difference value for pixels assigned to the one section other than the one section in the pixel order of the image data. An image data compression encoding apparatus that generates the encoded data by variable-length encoding each segment identification value indicating that the pixel is allocated to each segment. Determining a first deviation threshold and a second deviation threshold greater than the first deviation threshold based on the standard deviation of the pixel values of each pixel in the image data;
For each of the pixels, a pixel whose deviation value is less than the first deviation threshold is a first division, and a pixel whose deviation value is greater than or equal to the first deviation threshold and less than the second deviation threshold is the first. 2 divisions, a pixel whose deviation value is equal to or greater than the second deviation threshold is assigned to the third division, and encoded data is generated by encoding the assigned pixel value for each of the first to third divisions. A method for compressing and encoding image data. Determining a first deviation threshold and a second deviation threshold greater than the first deviation threshold based on the standard deviation of the pixel values of each pixel in the image data;
For each of the pixels, a pixel whose deviation value is less than the first deviation threshold is a first division, and a pixel whose deviation value is greater than or equal to the first deviation threshold and less than the second deviation threshold is the first. 2 divisions, a pixel whose deviation value is equal to or greater than the second deviation threshold is assigned to the third division, and encoded data is generated by encoding the assigned pixel value for each of the first to third divisions. An image data compression encoding program for causing a computer to execute the processing to perform.

Images