JP2010050697A - Apparatus and method for compressing/decompressing image - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an apparatus and method for compressing/decompressing an image for improving processing speed of decoding processing. <P>SOLUTION: An encoder 10 is provided with: an encoding part 101 which performs variable length encoding to a symbol value to generate encoded data; a first packing part 102 which packs N prefixes and N suffixes to generate first packing data for one pixel of the generated encoded data; and a second packing part 103 which packs the first packing data generated by the first packing part 102 by unit of prescribed pixels to generate second packing data. A decoder 30 is provided with: an unpacking part 301 which unpacks the second packing data generated by the second packing part 103 to segment the head of the encoded data; a register 302 which stores the encoded data whose head is segmented; and a decoding part 303 which decodes the encoded data stored in the register 302. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an image compression / decompression apparatus and an image compression / decompression method used for compressing and decompressing image data.

  When encoding a symbol with encoded data using a prefix that is a variable length code and a suffix that is a fixed length code, the encoding device alternately arranges the prefix and the suffix, and packs them. Generate packing data. A decoding apparatus for decoding such encoded data performs the following processing in order to process encoded data corresponding to N (N is a natural number) symbols in one cycle.

  It is assumed that encoded data to be processed in a certain cycle is packed on the MSB (Most Significant Bit) side and stored in a register. First, the decoding apparatus inputs the data packed and stored on the MSB side of the register to the decoding unit corresponding to the first (N = 1) symbol, and from this encoded data, The code length of the prefix corresponding to the first symbol is calculated, and the fixed length of the suffix corresponding to the first symbol is added to the calculation result to calculate the code length of the first symbol. Next, the decoding apparatus obtains a second value (N = N == left shift data) obtained by shifting the encoded data stored in the register to the left using the code length of the first symbol. Input to the decoding unit corresponding to the symbol of 2). At this time, the decoding apparatus packs and stores the Nth left shift data on the MSB side of the register.

However, in general image data, since one pixel is composed of three components (that is, N = 3), assuming that each component corresponds to each symbol, three symbols are decoded in one cycle. There is a need to. At this time, when the clock frequency of the decoding apparatus is high, the calculation of the symbol code length and the shift operation become a critical path of timing, so the processing speed of the decoding process cannot be improved.
JP-A-8-116266

  An object of the present invention is to provide an image compression / decompression apparatus and an image compression / decompression method for improving the processing speed of a decoding process.

According to the first aspect of the present invention,
An image compression / decompression apparatus comprising: an encoding device that encodes symbol values of image data to generate encoded data; and a decoding device that decodes encoded data generated by the encoding device. ,
The encoding device includes:
An encoding unit for variable-length encoding the symbol value according to a predetermined encoding method to generate the encoded data;
N prefixes and N suffixes corresponding to N (N is a natural number) symbols constituting one pixel of the encoded data generated by the encoding unit are set to the N prefixes after the N prefixes. A first packing unit that packs so that the suffix is arranged and generates first packing data;
A second packing unit that packs the first packing data generated by the first packing unit in units of predetermined pixels and generates second packing data;
The decoding device
An unpacking unit that unpacks the second packing data generated by the second packing unit and cuts out the beginning of the encoded data to be decoded;
A register for storing the encoded data cut out by the unpacking unit;
An image compression / decompression apparatus comprising a decoding unit that decodes encoded data stored in the register is provided.

According to a second aspect of the invention,
An image compression / decompression method for compressing and decompressing image data,
Variable-length encoding the symbol value of the image data according to a predetermined encoding method to generate encoded data;
N prefixes and N suffixes corresponding to N (N is a natural number) symbols constituting one pixel of the encoded data are arranged, and the N suffixes are arranged after the N prefixes. To generate first packing data,
Packing the first packing data in units of predetermined pixels to generate second packing data;
Unpacking the second packing data, generating encoded data, and storing the encoded data in a register;
An image compression / decompression method is provided, wherein the encoded data stored in the register is decoded.

  According to the present invention, the processing speed of the decoding process can be improved.

  Embodiments of the present invention will be described below with reference to the drawings. The following examples are one embodiment of the present invention and do not limit the scope of the present invention.

  First, Example 1 of the present invention will be described. The first embodiment of the present invention is an example of a decoding process performed using packing data including N prefixes and N suffixes arranged after the N prefixes. Here, in Embodiment 1 of the present invention, it is assumed that the code length of each prefix is unknown and the code length of each suffix is known.

  FIG. 1 is a block diagram illustrating a configuration of an image compression / decompression apparatus 1 according to Embodiment 1 of the present invention.

  As shown in FIG. 1, the image compression / decompression apparatus 1 according to the first embodiment of the present invention includes an encoding device 10, a memory subsystem 20, and a decoding device 30.

  As illustrated in FIG. 1, the encoding device 10 includes an encoding unit 101, a first packing unit 102, and a second packing unit 103.

  As shown in FIG. 1, the encoding unit 101 reads image data from an external device (not shown) (11), variable-codes the symbol value of the image data according to a predetermined encoding method, The prefix and suffix corresponding to are sent to the first packing unit 102 (12). For example, the encoding unit 101 includes a prediction encoding module and a quantization module such as DPCM (Differential Pulse Code Modulation) processing, and an inverse quantization module and an inverse prediction encoding module. In this case, the difference value between the plurality of image data calculated by the DPCM process becomes the symbol value.

  As shown in FIG. 1, the first packing unit 102 adds N prefixes and N suffixes to N symbols that constitute one pixel of the encoded data generated by the encoding unit 101. Packing is performed so that N suffixes are arranged after N prefixes, first packing data for one pixel (see FIG. 2) is generated, and sent to the second packing unit 103 (13). N is a value determined according to the encoding method of the encoding unit 101 and is a value indicating the number of components for one pixel. For example, when the relationship between the luminance signal Y and the color difference signals Cb and Cr is Y: Cb: Cr = 4: 4: 4, N = 3, and Y: Cb: Cr = 4: 2: 2. N = 2.

  FIG. 2 is a schematic diagram illustrating a data structure of the first packing data when N = 3 according to the first embodiment of the present invention.

  As shown in FIG. 2, the first packing data when N = 3 includes a prefix part including prefixes (Prefix (i), (i + 1), (i + 2)) corresponding to the i-th to i + 2th symbols, And a suffix part including suffixes (Suffix (i), (i + 1), (i + 2)) corresponding to the i-th to i + 2th symbols arranged after the prefix part. Each prefix is a variable length code, and each suffix is a fixed length code.

  Further, as shown in FIG. 1, the second packing unit 103 sequentially receives the first packing data for one pixel sent from the first packing unit 102, and the first packing data for M (M is an integer of 2 or more) pixels. One packing data is packed to generate second packing data (see FIG. 3). Then, when second packing data corresponding to the bus width of the bus 201 described later is generated, the second packing unit 103 outputs the second packing data to the bus 201 described later (14). M is a value indicating the number of pixels constituting the second packing data.

  FIG. 3 is a schematic diagram illustrating a data structure of the second packing data when M = 32 according to the first embodiment of the present invention.

  As shown in FIG. 3, the second packing data when M = 32 has a data structure composed of the first packing data for 32 pixels. Therefore, when N = 3, the second packing data when M = 32 includes 96 prefixes and 96 suffixes.

  As shown in FIG. 1, the memory subsystem 20 includes a bus 201, a system memory controller 202, a system memory 203, and a DMA (Direct Memory Access) controller (not shown).

  As shown in FIG. 1, the bus 201 is connected to the second packing unit 103, the system memory controller 202, and an unpacking unit 301 described later. The bus 201 transfers data for a predetermined bus width (for example, 64 bits or 32 bits) to the second packing unit 103, the system memory controller 202, and an unpacking unit 301 described later.

  Further, as shown in FIG. 1, the system memory controller 202 controls the system memory 203. Further, the system memory controller 202 receives the second packing data corresponding to the bus width output from the second packing unit 103 via the bus 201 (21), and writes the second packing data into the system memory 203 (22). The system memory controller 202 reads the second packing data stored in the system memory 203 (22), and outputs the second packing data to the unpacking unit 301 described later via the bus 201 (21, 31). .

  As shown in FIG. 1, the system memory 203 can store various data including the second packing data sent from the second packing unit 103 based on the control of the system memory controller 202.

  A DMA controller (not shown) controls data transfer among the second packing unit 103, the bus 201, the system memory controller 202, and the system memory 203.

  As illustrated in FIG. 1, the decoding device 30 includes an unpacking unit 301, a register 302, and a decoding unit 303.

  As shown in FIG. 1, the unpacking unit 301 includes a data buffer that stores encoded data. The unpacking unit 301 receives second packing data corresponding to the bus width from the bus 201 when the data buffer is empty (31), and stores the second packing data in the data buffer. In addition, the unpacking unit 301 generates N symbols based on the total code length of N prefixes and the total code length of N suffixes included in the encoded data stored in the register 302. The total code length is calculated. Also, the unpacking unit 301 concatenates the encoded data stored in the register 302 and the encoded data stored in the data buffer, and shifts the register 302 by the total value of the code lengths of N symbols. An operation is performed, and the beginning of the encoded data of N symbols to be decoded next is cut out and the encoded data is stored in the register 302 (32). Further, the unpacking unit 301 ensures that the encoded data for at least one pixel is always stored in the register 302. The unpacking unit 301 includes an adder and a shift calculator for realizing these processes, and a table indicating the relationship between the code and the code length.

  Also, as shown in FIG. 1, the register 302 stores encoded data for at least one pixel packed in the MSB side.

  FIG. 4 is a flowchart showing the processing procedure of the decoding process of the decoding device 30 according to the first embodiment of the present invention.

  As shown in FIG. 4, first, a decoding step (S401) is performed. In the decoding step (S401), the decoding unit 303 reads the encoded data stored in the register 302 from the MSB side, decodes the encoded data, and outputs it to an external device (not shown).

  Next, as shown in FIG. 4, a prefix code length calculation step (S402) is performed. In the prefix code length calculation step (S402), the unpacking unit 301 refers to the N prefixes that are packed and stored on the MSB side of the register 302, and calculates the total code length of the N prefixes. calculate. When N = 3, the calculation result of the prefix code length calculation step (S402) is the sum of the code lengths of Prefix (i) to Prefix (i + 2).

  Next, as shown in FIG. 4, a symbol code length calculation step (S403) is performed. In the symbol code length calculation step (S403), the unpacking unit 301 adds the total value of the code lengths of known suffixes to the calculation result of the prefix code length calculation step (S402).

  Next, as shown in FIG. 4, a shift calculation step (S404) is performed. In the shift calculation step (S404), the unpacking unit 301 performs a shift calculation for left-shifting the encoded data stored in the register 302 based on the calculation result of the symbol code length calculation step (S403). At this time, the left shift data obtained by shifting the encoded data stored in the register 302 to the left becomes the encoded data to be encoded next.

  S401 to S405 are repeated as long as there is encoded data to be decoded next (S405-NO).

  As shown in FIG. 4, the decoding process according to the first embodiment of the present invention ends when there is no encoded data to be decoded next (S405-YES).

  In the first embodiment of the present invention, the encoding unit 101 may generate encoded data by encoding a symbol value in accordance with the Golomb-Rice encoding method. In this case, the variable length code of the prefix is an Unary code.

  According to the first embodiment of the present invention, immediately before the prefix code length calculation step (S402 in FIG. 4), the register 302 stores Prefix (i) packed in the MSB side and stores the shift calculation step (S405 in FIG. 4). Immediately after (), Prefix (i + N) is packed and stored on the MSB side. That is, it is possible to perform decoding processing of encoded data corresponding to N symbols by performing the symbol code length calculation step (S403) and the shift calculation step (S404), which are critical paths, in the conventional technique only once. As a result, it is possible to reduce the influence of the critical path on the decoding process, and as a result, the processing speed of the decoding process can be improved.

  Next, a second embodiment of the present invention will be described. The second embodiment of the present invention is an example of a decoding process performed on encoded data in which the suffix code length is different for each symbol. In addition, the description about the content similar to Example 1 of this invention is abbreviate | omitted. Here, in the second embodiment of the present invention, it is assumed that the code length of each prefix and each suffix is unknown.

  As illustrated in FIG. 1, the second packing unit 103 according to the second embodiment of the present invention packs first packing data for M pixels and a header to be described later, and generates second packing data (see FIG. 5). Then, the data is output to the bus 201 (14).

  FIG. 5 is a schematic diagram illustrating a data structure of the second packing data when N = 3 and M = 32 according to the second embodiment of the present invention.

  As shown in FIG. 5, the second packing data according to the second embodiment of the present invention has a data structure including a header and first packing data for 32 pixels. The header includes the code length of N suffixes (Suffix (i), (i + 1), (i + 2)) of the first packing data.

  Also, as shown in FIG. 4, in the decoding process according to the second embodiment of the present invention, in the symbol code length calculation step (S403), the unpacking unit 301 stores the second packing data stored in the register 302. The total code length of the suffix is calculated with reference to the header, and the calculated code length of the suffix is added to the calculation result of the prefix code length calculation step (S402).

  According to the second embodiment of the present invention, since a header including data indicating the code length of N suffixes is added to the second packing data, the encoded data having a different code length of the suffix for each symbol. The processing speed of the decryption process performed in this way can be improved.

1 is a block diagram illustrating a configuration of an image compression / decompression device 1 according to Embodiment 1 of the present invention. It is the schematic which shows the data structure of the 1st packing data at the time of N = 3 concerning Example 1 of this invention. It is the schematic which shows the data structure of the 2nd packing data in case M = 32 which concerns on Example 1 of this invention. It is a flowchart which shows the process sequence of the decoding process of the decoding apparatus 30 which concerns on Example 1 of this invention. It is the schematic which shows the data structure of the 2nd packing data in case of N = 3 and M = 32 based on Example 2 of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Image compression / decompression apparatus 10 Encoding apparatus 101 Encoding part 102 1st packing part 103 2nd packing part 20 Memory subsystem 201 Bus 202 System memory controller 203 System memory 30 Decoding apparatus 301 Unpacking part 302 Register 303 Decoding Part

Claims (5)

An image compression / decompression apparatus comprising: an encoding device that encodes symbol values of image data to generate encoded data; and a decoding device that decodes encoded data generated by the encoding device. ,
The encoding device includes:
An encoding unit for variable-length encoding the symbol value according to a predetermined encoding method to generate the encoded data;
N prefixes and N suffixes corresponding to N (N is a natural number) symbols constituting one pixel of the encoded data generated by the encoding unit are set to the N prefixes after the N prefixes. A first packing unit that packs so that the suffix is arranged and generates first packing data;
A second packing unit that packs the first packing data generated by the first packing unit in units of predetermined pixels and generates second packing data;
The decoding device
An unpacking unit that unpacks the second packing data generated by the second packing unit and cuts out the beginning of the encoded data to be decoded;
A register for storing the encoded data cut out by the unpacking unit;
An image compression / decompression apparatus comprising: a decoding unit that decodes encoded data stored in the register.   The unpacking unit calculates a total value of code lengths of the N symbols based on a total value of code lengths of the N prefixes and a total value of code lengths of the N suffixes. The image compression / decompression apparatus according to claim 1, wherein a shift operation is performed on the register based on a result.   The image compression / decompression device according to claim 1, wherein the second packing unit adds the code length of the N suffixes as a header to generate the second packing data. An image compression / decompression method for compressing and decompressing image data,
Variable-length encoding the symbol value of the image data according to a predetermined encoding method to generate encoded data;
N prefixes and N suffixes corresponding to N (N is a natural number) symbols constituting one pixel of the encoded data are arranged, and the N suffixes are arranged after the N prefixes. To generate first packing data,
Packing the first packing data in units of predetermined pixels to generate second packing data;
Unpacking the second packing data, cutting out the beginning of the encoded data to be decoded, and storing the encoded data in a register;
An image compression / decompression method, wherein the encoded data stored in the register is decoded.   A total value of the code lengths of the N symbols is calculated based on a total value of the code lengths of the N prefixes and a total value of the code lengths of the N suffixes. 5. The image compression / decompression method according to claim 4, wherein a shift operation is performed.

Images