JP2011019097A - Decoder, encoder, image processing device and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a decoder which efficiently performs decoding processing.SOLUTION: A starting position information decoding part 332 decodes a starting position information code to acquire starting position information of a block. A DC component decoding part 334 decodes DC components on the basis of the starting position information decoded by the starting position information decoding part 332, and outputs code length of the DC components to a data configuration unit decoding part 36. An AC component head position calculation part 362 calculates the head position of the AC components on the basis of the decoded starting position information of the block and the code length of the DC components. An AC component decoding part 364 performs variable length decoding of a data configuration unit code included in each one block on the basis of the calculated head position of the AC components. An IDCT processing part 366 performs inverse discrete cosine transform to the DC components and the AC components to which the variable length decoding is performed to decode image data of a part corresponding to the processed block.

Description

  The present invention relates to a decoding device, an encoding device, an image processing device, and a program.

  In Patent Document 1, an encoded signal sequence is decomposed as a plurality of signal sequences from a specific fixed-length code to the next specific fixed-length code, each of the decomposed signal sequences is once written in a plurality of buffer memories, and a plurality of time-sequential sequences are written. An encoding / decoding device that reads a plurality of signal sequences from the buffer memory is disclosed.

Japanese Patent Application Laid-Open No. 11-215009

  An object of the present invention is to provide a decoding device that efficiently performs a decoding process.

[Decoding device]
According to the first aspect of the present invention, there is provided correlation information decoding means for decoding correlation information mutually correlated between a plurality of processing units each including a plurality of data constituent units constituting data, and the plurality of encoded data Distributing means for distributing the constituent units to a plurality of data constituent unit decoding means, and the plurality of data structures for decoding each of the distributed data constituent units in parallel based on the correlation information decoded by the correlation information decoding means A decoding device having unit decoding means.

  According to a second aspect of the present invention, the correlation information decoding unit decodes information related to a start position of each of the plurality of processing units, and the data component unit decoding unit is based on the decoded start position information. The decoding apparatus according to claim 1, wherein a start position of a data configuration unit within the processing unit is calculated.

  According to the third aspect of the present invention, the correlation information decoding unit decodes information related to a difference between the processing unit to be processed and the adjacent processing unit, and the data configuration unit decoding unit includes the decoded difference. The decoding device according to claim 1, wherein the data structural unit is decoded based on information.

According to a fourth aspect of the present invention, the correlation information decoding means indicates prediction method information indicating which processing unit of a plurality of other processing units is to be encoded with respect to the processing unit to be processed. The decoding unit according to claim 1, wherein the data component unit decoding unit decodes the data component unit based on the decoded prediction scheme information.
According to a fifth aspect of the present invention, the correlation information decoding means decodes the prediction method information according to the correlation information corresponding to the processed processing unit for the processing unit to be processed. It is a decoding apparatus of description.

[Encoding device]
According to a sixth aspect of the present invention, there is provided a data code for encoding a plurality of data constituent units constituting data to be processed and correlation information correlated with each other between the plurality of processing units including the plurality of data constituent units. And a start position encoding means for encoding information related to the start positions of each of the plurality of processing units.

  According to a seventh aspect of the present invention, there is provided a data code for encoding a plurality of data constituent units constituting data to be processed and information relating to a difference between processing units adjacent to the processing unit including the plurality of data constituent units. And a start position encoding means for encoding information related to the start positions of each of the plurality of processing units.

The present invention according to claim 8 divides the data to be processed into a plurality of processing units each including a plurality of data constituent units, and a plurality of other processing units for each of the divided processing units. A prediction method encoding means for encoding prediction method information indicating which processing unit is to be encoded, and for each of the plurality of data constituent units, a prediction error is calculated based on the prediction method information An encoding apparatus comprising: a prediction error calculating means for encoding; a prediction error encoding means for encoding the calculated prediction error; and a start position information encoding means for encoding the start position information of the processing unit.
The present invention according to claim 9 is characterized in that the prediction method encoding means encodes the prediction method information according to the correlation information corresponding to the processed processing unit for the processing unit to be processed. 8. The encoding device according to 8.

[Image processing device]
According to the tenth aspect of the present invention, there is provided correlation information decoding means for decoding correlation information mutually correlated between a plurality of processing units each including a plurality of data constituting units constituting image data, and the plurality of encoded plural pieces Distributing means for distributing the data constituent unit to a plurality of data constituent unit decoding means, and the plurality of data for decoding each of the distributed data constituent units in parallel based on the correlation information decoded by the correlation information decoding means An image processing apparatus comprising: a constituent unit decoding unit; and an image output unit that outputs the image data based on the plurality of data constituent units decoded by the data unit decoding unit.

[Decryption program]
The present invention according to claim 11 is a correlation information decoding step for decoding correlation information mutually correlated between a plurality of processing units each including a plurality of data constituent units constituting data, and the plurality of encoded data A decoding program for causing a computer to execute a distribution step of distributing a constituent unit and a data constituent unit decoding step of decoding each of the distributed data constituent units in parallel based on the decoded correlation information.

  According to the first aspect of the present invention, it is possible to provide a decoding device that performs a decoding process more efficiently than when the present configuration is not provided.

  According to the second aspect of the present invention, in addition to the effect obtained by the first aspect of the present invention, the parallel decoding processing of the data structure unit is efficiently performed as compared with the case where the present structure is not provided. It can be carried out.

  According to the third aspect of the present invention, in addition to the effect obtained by the first or second aspect of the present invention, in addition to the encoding method that assigns a short code to a data constituent unit having a high appearance frequency within each processing unit The encoded information can be decoded.

According to the fourth aspect of the present invention, in addition to the effect obtained by the first or second aspect of the present invention, the encoded information encoded by the encoding method using the difference between adjacent data constituent units. Can be decrypted.
According to the fifth aspect of the present invention, in addition to the effect obtained by the present invention according to the fourth aspect, the encoding is performed by the encoding scheme that assigns a shorter code to a higher rank when coding the rank of the prediction scheme. The encoded information can be decoded.

  According to the sixth aspect of the present invention, decoding processing can be performed in parallel for each processing unit when data is decoded.

  According to the present invention of claim 7, at the time of data decoding, it is possible to decode encoded information encoded by an encoding method in which a short code is assigned to a data constituent unit having a high appearance frequency within each processing unit.

According to the eighth aspect of the present invention, when decoding data, it is possible to decode the encoded information encoded by the encoding method using the difference between the adjacent data constituent units.
According to the present invention of claim 9, in addition to the effect obtained by the present invention of claim 8, the data compression rate can be improved as compared with the case where the present configuration is not provided.

  According to the tenth aspect of the present invention, it is possible to provide an image processing apparatus that performs a decoding process more efficiently than when the present configuration is not provided.

  According to the present invention of claim 11, it is possible to provide a decoding program that performs a decoding process more efficiently than when the present configuration is not provided.

It is a figure which illustrates the hardware constitutions of the image processing apparatus concerning this embodiment. 2 is a processing program that operates in the image processing apparatus illustrated in FIG. 1. It is a figure which shows the structure of the encoding program concerning 1st Embodiment. It is a figure which illustrates image data code. It is a figure which shows the structure of the decoding program concerning 1st Embodiment. It is a figure which illustrates the process of the decoding program concerning 1st Embodiment. It is a figure which shows the structure of the encoding program concerning 2nd Embodiment. It is a figure which illustrates a prediction system, (A) shows an example of the pixel column made into blocks, (B) shows the prediction system which refers to a block immediately before, (C) is the prediction used by this embodiment. Indicates the method. It is a figure which illustrates the process of a prediction system order | rank encoding part, (A) illustrates a prediction system / ranking correspondence table, (B) illustrates the process with respect to block # 1, (C) illustrates with respect to block # 2 The process is illustrated. It is a figure which illustrates the encoding information which the code | symbol program concerning 2nd Embodiment outputs. It is a figure which shows the structure of the decoding program concerning 2nd Embodiment. It is a figure which illustrates the process of the decoding program concerning 2nd Embodiment.

[First Embodiment]
First, the hardware configuration of the image processing apparatus 2 in the first embodiment will be described.
FIG. 1 is a diagram illustrating a hardware configuration of an image processing apparatus 2 according to the present embodiment.
As illustrated in FIG. 1, the image processing apparatus 2 includes a control device 21 including a calculation unit 212 such as a CPU and a storage unit 214 such as a memory, a communication device 22, a recording device 24, and a user interface device (UI). Device) 25, a printing device 26, and an image reading device 27.

The UI device 25 includes a display device such as an LCD (Liquid Crystal Display) display device or a CRT (Cathode Ray Tube) display device, and a keyboard / touch panel.
The printing device 26 is, for example, a printer, and prints character data or image data on a recording medium such as paper.
The image reading device 27 is a scanner or the like, for example, and reads an image or the like from a recording medium and converts it into image data.
That is, the image processing apparatus 2 has a hardware configuration part as a computer capable of information processing and communication with other image processing apparatuses or terminals.
In the following drawings, substantially the same components and processes are denoted by the same reference numerals.

FIG. 2 shows a processing program 3 that operates in the image processing apparatus 2 shown in FIG.
As shown in FIG. 2, the processing program 3 includes an encoding program 30 and a decoding program 32.
The processing program 3 is supplied to the image processing apparatus 2 via, for example, the storage medium 240 (FIG. 1), loaded into the storage unit 214, and installed on the image processing apparatus 2 on the OS (not shown). It is executed by specifically using the hardware resources of the processing device 2 (the same applies to the software shown below).
In the present embodiment, the encoding program 30 and the decoding program 32 are realized by software, but all or part of the encoding program 30 and the decoding program 32 may be a DSP (Digital Signal Processor) or FPGA. It may be realized by hardware such as (Field Programmable Gate Array).

In the processing program 3, the encoding program 30 performs encoding (compression) processing on the input image data to generate encoding information.
The decoding program 32 performs a decoding (decompression) process on the encoded information to generate a decoded image.

FIG. 3 is a diagram showing a configuration of the encoding program 30 shown in FIG.
As shown in FIG. 3, the encoding program 30 includes an image receiving unit 302, an image data encoding unit 304, a start position encoding unit 306, an image data code output unit 308, and a start position code output unit 310. .
The image receiving unit 302 receives, for example, image data of an image read by the image reading device 27 (FIG. 1) and outputs the image data to the image data encoding unit 304.

The image data encoding unit 304 encodes image data by an image compression method such as JPEG (Joint Photographic Experts Group) format, for example, and an alternating current (AC) component code (data configuration) that is a code in pixel units (data configuration units). A unit code) and a direct current (DC) component code (inter-block correlation information) that correlates with each other between the block including the AC component code and a block adjacent to the block.
Here, a block is a processing unit when image data is encoded (compressed), that is, encoding processing is performed in units of blocks.
For example, the image data encoding unit 304 divides the image data into blocks of a fixed size (for example, 8 × 8 pixels), and uses a discrete cosine transform (DCT) or the like in units of the blocks. To frequency domain.
Further, the image data encoding unit 304 compresses the converted data by performing entropy encoding such as Huffman encoding after reducing the amount of information by quantization.

The direct current component includes information on the luminance of the block and has a high correlation with the adjacent block. Therefore, a DC component code is generated by encoding the difference from the immediately preceding block by DPCM (differential pulse code modulation) or the like, for example.
As for the encoding of AC components, the AC components are arranged in a single signal sequence by zigzag scanning, and encoded by a variable length encoding method such as Huffman encoding.

The start position encoding unit 306 acquires block position information from the image data encoding unit 304, and encodes information related to the start position of the block (start position information).
For example, the start position encoding unit 306 encodes the start address (offset) of the block or encodes the code length of the block.

The image data code output unit 308 outputs the AC component code and the DC component code generated by the image data encoding unit 304 to the decoding program 32.
The start position code output unit 310 outputs the start position information (start position code) encoded by the start position encoding unit 306 to the decoding program 32.
Note that the direct current component and the start position information correlate with other blocks as described above, and thus may be collectively referred to as correlation information in the present embodiment.

FIG. 4 is a diagram illustrating image data codes.
As shown in FIG. 4, the image data code includes a DC component code and a data component unit code.
The data constituent unit code includes AC component codes # 1 to #M (M is an integer equal to or greater than 1. However, M does not always indicate the same number).
Further, the image data code is variable-length coded, and the code length differs in each block, but the decoding program 32 determines the start position of each block using the above-described start position code.
In the present embodiment, the AC component code, the DC component code, and the start position code are output through separate paths, but may be configured to be output collectively.

FIG. 5 is a diagram showing the configuration of the decryption program 32 shown in FIG.
As shown in FIG. 5, the decoding program 32 includes an image data code receiving unit 322, a start position code receiving unit 324, a correlation information decoding unit 33, a distribution unit 340, and buffer units 342-1 to 342-N (N is 1 or more). (However, N does not always indicate the same number), and is composed of data constituent unit decoding units 36-1 to 36-N and an image output unit 370.
In the following description, when any one of a plurality of constituent parts such as the data constituent unit decoding units 36-1 to 36 -N is indicated without being specified, the data constituent unit decoding unit 36 may be simply abbreviated.

In the decoding program 32, the image data code receiving unit 322 receives the image data code from the code program 30 and outputs it to the correlation information decoding unit 33.
The start position code receiving unit 324 receives the start position code from the code program 30 and outputs it to the correlation information decoding unit 33.

The correlation information decoding unit 33 includes a start position information decoding unit 332 and a DC component decoding unit 334.
The correlation information decoding unit 33 decodes the DC component and the start position information (correlation information) using these components.

The start position information decoding unit 332 decodes the start position information code from the start position code reception unit 324, and acquires the start position information (start offset) of the block.
The DC component decoding unit 334 decodes the DC component based on the start position information decoded by the start position information decoding unit 332, and outputs the code length of the DC component to the data component unit decoding unit 36.
The DC component decoding unit 334 performs variable length decoding on the DC difference coefficient by Huffman decoding or the like, and performs addition processing with the DC component of the immediately preceding block, thereby decoding the DC component coefficient.

The distribution unit 340 distributes the data configuration unit code included in each block to the buffer units 342-1 to 342-N.
The buffer unit 342 is, for example, a buffer or the like, and temporarily stores the distributed data unit code until the processing of the downstream data unit decoding unit 36 is completed.

Each of the data structural unit decoding units 36-1 to 36-N includes an AC component head position calculation unit 362, an AC component decoding unit 364, and an IDCT processing unit 366.
For example, the data component unit decoding unit 36-1 includes an AC component head position calculation unit 362-1, an AC component decoding unit 364-1, and an IDCT processing unit 366-1. An AC component head position calculation unit 362-2, an AC component decoding unit 364-2, and an IDCT processing unit 366-2 (not shown) are included.
Based on the correlation information decoded by the correlation information decoding unit 33, the data unit decoding units 36-1 to 36-N decode the data unit code in parallel for each block. Restore the original image data.
That is, each data composition unit decoding unit 36 processes a data composition unit code included in one block.

In the data component unit decoding unit 36, the AC component head position calculation unit 362 calculates the AC component head position (offset) based on the decoded block start position information and the DC component code length.
The AC component decoding unit 364 performs variable-length decoding on the data constituent unit code included in each block based on the calculated head position of the AC component, for example, by Huffman decoding.
The IDCT processing unit 366 performs an inverse discrete cosine transform (IDCT) process on the DC component and the AC component subjected to variable length decoding, and outputs image data (pixel value) of a portion corresponding to the processed block. Restore.

  The image output unit 370 generates original image data by integrating the restored image data of each block, and outputs the original image data to the UI device 25 or the printing device 26.

FIG. 6 is a diagram illustrating the processing of the decryption program 32.
FIG. 6 illustrates the case where the number of data structural unit decoding units 36 is 2 (that is, N = 2) for the sake of concreteness and clarification of explanation, and includes a DC component decoding unit 334, an AC component, and the like. The processes in the decoding units 364-1 and 364-2 and the IDCT processing units 366-1 and 366-2 are illustrated.

The DC component decoding unit 334 performs variable length decoding on the DC component code # 1 of the block # 1 (FIG. 4), performs restoration addition with the DC component of the immediately preceding block, and the decoding of the DC component # 1 is completed.
Thereafter, similarly, the DC component decoding unit 334 performs variable length decoding on the DC component code # 2 of the block # 2, performs restoration addition with the DC component of the immediately preceding block (block # 1), and performs the DC component # 2. 2 decoding is completed.

When the DC component code # 1 is subjected to variable length decoding by the DC component decoding unit 334, the AC component decoding unit 364-1 performs variable length decoding of the AC component code # 1-1, and similarly, the AC component code # 1. Variable length decoding of −2 to # 1-M is performed.
At this time, since the AC component head position is calculated by the AC component head position calculation unit 362-1, the AC component code can be cut out.
On the other hand, when DC component code # 2 is variable-length decoded by DC component decoding unit 334, AC component decoding unit 364-2 performs variable-length decoding on AC component code # 2-1, and so on. Variable length decoding of # 2-2 to # 2-M is performed.

When the AC component code # 1-M is variable-length decoded by the AC component decoding unit 364-1, the IDCT processing unit 366-1 has already decoded the decoded AC components # 1-1 to # 1-M. An inverse discrete cosine transform (IDCT) process is performed on the DC component # 1.
On the other hand, when AC component code # 2-M is variable-length decoded by AC component decoding unit 364-2, IDCT processing unit 366-2 has already decoded AC components # 2-1 to # 2-M and An inverse discrete cosine transform (IDCT) process is performed on the decoded DC component # 2.

  In the present embodiment, as shown in FIG. 6, the process for block # 1 and the process for block # 2 are parallelized, and the processes of the DC component decoding unit 334, the AC component decoding unit 364, and the IDCT processing unit 366 are piped. Lined up.

In the first embodiment, the processing unit is a block, but an MCU (Minimum Coded Unit) including one or more blocks may be a processing unit.
In this case, the start position encoding unit 306 encodes the start address (offset) of the MCU or the code length of the MCU, and the start position information decoding unit 332 decodes the start position information code to start the MCU start position information ( Get start offset).

[Second Embodiment]
Next, a second embodiment will be described.
The processing program of the present embodiment is configured by replacing the encoding program 30 with the encoding program 40 and the decoding program 32 with the decoding program 42 in the processing program of the first embodiment described above. .

FIG. 7 is a diagram showing a configuration of the encoding program 40 according to the second embodiment.
As shown in FIG. 7, the encoding program 40 includes an image receiving unit 302, a block dividing unit 404, a prediction scheme encoding unit 406, a prediction error encoding unit 408, a start position information encoding unit 410, and a code output unit 412. Composed.

The block dividing unit 404 divides the image data from the image receiving unit 302 into blocks that are processing units of the encoding process.
This block includes a predetermined number of pixels (data configuration unit) (for example, if the size is 1 × 4 pixels, it includes 4 pixels).

For each block, the prediction method rank encoding unit 406 predicts the pixel values included in the block using a plurality of prediction methods, and determines the prediction method with the smallest code length (that is, the compression rate is the best). To do.
Also, the prediction method rank encoding unit 406 refers to a preset prediction method / rank correspondence table, identifies the rank (prediction method rank) corresponding to the determined prediction scheme, and codes the rank. Then, a prediction method rank code is generated.
Furthermore, the prediction method rank encoding unit 406 performs differential pulse code modulation (DPCM) processing on each pixel with the determined prediction method, and calculates a prediction error (difference).
Here, the DPCM processing is a method of taking the difference between a pixel to be processed and an adjacent (for example, immediately preceding) pixel and deleting redundancy.
Furthermore, the prediction method order encoding unit 406 updates the prediction method / order correspondence table based on the adopted prediction method.

Hereinafter, the processing of the prediction method rank encoding unit 406 will be further described with reference to FIGS. 8 and 9.
FIG. 8 is a diagram illustrating a prediction method.
In the example shown in FIG. 8A, X1 to X4 indicate the target pixel, and Y1 to Y4, A1 to A4, and C1 to C4 indicate the pixels used when calculating the prediction error.
In the prediction method X referring to the immediately preceding block as shown in FIG. 8B, for example, the prediction error in X1 is calculated as P (X1) = X1-A3, and the prediction error in X2 is P (X2) = Calculated as X2-X1.
Here, in this embodiment, as will be described later, since the decoding process for each block is performed in parallel in the decoding program 42, when this prediction method X is adopted, the first pixel of each block is the target of the processing. When decoding a block, there is a possibility that the process for the last pixel of the immediately preceding block may not be completed, and thus there is a possibility that the prediction error cannot be calculated.
Therefore, in the present embodiment, the prediction error may be calculated using a prediction method illustrated in FIG.

For example, in the prediction method A, the prediction error in the first pixel X1 is calculated as P (X1) = X1-Y1, and the pixel at the corresponding position in the upper block is referred to.
In the prediction method B, the prediction error in the first pixel X1 is calculated as P (X1) = X1- (A3 + Y1-C3), and the pixel at the corresponding position in the diagonally upper block is referred to.
In the prediction method C, the prediction error in the first pixel X1 is calculated as P (X1) = X1-A1, and the first pixel in the immediately preceding block is referred to.
Further, in the prediction method D, the prediction error in the leading pixel X1 is calculated as P (X1) = X1-Y1, and the prediction error in the pixel X2 is also calculated as P (X2) = X2-Y2, and for all the pixels. Reference the corresponding pixel in the upper block.

FIG. 9 is a diagram illustrating processing of the prediction method rank encoding unit 406.
For example, as illustrated in FIG. 9A, a prediction method / order correspondence table in which prediction method A is assigned to rank 1, prediction method B is assigned to rank 2, prediction method C is assigned to rank 3, and prediction method D is assigned to rank 4. Is set.

First, as illustrated in FIG. 9B, the prediction scheme order encoding unit 406 calculates a prediction error for the block # 1 using the prediction schemes A to D, and determines a prediction scheme with the smallest code length. To do.
Here, for example, when the prediction method B is the best, the prediction method rank encoding unit 406 refers to the prediction method / rank correspondence table illustrated in FIG. Identify what to do.
Then, the prediction method rank encoding unit 406 encodes the specified rank 2.
Further, as shown in FIG. 9B, the prediction method rank encoding unit 406 updates the prediction method / rank correspondence table, sets the prediction method B adopted in the processing for the block # 1 to rank 1, and is originally higher. The order of the prediction method A is dropped to rank 2.

Next, as illustrated in FIG. 9C, the prediction scheme order encoding unit 406 calculates a prediction error with the prediction schemes A to D for the block # 2, and selects a prediction scheme with the smallest code length. decide.
Here, for example, when the prediction method B is the best, the prediction method rank encoding unit 406 refers to the updated prediction method / rank correspondence table in FIG. Identify what to do.
Then, the prediction method rank encoding unit 406 encodes the specified rank 1.
Furthermore, the prediction method rank encoding unit 406 updates the prediction method / rank correspondence table as shown in FIG. 9C (however, since the prediction method corresponding to rank 1 is adopted, the rank change is not performed). Not)
In order to encode the rank, a small number of bits is assigned to the higher rank (for example, “0” for the first rank, “01” for the second rank, etc.), so a prediction method corresponding to the higher rank is adopted. If so, the overall code length is shortened, and the compression rate is improved.

The prediction error encoding unit 408 (FIG. 7) performs variable length encoding on the prediction error obtained by the prediction method determined in 406 by the prediction method encoding unit, and generates a prediction error code.
The start position information encoding unit 410 encodes the start address of each divided block and generates a start position code.
The code output unit 412 is generated by the prediction method order code generated by the prediction method encoding unit 406, the prediction error code of each pixel generated by the prediction error encoding unit 408, and the start position information encoding unit 410. The head position code is output to the decoding program 42.

FIG. 10 is a diagram illustrating information (encoded information) output from the code program 40.
As shown in FIG. 10, the encoded information includes header information, correlation information corresponding to each block, and a data configuration unit code.
Here, the header information includes information corresponding to the prediction method / ranking correspondence table before update shown in FIG.

The correlation information is information correlated with other blocks, and includes a start position code and a prediction method rank code.
The data constituent unit code is a set of prediction error codes for pixels (data constituent units) included in each block, and includes prediction error codes # 1 to #M.
Here, M is, for example, the number of pixels included in one block (for example, 4 in the case of 1 × 4 pixels).
Also, the prediction error information is variable-length coded, and the code length differs for each block, but the decoding program 42 determines the start position of each block using the above-described start position code.
In the present embodiment, the prediction method rank code, the start position code, and the prediction error code are output on the same route, but may be configured to be output on different routes.

FIG. 11 is a diagram showing the configuration of the decryption program 42.
As shown in FIG. 11, the decoding program 42 includes a code receiving unit 422, a correlation information decoding unit 43, a distribution unit 340, buffer units 342-1 to 342-N, data configuration unit decoding units 46-1 to 46-N, And an image output unit 370.
In the decoding program 42, the code receiving unit 422 receives the encoded information from the code program 40 and outputs it to the correlation information decoding unit 43.

The correlation information decoding unit 43 includes a start position information decoding unit 432 and a prediction method rank decoding unit 434.
The correlation information decoding unit 43 decodes the prediction method rank information and the start position information using these components.
The start position information decoding unit 432 decodes the start position code from the code receiving unit 422, and acquires start position information (start address) of each block.
The prediction method order decoding unit 434 decodes the prediction method order code from the code receiving unit 422 based on the start position information, and outputs the code length of the prediction method order information to the data component unit decoding unit 46.
When the processing target is the first block (block # 1 in the example of FIG. 10), the prediction method order decoding unit 434 performs decoding with reference to the prediction method / order correspondence table included in the header information shown in FIG. Prediction method information corresponding to the rank is acquired and output to the data component decoding unit 46.
Furthermore, the prediction method order decoding unit 434 updates the prediction method / order correspondence table based on the acquired prediction method information.
In addition, when the processing target is a block other than the top block (in the example of FIG. 10, blocks after block # 2), the prediction method order decoding unit 434 refers to the updated prediction method / order correspondence table and performs the decoding order. Is obtained and output to the data component decoding unit 46.
Further, the prediction method order decoding unit 434 updates the prediction method / order correspondence table based on the acquired prediction method information in the same manner as described above.
As described above, the prediction method order decoding unit 434 refers to the prediction method / order correspondence table updated in the process for the immediately processed block, and acquires prediction method information regarding the processing target block.

Each of the data constituent unit decoding units 46-1 to 46-N includes a prediction error head position calculation unit 462, a prediction error decoding unit 464, and an IDPCM processing unit 466.
For example, the data component unit decoding unit 46-1 includes a prediction error head position calculation unit 462-1, a prediction error decoding unit 464-1, and an IDPCM processing unit 466-1. A prediction error head position calculation unit 462-2, a prediction error decoding unit 464-2, and an IDPCM processing unit 466-2 are included.
Based on the correlation information decoded by the correlation information decoding unit 43, the data unit decoding units 46-1 to 46-N decode the data unit code in parallel for each block. Restore the original image data.

In the data composition unit decoding unit 46, the prediction error start position calculation unit 462 calculates the start position (start address) of the prediction error code based on the decoded start position information and the code length of the prediction scheme order information.
The prediction error decoding unit 464 performs variable-length decoding on the prediction error of each pixel included in one block based on the calculated head position of the prediction error code.
The IDPCM processing unit 466 performs an inverse differential pulse code modulation (IDPCM) process on the decoded prediction error in a prediction method corresponding to the prediction method information from the prediction method order decoding unit 434. The value of each pixel (data configuration unit) is calculated, and the image data (pixel value) of the part corresponding to the processed block is restored.

FIG. 12 is a diagram illustrating the processing of the decryption program 42.
12 illustrates the case where the number of data component unit decoding units 46 is 2 (that is, N = 2) for the sake of concreteness and clarification of the description. The processes in the error decoding units 464-1 and 464-2 and the IDPCM processing units 466-1 and 466-2 are illustrated.

The prediction method order decoding unit 434 acquires prediction method order information by decoding the prediction method order code of block # 1 (FIG. 10), which is the first block, and refers to the prediction method / order correspondence table included in the header information. Then, the prediction method information corresponding to the decoded rank is restored.
Furthermore, the prediction method order decoding unit 434 updates the prediction method / order correspondence table based on the restored prediction method information.
Thereafter, similarly, the prediction method order decoding unit 434 decodes the prediction method order code of block # 2 to obtain prediction method order information.
Also, the prediction method rank decoding unit 434 refers to the prediction method / rank correspondence table updated in the process for the block # 1, restores the prediction method information corresponding to the decoded rank, and based on this prediction method information The prediction method / ranking correspondence table is updated.

The prediction error decoding unit 464-1 decodes the prediction error code # 1-1 when the prediction method rank code corresponding to the block # 1 is decoded by the prediction method rank decoding unit 434, and similarly, the prediction error code Decode # 1-2 to # 1-M.
At this time, since the start position of the prediction error code is calculated by the prediction error start position calculation unit 462-1, the prediction error code can be cut out.
On the other hand, the prediction error decoding unit 464-2 decodes the prediction error code # 2-1 when the prediction method rank code corresponding to the block # 2 is decoded by the prediction method rank decoding unit 434, and so on. Error codes # 2-2 to # 2-M are decoded.

When the prediction error codes # 1-1 to # 1-M are decoded by the prediction error decoding unit 464-1, the IDPCM processing unit 466-1 converts the decoded prediction error into the restored prediction method information. Inverse differential pulse code modulation (IDPCM) processing is performed using a corresponding prediction method.
On the other hand, when the prediction error codes # 2-1 to # 2-M are decoded by the prediction error decoding unit 464-2, the IDPCM processing unit 466-2 restores the predicted prediction method for the decoded prediction error. Inverse differential pulse code modulation (IDPCM) processing is performed using a prediction method corresponding to information.

In this case, as shown in FIG. 12, the process for block # 1 and the process for block # 2 are parallelized, and the processes of the prediction method rank decoding unit 434, the prediction error decoding unit 464, and the IDPCM processing unit 466 are pipelined. Has been.
In the second embodiment described above, the order of prediction methods is encoded (or decoded), but the prediction method information itself may be encoded (or decoded).

Hereinafter, the operation of the present embodiment will be described.
When data is to be efficiently compressed for each processing unit (block), as described above, encoding is performed based on correlation information correlated with each processing unit, but in the conventional example, decoding is performed. If the units are simply parallelized, the data cannot be encoded and decoded using the correlation information when the processing is parallelized for each processing unit.

On the other hand, in the present embodiment, a correlation information decoding unit that decodes correlation information mutually associated with each processing unit and a data constituent unit decoding unit are separated, and a plurality of data constituent unit decoding units are parallelized. Since it is configured, encoding and decoding using correlation information can be realized, and therefore decoding processing can be performed efficiently.
Furthermore, in this embodiment, since the correlation information decoding unit is not parallelized, an increase in the mounting scale can be prevented.
In this embodiment, since the start position information of each processing unit can be decoded and the head position of the data configuration unit can be specified, decoding processing can be parallelized without inserting an EOB code into the encoded information.

2 ... Image processing device,
3 ... Processing program,
30 ... encoding program,
302 ... Image reception unit,
304: Image data encoding unit,
306 ... start position encoding unit,
308: Image data code output unit,
310 ... start position code output unit,
32 ... Decryption program,
322 ... Image data code receiving unit,
324 ... Start position code receiving unit,
33 ... correlation information decoding unit,
332 ... start position information decoding unit,
334 ... DC component decoding unit,
340 ... distribution unit,
342 ... buffer part,
36... Data structure unit decoding unit,
362 ... AC component head position calculation unit,
364 ... AC component decoding unit,
366... IDCT processing unit,
370: Image output unit,
40: Encoding program,
404 ... block division unit,
406 ... Prediction encoding unit,
408 ... Prediction error encoding unit,
410 ... start position information encoding unit,
412 ... code output unit,
42. Decoding program,
422 ... code receiving part,
43 ... correlation information decoding unit,
432 ... start position information decoding unit,
434 ... Prediction method order decoding unit,
46: Data unit decoding unit,
462 ... Prediction error head position calculation unit,
464 ... Prediction error decoding unit,
466... IDPCM processing unit,

Claims (11)

Correlation information decoding means for decoding correlation information mutually correlated between a plurality of processing units each including a plurality of data constituting units constituting data;
Distributing means for distributing the plurality of encoded data constituent units to a plurality of data constituent unit decoding means;
And a plurality of data constituent unit decoding means for decoding each of the distributed data constituent units in parallel based on the correlation information decoded by the correlation information decoding means. The correlation information decoding means decodes information related to a start position of each of the plurality of processing units,
The decoding apparatus according to claim 1, wherein the data component unit decoding unit calculates a start position of a data component unit in the processing unit based on the decoded start position information. The correlation information decoding unit decodes information related to a difference between the processing unit to be processed and the adjacent processing unit,
The decoding device according to claim 1, wherein the data structural unit decoding unit decodes the data structural unit based on the decoded difference information. The correlation information decoding means decodes prediction scheme information indicating which processing unit of a plurality of other processing units is to be encoded with respect to the processing unit to be processed,
The decoding device according to claim 1 or 2, wherein the data constituent unit decoding means decodes the data constituent unit based on the decoded prediction method information. The decoding apparatus according to claim 4, wherein the correlation information decoding unit decodes the prediction method information according to the correlation information corresponding to the processed processing unit for the processing unit to be processed. A data encoding means for encoding a plurality of data constituent units constituting the data to be processed and correlation information correlated with each other between the plurality of processing units including the plurality of data constituent units;
An encoding apparatus comprising: start position encoding means for encoding information relating to a start position of each of the plurality of processing units. A data encoding means for encoding a plurality of data constituent units constituting data to be processed and information relating to a difference between processing units adjacent to the processing units including the plurality of data constituent units;
An encoding apparatus comprising: start position encoding means for encoding information relating to a start position of each of the plurality of processing units. Dividing means for dividing the data to be processed into a plurality of processing units each including a plurality of data constituent units;
For each of the divided processing units, a prediction method encoding unit that encodes prediction method information indicating which processing unit of a plurality of other processing units is to be encoded, and
Prediction error calculation means for calculating a prediction error based on the prediction method information for each of the plurality of data constituent units;
Prediction error encoding means for encoding the calculated prediction error;
An encoding apparatus comprising: start position information encoding means for encoding start position information of the processing unit. The encoding apparatus according to claim 8, wherein the prediction method encoding unit encodes the prediction method information according to the correlation information corresponding to the processed processing unit for the processing unit to be processed. Correlation information decoding means for decoding correlation information that correlates to each other between a plurality of processing units each including a plurality of data constituting units constituting image data;
Distributing means for distributing the plurality of encoded data constituent units to a plurality of data constituent unit decoding means;
A plurality of data constituent unit decoding means for decoding each of the distributed data constituent units in parallel based on the correlation information decoded by the correlation information decoding means;
An image processing apparatus comprising: an image output unit that outputs the image data based on the plurality of data configuration units decoded by the data unit decoding unit. A correlation information decoding step for decoding correlation information correlated with each other between a plurality of processing units each including a plurality of data constituent units constituting data;
A distributing step of distributing the plurality of encoded data constituent units;
A decoding program causing a computer to execute a data composition unit decoding step of decoding each of the distributed data composition units in parallel based on the decoded correlation information.

Images