JP6098703B2 - Image decoding apparatus, image decoding method, image decoding program, receiving apparatus, receiving method, and receiving program - Google Patents

Description

  The present invention relates to an image decoding technique, and more particularly to an entropy decoding technique for a residual signal.

  In MPEG-4 AVC, which is an international standard for moving picture coding, context-switching arithmetic coding called CABAC is adopted as an entropy coding method. CABAC has a plurality of variables called contexts that store the probability of occurrence of information to be encoded. An optimum context is selected from surrounding coding information and used for coding. Also, since the occurrence probability is updated by the encoding process in each context, the occurrence probability of the encoded information can be estimated with high accuracy, and efficient encoding is possible.

JP 2007-300517 A

  In MPEG-4 AVC, in addition to estimation of information occurrence probability by switching contexts based on peripheral decoded information, learning of occurrence probability based on a decoding result is performed. The probability of occurrence of information to be decoded for each context can be optimized, and the encoding efficiency is improved. However, it is necessary to sequentially process the calculation of the context index and the decoding of the significant difference coefficient information with respect to all the significant difference coefficient information in the processing target block, which requires a calculation time.

  Patent Document 1 discloses a technique for reducing processing delays related to decoding by arranging contexts for frequently occurring syntax elements on a memory with low access latency. However, the method of Patent Document 1 does not eliminate the dependency between the calculation of the context index and the decoding of the syntax element, and is not an essential solution to the processing delay in that these cannot be executed in parallel.

  The present invention has been made in view of such a situation, and an object of the present invention is to realize a context index calculation method that can be processed in parallel and has a small amount of calculation in differential coefficient encoding / decoding, and has a simple circuit configuration and real time. An object is to provide an image decoding technique suitable for processing. Another object is to provide an image decoding technique with high encoding efficiency by realizing calculation of a context index with reference to a peripheral difference coefficient appropriate for correlation.

An image decoding device that decodes an encoded stream in which difference information between an image to be decoded and an image to be predicted is divided into a plurality of sub-blocks and the divided sub-blocks are encoded in a predetermined order. A significant sub-block information decoding unit that decodes significant sub-block information indicating whether or not all the values of the difference coefficients belonging to the sub-block are zero, and whether or not the value of the difference coefficient is zero A significant difference coefficient information decoding unit that decodes the significant difference coefficient information indicating the difference coefficient value decoding unit that decodes the value of the difference coefficient, and the significant difference coefficient information of the difference coefficient that belongs to the sub-block to be decoded. based not, based on the sum of the significant difference coefficient information which is adjacent to the sub-blocks composed has and the decoded belongs to decoded sub-blocks adjacent to the sub-blocks to be decoded Te provides an image decoding apparatus characterized by comprising a context deriving unit that derives a context for decoding the significant difference coefficient information of the sub-block to be the decoding target.
An image decoding method in which difference information between an image to be decoded and an image to be predicted is divided into a plurality of subblocks, and an encoded stream in which the divided subblocks are encoded in a predetermined order is decoded. And a significant sub-block information decoding step for decoding significant sub-block information indicating whether or not all the values of the difference coefficients belonging to the sub-block are zero, and whether or not the value of the difference coefficient is zero. A significant difference coefficient information decoding step for decoding the significant difference coefficient information, a difference coefficient value decoding step for decoding the difference coefficient value, and the significant difference coefficient information of the difference coefficient belonging to the sub-block to be decoded. based not, the significant difference that is adjacent to the sub-blocks composed has and the decoded belongs to decoded sub-blocks adjacent to the sub-blocks to be decoded Based on the sum of the number information, to provide an image decoding method characterized by having a context derivation step of deriving a context for decoding the significant difference coefficient information of the sub-block to be the decoding target.
An image decoding program for decoding an encoded stream in which difference information between an image to be decoded and an image to be predicted is divided into a plurality of sub-blocks and the divided sub-blocks are encoded in a predetermined order And a significant sub-block information decoding step for decoding significant sub-block information indicating whether or not all the values of the difference coefficients belonging to the sub-block are zero, and whether or not the value of the difference coefficient is zero. A significant difference coefficient information decoding step for decoding the significant difference coefficient information, a difference coefficient value decoding step for decoding the difference coefficient value, and the significant difference coefficient information of the difference coefficient belonging to the sub-block to be decoded. based not, the chromatic adjacent subblocks composed has and the decoded belongs to decoded sub-blocks adjacent to the sub-blocks to be decoded Provided is an image decoding program that causes a computer to execute a context deriving step for deriving a context for decoding the significant difference coefficient information of the sub-block to be decoded based on a sum of difference coefficient information To do.
A receiving device that receives an encoded stream in which a moving image is encoded and decodes the received encoded stream, and includes difference information between an image to be decoded and an image to be predicted. A receiving unit that receives encoded data obtained by packetizing an encoded stream that is divided into sub-blocks and the divided sub-blocks are encoded in a predetermined order; and the received packetized encoded data And a sub-block information indicating whether or not all of the values of the difference coefficients belonging to the sub-block are all zero from the restored encoded stream. A significant sub-block information decoding unit that decodes the significant difference coefficient information indicating whether or not the value of the difference coefficient is zero from the restored encoded stream Significant difference coefficient information decoding unit, difference coefficient value decoding unit for decoding the difference coefficient value from the restored encoded stream, and the significant difference coefficient information of the difference coefficient belonging to the sub-block to be decoded The sub-block to be decoded based on the sum of the significant difference coefficient information that belongs to the decoded sub-block adjacent to the sub-block to be decoded and is adjacent to the sub-block to be decoded. And a context deriving unit for deriving a context for decoding the significant difference coefficient information.
A reception method for receiving an encoded stream in which a moving image is encoded and decoding the received encoded stream, wherein difference information between an image to be decoded and an image to be predicted includes a plurality of information A receiving step of receiving encoded data obtained by packetizing an encoded stream obtained by dividing the divided sub-blocks into encoded sub-blocks in a predetermined order; and the received packetized encoded data. A restoration step of restoring the encoded stream by packet processing, and significant sub-block information indicating whether or not all the values of the difference coefficients belonging to the sub-block are all zero from the restored encoded stream A significant sub-block information decoding step for decoding and a value of the difference coefficient from the restored encoded stream indicates whether or not Significant difference coefficient information decoding step for decoding the difference coefficient information, a difference coefficient value decoding step for decoding the value of the difference coefficient from the restored encoded stream, and a difference coefficient belonging to the sub-block to be decoded The decoding based on the sum of the significant difference coefficient information that belongs to the decoded subblock adjacent to the decoding target subblock and that is adjacent to the decoding target subblock , not based on the significant difference coefficient information. And a context deriving step for deriving a context for decoding the significant difference coefficient information of the target sub-block.
A reception program that receives an encoded stream in which a moving image is encoded and decodes the received encoded stream, and includes difference information between an image to be decoded and an image to be predicted. A receiving step of receiving encoded data obtained by packetizing an encoded stream obtained by dividing the divided sub-blocks into encoded sub-blocks in a predetermined order; and the received packetized encoded data. A restoration step of restoring the encoded stream by packet processing, and significant sub-block information indicating whether or not all the values of the difference coefficients belonging to the sub-block are all zero from the restored encoded stream Whether or not the value of the difference coefficient is zero from the significant sub-block information decoding step of decoding and the restored encoded stream A significant difference coefficient information decoding step for decoding the indicated significant difference coefficient information, a difference coefficient value decoding step for decoding the value of the difference coefficient from the restored encoded stream, and a difference coefficient belonging to a sub-block to be decoded wherein not based on significant difference coefficient information, based on a sum of said significant difference coefficient information which is adjacent to the sub-blocks composed has and the decoded belongs to decoded sub-blocks adjacent to the sub-block to be decoded of said There is provided a reception program that causes a computer to execute a context deriving step for deriving a context for decoding the significant difference coefficient information of a sub-block to be decoded.

  It should be noted that any combination of the above-described constituent elements and a conversion of the expression of the present invention between a method, an apparatus, a system, a recording medium, a computer program, etc. are also effective as an aspect of the present invention.

  ADVANTAGE OF THE INVENTION According to this invention, the circuit structure is simple and the decoding of the difference signal suitable for real time processing is realizable.

It is a flowchart for demonstrating the decoding procedure of the conventional difference coefficient. It is a flowchart for demonstrating the decoding procedure of the conventional subblock difference coefficient. It is a flowchart for demonstrating the decoding procedure of the conventional significant difference coefficient. It is a flowchart for demonstrating the decoding procedure of the conventional difference coefficient value. It is a block diagram which shows the structure of the image coding apparatus for performing the encoding method of the difference coefficient which concerns on embodiment. It is a block diagram which shows the structure of the image decoding apparatus for performing the decoding method of the difference coefficient which concerns on embodiment. It is a figure for demonstrating the scanning order of a subblock and a difference coefficient. It is a block diagram which shows the detailed structure of the 1st Example of the image decoding apparatus of FIG. It is a figure for demonstrating the definition of the periphery difference coefficient in the decoding procedure of the significant difference coefficient of FIG. It is a figure for demonstrating the definition of the periphery difference coefficient in the decoding procedure of the significant difference coefficient of FIG. It is a figure for demonstrating the context definition in the decoding procedure of the significant difference coefficient of FIG. It is a figure for demonstrating the subblock division | segmentation of a difference coefficient. It is a flowchart for demonstrating the decoding procedure of the difference coefficient value in a 1st Example. It is a figure for demonstrating an encoding block size. It is a block diagram which shows the detailed structure of the 1st Example of the image coding apparatus of FIG. It is a flowchart for demonstrating the encoding procedure of the difference coefficient in a 1st Example. It is a flowchart for demonstrating the encoding procedure of the subblock difference coefficient in 1st Example. It is a flowchart for demonstrating the encoding procedure of the significant difference coefficient in a 1st Example. It is a flowchart for demonstrating the encoding procedure of the difference coefficient value in a 1st Example. It is a figure for demonstrating the structure which includes a subblock position in the context calculation of significant difference coefficient information.

  First, a technique that is a premise of the embodiment of the present invention will be described.

  A method of associating a plurality of contexts with each coding syntax and selecting a context based on the correlation of syntax elements can optimize code allocation, and thus can perform efficient coding.

  As an example of context-switching entropy encoding, a procedure for decoding quantized orthogonal transform coefficients of a differential signal encoded in 16 × 16 size will be described using the flowchart of FIG. FIG. 12 shows quantized orthogonal transform coefficients to be processed. Hereinafter, the quantized orthogonal transform coefficient is referred to as a difference coefficient. In this procedure, the 16 × 16 difference coefficient to be processed is divided into 4 × 4 size sub-blocks 401 to 416, and scanning in units of sub-blocks is preferentially performed.

  A sub-block to be processed is determined according to the scanning order described later (S101). If all the sub-blocks have been scanned, the differential coefficient decoding process ends. Reference numeral 902 in FIG. 7 shows the scanning order of the sub-blocks. In this procedure, scanning is started from the lower right sub-block of the difference coefficient area, scanning is performed according to the rule from lower right to upper left, and further from upper right to upper left, and scanning is performed in the upper left sub-block. Finish. Reference numeral 901 in FIG. 7 represents the sub-block scanning order using arrows. When the scanning order shown in FIG. 7 is followed, in all the processing target sub-blocks, scanning of the sub-blocks spatially located on the right side and the lower side is completed.

  Returning to the flowchart of FIG. 1, decoding processing is performed on all the difference coefficient values of the processing target sub-block (S102). After decoding of the sub-block difference coefficient value is completed, the process proceeds to step S101.

  Details of the sub-block difference coefficient value decoding processing will be described with reference to the flowchart of FIG.

  Significant sub-block information is decoded (S201). The significant sub-block information is a 1-bit flag for indicating that a non-zero difference coefficient exists in the processing target sub-block. Significant sub-block information of 1 indicates that there is at least one non-zero difference coefficient in the processing target sub-block. Significant sub-block information of 0 indicates that all the difference coefficients of the processing target sub-block are 0.

  The value of significant sub-block information is determined (S202). When the significant sub-block information is 0, all difference coefficient values of the processing target sub-block are set to 0 (S209), and the sub-block difference coefficient value decoding process is terminated.

  When the significant sub-block information is 1, all significant difference coefficient information of the processing target sub-block is decoded (S203). The significant difference coefficient information is a 1-bit flag for indicating that the difference coefficient value at the processing target position is not 0. Significant coefficient information of 1 indicates that the difference coefficient value at the processing target position is not 0, and significance coefficient information of 0 indicates that the difference coefficient value at the processing target position is 0. Details of the decoding procedure of the sub block significant difference coefficient information will be described later. After decoding of all significant difference coefficient information of the sub-block, the process proceeds to decoding of the difference coefficient value in step S204.

  The differential coefficient value is decoded (S204). Details of the differential coefficient value decoding process will be described later. After completing the decoding process of the difference coefficient value, the process proceeds to step S101, and the next sub-block is scanned.

[Decoding procedure of significant difference coefficient information]
The decoding procedure of the sub block significant difference coefficient information in step S203 will be described with reference to the flowchart of FIG.

  A sub-block to be processed is determined according to a predetermined scanning order (S301). It is assumed that the scanning order of the difference coefficients in the sub-block follows the rules shown in FIG.

  A marginal significant difference coefficient sum countCoeff, which is the sum of the number of difference coefficients that are adjacent to the processing target difference coefficient position and has been decoded, is calculated (S302). FIG. 9 shows an example of the difference coefficient position for calculating the peripheral significant difference coefficient sum countCoeff. Reference numeral 202 denotes a peripheral difference coefficient when the processing target position is the reference numeral 201, and reference numeral 204 denotes a peripheral difference coefficient when the processing target position is the reference numeral 203. As shown in FIG. 9, five difference coefficients that are on the right side and the lower side of the processing target difference coefficient position and adjacent to the processing target difference coefficient position are set as peripheral difference coefficients. Since the scanning order of the difference coefficients follows FIG. 7, the difference coefficients that belong to the same sub-block as the processing target difference coefficient and that are on the right side and the lower side of the processing target difference coefficient position have been decoded. Similarly, significant difference coefficients belonging to sub-blocks on the right side and lower side of the sub-block to which the processing target position belongs have also been decoded. The peripheral difference coefficient sum countCoeff is a variable for estimating the occurrence probability of a significant difference coefficient. In terms of image characteristics and visual characteristics, significant difference coefficients tend to concentrate 1 in the low frequency range and concentrate 0 in the high frequency range. Since the significant difference coefficient has a spatial correlation, the difference coefficient adjacent to the processing target position is set as the calculation target of the peripheral difference coefficient sum countCoeff. The peripheral difference coefficient indicating the outside of the difference coefficient area is excluded from the calculation of the peripheral significant coefficient sum countCoeff.

  Returning to the flowchart of FIG. 3, it is determined whether the marginal significant coefficient sum countCoeff is 0 (S303). When the peripheral significant coefficient sum countCoeff is 0, the context index ctxIdx for decoding the significant difference coefficient information is set to 0 (S304), and the significant difference coefficient information is decoded using the context corresponding to the context index ctxIdx. . Then, significant difference coefficient information is set in the difference coefficient value (S308).

  If the marginal significant coefficient sum countCoeff is not 0, it is determined whether the marginal significant coefficient sum countCoeff is 2 or less (S305). If the peripheral significant coefficient sum countCoeff is 2 or less, the context index ctxIdx for decoding the significant difference coefficient information is set to 1 (S306), and the significant difference coefficient information is decoded using the context corresponding to the context index ctxIdx. To do. Then, significant difference coefficient information is set in the difference coefficient value (S308).

  When the marginal significant coefficient sum countCoeff is not 2 or less, that is, when the marginal significant coefficient sum countCoeff is 3 or more, the context index ctxIdx for decoding the significant difference coefficient information is set to 2 (S307), and the context index ctxIdx Significant difference coefficient information is decoded using the context corresponding to. Then, significant difference coefficient information is set in the difference coefficient value (S308).

  The context is a variable for storing the occurrence probability of the information to be decoded, and the codeword assignment is switched based on the occurrence probability indicated by the context. In the above example, three contexts for encoding significant difference coefficients are defined, and the context for decoding significant difference coefficients is determined by the size of the peripheral significant difference coefficient sum. The context corresponding to the context index ctxIdx = 0 when the peripheral significant coefficient sum countCoeff is 0 has a high occurrence probability of significant coefficient information of 0, and the context when the peripheral significant coefficient sum countCoeff is 3 or more. In the context corresponding to the index ctxIdx = 2, the occurrence probability of significant coefficient information of 1 is set high. For information with a high probability of occurrence, the amount of codes can be shortened, so that the coding efficiency can be improved by increasing the accuracy of estimating the probability of occurrence.

  In MPEG-4 AVC, in addition to estimation of information occurrence probability by switching contexts based on peripheral decoded information, learning of occurrence probability based on a decoding result is performed. The probability of occurrence of information to be decoded for each context can be optimized, and the encoding efficiency is improved.

  In general, information tends to concentrate on the low-frequency range of the orthogonal transform component of an image. In addition, since the deterioration of the high frequency component has little influence on the visual characteristics, the high frequency component is often roughly quantized in practice. Therefore, the significant coefficient information tends to concentrate on the low frequency components. Significant coefficient information has a high correlation with surrounding significant coefficients, and it is reasonable from the viewpoint of coding efficiency to switch contexts depending on the number of surrounding significant coefficient information.

[Difference coefficient value decoding process]
The decoding procedure of the difference coefficient value of the sub-block in step S204 in the flowchart of FIG. 2 will be described using the flowchart of FIG.

  A sub-block to be processed is determined according to a predetermined scanning order (S501). It is assumed that the scanning order of the difference coefficients within the sub-block follows the rules shown in FIG. If all the difference coefficients of the sub-block have been scanned, the decoding process of the difference coefficient value is completed, and the process proceeds to the next sub-block determination procedure (S101).

  It is determined whether or not the difference coefficient value at the processing target difference coefficient position is 0 (S502). When the difference coefficient value at the processing target difference coefficient position is 0, the decoding of the difference coefficient value at the processing target difference coefficient position is completed, and the process proceeds to step S501.

  When the difference coefficient value at the processing target difference coefficient position is 1, the absolute value of the difference coefficient at the processing target difference coefficient position is decoded (S503). When this procedure is executed, it is determined that the difference coefficient value is not 0, and a code word corresponding to a value obtained by subtracting 1 from the absolute value of the difference coefficient is encoded as an encoded sequence. Therefore, a value obtained by adding 1 to the value obtained by entropy decoding the code word is set as the absolute value of the difference coefficient.

  The sign of the difference coefficient at the processing target difference coefficient position is decoded (S504). The difference coefficient value is determined from the absolute value of the difference coefficient and the sign of the difference coefficient.

  In the significant difference coefficient information decoding procedure, 201 in FIG. 9 is scanned last in the sub-block as shown in the scanning order of 902 in FIG. 7, and the scanning order is 16 as shown in 902 in FIG. Of the peripheral difference coefficients 202 of 201, the scanning order of the position adjacent to 201 is 15, and scanning is performed immediately before 201. Since the context index ctxIdx necessary for decoding 201 significant difference coefficient information is calculated based on the sum of 202 significant difference coefficient information, until the decoding of 202 significant difference coefficient information is completed, the context of 201 Intex ctxIdx cannot be determined. This requires that the calculation of ctxIdx and the decoding of the significant difference coefficient information be sequentially processed for all significant difference coefficient information in the sub-block, and the time calculation amount cannot be reduced by parallelization. Means that. On the other hand, the difference coefficient occupies a high ratio in the encoded sequence, and the context index calculation process and the decoding process of the significant difference coefficient information occupy a large amount of time calculation in the entire decoding process. That is, the decoding process of significant coefficient information becomes the largest bottleneck in the real-time decoding process.

  Patent Document 1 discloses a technique for reducing processing delays related to decoding by arranging contexts for frequently occurring syntax elements on a memory with low access latency. However, the method of Patent Document 1 does not eliminate the dependency between the calculation of the context index and the decoding of the syntax element, and is not an essential solution to the processing delay in that these cannot be executed in parallel.

  Therefore, in the embodiment of the present invention, in the difference coefficient encoding / decoding, the dependency between the calculation of the context index and the encoding / decoding of the significant difference coefficient information is eliminated, and the context index can be processed in parallel and has a small amount of calculation. An image encoding technique that realizes a calculation method, has a simple circuit configuration, and is suitable for real-time processing is provided. In addition, an image encoding technique with high encoding efficiency is provided by realizing calculation of a context index with reference to a peripheral difference coefficient appropriate for correlation. Embodiments of the present invention will be described below.

  In the following description, “processing target block” refers to an encoding target block in the case of encoding processing by an image encoding device, and refers to a decoding target block in the case of decoding processing by an image decoding device. is there. “Processed block” refers to a decoded block that has been encoded in the case of encoding processing by the image encoding device, and a decoded block in the case of decoding processing by the image decoding device. is there. Hereinafter, unless otherwise noted, this meaning is used.

[Encoding device]
A preferred image encoding apparatus for carrying out the present invention will be described with reference to the drawings. FIG. 5 is a block diagram showing a configuration of the image coding apparatus according to the embodiment. The image coding apparatus according to the embodiment includes a subtracting unit 501, an orthogonal transform / quantization unit 502, an inverse quantization / inverse transform unit 503, an adder 504, a decoded image memory 505, a prediction unit 506, A difference information encoding unit 507, a prediction information encoding unit 508, and a mode determination unit 509 are provided.

  The mode determination unit 509 tries to encode all the prediction candidates and determines optimal prediction information for each block of the image. The prediction information includes a prediction mode indicating a divided block size and inter-frame prediction / intra-screen prediction. Further, when the prediction mode is inter-screen prediction, motion information such as a motion vector and a reference picture index is included in the prediction mode. In the case of intra prediction, the intra prediction mode is included. The determined prediction information is provided to the prediction unit 506 and the prediction information encoding unit 508.

  The prediction information coding unit 508 performs variable length coding on the input prediction information and outputs a coded sequence of prediction information.

  The prediction unit 506 generates a prediction image using the input prediction information and the already decoded image stored in the decoded image memory 505, and provides the generated prediction image to the subtraction unit 501.

  The subtraction unit 501 generates a difference image by subtracting the predicted image from the original image to be encoded, and provides the generated difference signal to the orthogonal transform / quantization unit 502.

  The orthogonal transformation / quantization unit 502 performs orthogonal transformation / quantization on the difference image to generate a difference coefficient, and provides the generated difference coefficient to the inverse quantization / inverse transformation unit 503 and the difference information encoding unit 507.

  The difference information encoding unit 507 entropy encodes the difference coefficient and outputs an encoded sequence of difference information.

  The inverse quantization / inverse transform unit 503 generates a decoded differential signal by performing inverse quantization / inverse orthogonal transformation on the difference coefficient received from the orthogonal transform / quantization unit 502, and adds the generated decoded differential signal to the addition unit 504. To give.

  The adding unit 504 generates a decoded image by adding the predicted image and the decoded difference signal, and stores the generated decoded image in the decoded image memory 505.

[Decoding device]
A preferred image decoding apparatus for implementing the present invention will be described with reference to the drawings. FIG. 6 is a block diagram showing the configuration of the video decoding apparatus according to the embodiment. The image decoding apparatus according to the embodiment includes a difference information decoding unit 801, an inverse quantization / inverse conversion unit 802, a prediction information decoding unit 803, an addition unit 804, a decoded image memory 805, and a prediction unit 806. .

  Since the decoding process of the image decoding apparatus in FIG. 6 corresponds to the decoding process provided in the image encoding apparatus in FIG. 5, the inverse quantization / inverse conversion unit 802 and the addition unit 804 in FIG. 8. Each of the configurations of the decoded image memory 805 and the prediction unit 806 is the same as each configuration of the inverse quantization / inverse conversion unit 503, the addition unit 504, the decoded image memory 505, and the prediction unit 506 of the image encoding device in FIG. Has a corresponding function.

  The prediction information decoding unit 803 generates prediction information by entropy decoding the input prediction information encoded sequence, and provides the generated prediction information to the prediction unit 806.

  The prediction unit 806 generates a prediction image using the input prediction information and the already decoded image stored in the decoded image memory 805, and gives the generated prediction image to the addition unit 804.

  The difference information decoding unit 801 generates difference information by entropy decoding the difference information. The generated difference information is given to the inverse quantization / inverse transform unit 802.

  The inverse quantization / inverse transform unit 802 performs inverse quantization / inverse orthogonal transform on the difference information received from the difference information decoding unit 801 to generate a decoded differential signal, and provides the generated decoded differential signal to the adding unit 804 .

  The adding unit 804 adds the predicted image and the decoded difference signal to generate a decoded image, stores the generated decoded image in the decoded image memory 805, and outputs the decoded image.

  The difference coefficient encoding and decoding processing according to the embodiment of the present invention is performed in the difference information encoding unit 507 of the moving image encoding device in FIG. 5 and the difference information decoding unit 801 of the moving image decoding device in FIG. . The details of the difference information encoding and decoding process according to the embodiment will be described below.

[Encoding block]
In the embodiment, as shown in FIG. 14, the screen is hierarchically divided into rectangular blocks, and each block is sequentially processed in a predetermined processing order. Each block to be divided is called a coding block. A block 1817 in FIG. 18 is a maximum unit of division in the embodiment, and this is called a maximum encoding block. A block 1816 in FIG. 14 is a minimum unit of division in the embodiment, and this is called a minimum coding block. In the following description, the minimum encoding block is 4 × 4 pixels and the maximum encoding block is 16 × 16 pixels.

[Predicted block]
Of the encoded blocks, a unit for performing intra prediction is called a prediction block. The prediction block has a size of not less than the minimum coding block and not more than the maximum coding block. In FIG. 14, blocks 1802, 1803 and 1804 are 16 × 16 blocks, blocks 1805, 1810, 1811 and 1801 are 8 × 8 blocks, and blocks 1806, 1807, 1808 and 1809 are 4 × 4 blocks. Blocks 1812, 1813, 1814, and 1815 are unprocessed blocks, and the encoding block size is not fixed. In the encoding procedure, an optimal prediction block size is determined, and the prediction block size is encoded. In the decoding procedure, the predicted block size is obtained from the bit stream. Hereinafter, the prediction block is described as a processing unit.

[Difference coefficient processing unit]
The unit for performing the quantization / orthogonal transform is the same as the unit of the prediction block. However, in the encoding / decoding process, the difference coefficient region is divided into a plurality of sub-blocks and scanned. The size of the sub-block is 4 × 4 size. FIG. 12 shows a 16 × 16 size difference coefficient region. 401 to 416 are sub-blocks. However, the unit for performing the quantization / orthogonal transformation may be determined independently of the unit of the prediction block.

(First embodiment)
[Encoding procedure]
A first example of the differential information encoding method according to the embodiment of the present invention will be described. FIG. 15 is a block diagram of a detailed configuration of the first embodiment of the difference information encoding unit 507 of FIG. The difference information encoding unit 507 according to the first embodiment includes an arithmetic encoding unit 701, a difference coefficient buffer 702, an encoding control unit 703, a context memory 704, and a scanning control unit 705. The encoding control unit 703 further includes: , A significant coefficient information encoding control unit 706, a differential coefficient value encoding control unit 707, and a significant sub-block information encoding control unit 708.

  Hereinafter, the difference coefficient encoding procedure will be described with reference to the flowcharts of FIGS. 16, 17, 18, and 19.

  The scanning control unit 705 determines a processing target sub-block (S601). If all the sub-blocks have been scanned, the differential coefficient decoding process ends. The scanning order of the sub-blocks is shown at 902 in FIG. In this procedure, scanning is started from the lower right sub-block of the difference coefficient area, scanning is performed according to the rule from lower right to upper left, and further from upper right to upper left, and scanning is performed in the upper left sub-block. Finish. As described above, the context is updated by the encoding process. Taking this scan order is a process of improving the estimation accuracy of the occurrence probability of the low frequency component difference coefficient by performing the coding of the low frequency component in which the differential coefficient is likely to occur after the high frequency component. There are advantages. Reference numeral 901 in FIG. 7 represents the sub-block scanning order using arrows. When the scanning order shown in FIG. 7 is followed, scanning of sub-blocks spatially located on the right side and the lower side of the processing target sub-block is completed. A sub-block encoding process is performed on the processing target sub-block (S602).

[Sub-block encoding procedure (S602)]
The significant subblock information encoding control unit 708 acquires the processing target subblock from the difference coefficient buffer 702. When all the difference coefficients of the sub-block are scanned and all the difference coefficient values are 0, the significant sub-block information is set to 0. Otherwise (when there is at least one non-zero difference coefficient value), significant sub-block information is set to 1 (S701).

  The significant sub-block information encoding control unit 708 refers to the difference coefficient included in the sub-block that is adjacent to the processing target sub-block and has been decoded from the difference coefficient buffer 702, and encodes the significant sub-block information. Determine the context index ctxIdx. A context corresponding to the context index ctxIdx is read from the context memory 704. Significant sub-block information and context are sent to the arithmetic coding unit 701. The arithmetic encoding unit 701 encodes significant sub-block information using the context (S702).

  The significant subblock information encoding control unit 708 determines the value of significant subblock information (S703). If the significant sub-block information is 0, the sub-block difference coefficient value encoding process is terminated, and the process proceeds to step S601.

  When the significant sub-block information is 1, all significant difference coefficient information of the processing target sub-block is encoded (S704). Details of the encoding procedure of the significant difference coefficient information will be described later. After encoding all significant difference coefficient information of the sub-block, the process proceeds to encoding of difference coefficient values in step S704.

  The difference coefficient value encoding control unit 707 performs an encoding process on all the difference coefficient values of the processing target sub-block (S705). Details of the encoding procedure of the sub block difference coefficient values will be described later. After encoding all the difference coefficient values of the sub-block, the process proceeds to step S601.

[Significant Difference Coefficient Information Encoding Processing Procedure (S704)]
The significant coefficient information encoding control unit 706 calculates the sum of the number of non-zero difference coefficients around the processing target sub-block, that is, the peripheral significant coefficient sum countCoeff (S801). In this procedure, a difference coefficient that belongs to a sub-block spatially on the right side and lower side of the processing target sub-block and is adjacent to the processing target sub-block is defined as a peripheral difference coefficient.

  FIG. 10 shows the peripheral difference coefficient positions. Reference numeral 301 denotes a processing target sub-block, and reference numeral 302 denotes a peripheral difference coefficient. The peripheral difference coefficient indicating the outside of the difference coefficient area is excluded from the calculation of the peripheral significant coefficient sum countCoeff. The difference coefficient 303 belonging to the right side and lower side sub-block of the processing target sub-block can take either a configuration included in the peripheral difference coefficient or a configuration not included. In the configuration in which the code 303 is included in the peripheral difference coefficient, the number of the peripheral difference coefficients is increased, and the significant difference coefficient information occurrence probability can be estimated with high accuracy. In the configuration in which the code 303 is not included in the peripheral difference coefficient, The amount of calculation and the circuit scale can be reduced by reducing the addition processing related to countCoeff and the boundary determination processing of the difference coefficient region.

  The significant coefficient information encoding control unit 706 determines a difference coefficient to be processed (S802). The scanning order of the difference coefficients in the sub-block follows the rules shown in FIG. 7 as in the scanning order of the sub-blocks in the difference coefficient area. If all the significant difference coefficients of the sub-block have been scanned, the encoding process of the significant difference coefficient is terminated, and the process proceeds to the difference coefficient value encoding procedure (S704).

  The significant coefficient information encoding control unit 706 determines whether the marginal significant coefficient sum countCoeff is 0 (S803).

  If the peripheral significant coefficient sum countCoeff is 0, the processing target difference coefficient position in the processing target sub-block is determined (S804). The horizontal difference coefficient position is posX, the vertical difference coefficient position is posY, and the processing target difference coefficient position is pos = posX + posY. If pos <= 2, the context index ctxIdx for encoding significant coefficient information is set to 1 (S805). Otherwise (pos> 2), the context index ctxIdx is set to 0 (S806). . The definition of the context index ctxIdx when countCoeff = 0 is indicated by reference numeral 601 in FIG.

  If the marginal significant coefficient sum countCoeff is not 0, it is determined whether the marginal significant coefficient sum countCoeff is 1 or less (S807). When the peripheral significant coefficient sum countCoeff is 1 or less, the processing target difference coefficient position in the processing target sub-block is determined (S408). If pos <= 3, the context index ctxIdx for decoding significant coefficient information is set to 1 (S809). Otherwise (pos> 3), the context index ctxIdx is set to 0 (S810). The definition of the context index ctxIdx when countCoeff = 1 is indicated by reference numeral 602 in FIG.

  If the marginal significant coefficient sum countCoeff is not 1 or less, it is determined whether the marginal significant coefficient sum countCoeff is 2 or less (S811). If the peripheral significant coefficient sum countCoeff is 2 or less, the processing target difference coefficient position in the processing target sub-block is determined (S812). If pos <= 2, the context index ctxIdx for decoding significant coefficient information is set to 2 (S813). Otherwise (pos> 2), the context index ctxIdx is set to 1 (S814). The definition of the context index ctxIdx when countCoeff = 2 is indicated by reference numeral 603 in FIG.

  If the marginal significant coefficient sum countCoeff is not 2 or less, the context index ctxIdx for decoding significant coefficient information is set to 2 (S815). The definition of the context index ctxIdx when countCoeff> 2 is indicated by reference numeral 605 in FIG.

  The significant coefficient information encoding control unit 706 acquires the difference coefficient at the processing target position from the difference coefficient buffer 702. If the difference coefficient value is not 0, the significant difference coefficient information is set to 1, and if not (when the difference coefficient value is 0), the significant difference coefficient information is set to 0 (S816).

  The significant coefficient information encoding control unit 706 reads the context corresponding to the determined context index ctxIdx from the context memory 704, and then sends the significant difference coefficient information and the context to the arithmetic encoding unit 701. The arithmetic encoding unit 701 encodes significant difference coefficient information using the context (S817).

[Difference coefficient value encoding process (S705)]
The difference coefficient value encoding control unit 707 determines a difference coefficient to be processed (S901). It is assumed that the scanning order of the difference coefficients within the sub-block follows the rules shown in FIG. 7 in the same manner as the scanning order of the significant difference coefficients. If all the difference coefficients of the subblock have been scanned, the encoding process of the difference coefficient value ends, and the process proceeds to the next subblock determination procedure (S601).

  The difference coefficient value encoding control unit 707 determines whether or not the difference coefficient value at the processing target difference coefficient position is 0 (S902). If the difference coefficient value at the processing target difference coefficient position is 0, the encoding of the difference coefficient value at the processing target difference coefficient position is completed, and the process proceeds to step S901.

  If the difference coefficient value at the processing target difference coefficient position is not 0, the encoded difference coefficient absolute value and the code at the processing target difference coefficient position are calculated (S903, S904). Since it is determined that the difference coefficient value is not 0 when this procedure is executed, the encoded difference coefficient absolute value is a value obtained by subtracting 1 from the absolute value of the difference coefficient. If the difference coefficient is positive, the sign is set to 0, and if the difference coefficient is negative, the sign is set to 1.

  The difference coefficient value encoding control unit 707 reads the context from the context memory 704 and then sends the encoded absolute value and the context to the arithmetic encoding unit 701. The arithmetic encoding unit 701 encodes the encoded absolute value using the context (S905).

  The difference coefficient value encoding control unit 707 reads the context from the context memory 704 and then sends the code and context to the arithmetic encoding unit 701. The arithmetic encoding unit 701 encodes the encoded absolute value using the context (S905).

[Decryption procedure]
A first example of the differential coefficient decoding method according to the embodiment of the present invention will be described. FIG. 8 is a block diagram of a detailed configuration of the first embodiment of the difference information decoding unit 801 in FIG. The difference information decoding unit 801 according to the first embodiment includes an arithmetic decoding unit 1001, a difference coefficient buffer 1002, a decoding control unit 1003, a context memory 1004, and a scanning control unit 1005. The decoding control unit 1003 further includes significant coefficient information. A decoding control unit 1006, a difference coefficient value decoding control unit 1007, and a significant sub-block information decoding control unit 1008 are provided.

  Since the difference information decoding process in the difference information decoding unit 801 in FIG. 8 corresponds to the difference information encoding process in the difference information encoding unit 507 in FIG. 5, the difference coefficient buffer in the difference information encoding unit in FIG. Each configuration of 1002, the context memory 1004, and the scan control unit 1005 has a function corresponding to each configuration of the difference coefficient buffer 702, the context memory 704, and the scan control unit 705 in FIG.

  The difference information decoding procedure will be described below with reference to the flowcharts of FIGS. 1, 2, 4, and 13.

  The scanning control unit 1005 determines a processing target sub-block (S101). If all the sub-blocks have been scanned, the differential coefficient decoding process ends. The scanning order of the sub-blocks is shown at 902 in FIG. In this procedure, scanning is started from the lower right sub-block of the difference coefficient area, scanning is performed according to the rule from lower right to upper left, and further from upper right to upper left, and scanning is performed in the upper left sub-block. Finish. In FIG. 7, reference numeral 901 represents the sub-block scanning order using arrows. When the scanning order shown in FIG. 7 is followed, scanning of sub-blocks spatially located on the right side and the lower side of the processing target sub-block is completed. A sub-block decoding process is performed on the processing target sub-block (S102).

[Sub-block decoding (S102)]
The significant sub-block information decoding control unit 1008 refers to the difference coefficient included in the sub-block that is adjacent to the processing target sub-block and has been decoded from the difference coefficient buffer 1002, and sets a context for decoding the significant sub-block information. The determined context is read from the context memory 1004. A decoding instruction is sent together with the context to the arithmetic decoding unit 1001. The arithmetic decoding unit 1001 performs decoding processing of the encoded sequence using the context, and decodes significant sub-block information (S201).

  The significant sub-block information decoding control unit 1008 determines the value of significant sub-block information (S202). When the significant subblock information is 0, all the difference coefficient values of the processing target subblock of the difference coefficient buffer 1002 are set to 0 (S209), and the subblock difference coefficient value decoding process is terminated.

  When the significant sub-block information is 1, all significant difference coefficient information of the processing target sub-block is decoded (S203). Details of the decoding procedure of the sub block significant difference coefficient information will be described later. After decoding of all significant difference coefficient information of the sub-block, the process proceeds to decoding of the difference coefficient value in step S204.

  Decoding processing is performed on all the difference coefficient values of the processing target sub-block (S204). Details of the decoding procedure of the sub block difference coefficient value will be described later. After decoding all the difference coefficient values of the sub-block, the process proceeds to step S101.

[Procedure for Decoding Significant Difference Coefficient Information (S203)]
The significant coefficient information decoding control unit 1006 calculates the sum countCoeff of the number of significant difference coefficients around the processing target difference coefficient position (S401). In this procedure, a difference coefficient that belongs to a sub-block spatially on the right side and lower side of the processing target sub-block and is adjacent to the processing target sub-block is defined as a peripheral difference coefficient.

  FIG. 10 shows the peripheral difference coefficient positions. Reference numeral 301 denotes a processing target sub-block, and reference numeral 302 denotes a peripheral difference coefficient. The peripheral difference coefficient indicating the outside of the difference coefficient area is excluded from the calculation of the peripheral significant coefficient sum countCoeff. The difference coefficient 303 belonging to the right side and lower side sub-block of the processing target sub-block can take either a configuration included in the peripheral difference coefficient or a configuration not included. In the configuration in which the code 303 is included in the peripheral difference coefficient, the number of the peripheral difference coefficients is increased, and the significant difference coefficient information occurrence probability can be estimated with high accuracy. In the configuration in which the code 303 is not included in the peripheral difference coefficient, The amount of calculation and the circuit scale can be reduced by reducing the addition processing related to countCoeff and the boundary determination processing of the difference coefficient region.

  The significant coefficient information decoding control unit 1006 determines a difference coefficient to be processed (S402). It is assumed that the scanning order of the difference coefficients in the sub-block follows the rules shown in FIG. 7 as the scanning order of the sub-blocks in the difference coefficient area. If all the significant difference coefficients of the sub-block have been scanned, the significant difference coefficient decoding process is completed, and the process proceeds to the difference coefficient value decoding procedure (S204). The significant coefficient information decoding control unit 1006 determines whether or not the peripheral significant coefficient sum countCoeff is 0 (S403). When the peripheral significant coefficient sum countCoeff is 0, the processing target difference coefficient position in the processing target sub-block is determined (S404). The horizontal difference coefficient position is posX, the vertical difference coefficient position is posY, and the processing target difference coefficient position is pos = posX + posY. If pos <= 2, the context ctxIdx for decoding significant coefficient information is set to 1 (S405), otherwise (pos> 2), the context ctxIdx is set to 0 (S406). The definition of the context when countCoeff = 0 is indicated by reference numeral 601 in FIG. After the determined context is read from the context memory 1004, a decoding instruction is sent to the arithmetic decoding unit 1001 along with the context. The arithmetic decoding unit 1001 performs decoding processing of the encoded sequence using the context, and decodes significant difference coefficient information (S416).

  When the marginal significant coefficient sum countCoeff is not 0, it is determined whether the marginal significant coefficient sum countCoeff is 1 or less (S407). When the peripheral significant coefficient sum countCoeff is 1 or less, the processing target difference coefficient position in the processing target sub-block is determined (S408). If pos <= 3, the context index ctxIdx for decoding significant coefficient information is set to 1 (S409). Otherwise (pos> 3), the context index ctxIdx is set to 0 (S410). The definition of the context when countCoeff = 1 is indicated by reference numeral 602 in FIG. After the determined context is read from the context memory 1004, a decoding instruction is sent to the arithmetic decoding unit 1001 along with the context. The arithmetic decoding unit 1001 performs decoding processing of the encoded sequence using the context, and decodes significant difference coefficient information (S416).

  If the marginal significant coefficient sum countCoeff is not 1 or less, it is determined whether the marginal significant coefficient sum countCoeff is 2 or less (S411). If the peripheral significant coefficient sum countCoeff is 2 or less, the processing target difference coefficient position in the processing target sub-block is determined (S412). If pos <= 2, the context index ctxIdx for decoding significant coefficient information is set to 2 (S413). Otherwise (pos> 2), the context index ctxIdx is set to 1 (S414). The definition of the context when countCoeff = 2 is indicated by reference numeral 603 in FIG. After the determined context is read from the context memory 1004, a decoding instruction is sent to the arithmetic decoding unit 1001 along with the context. The arithmetic decoding unit 1001 performs decoding processing of the encoded sequence using the context, and decodes significant difference coefficient information (S416).

  If the marginal significant coefficient sum countCoeff is not 2 or less, the context index ctxIdx for decoding the significant coefficient information is set to 2 (S415). The definition of the context when countCoeff> 2 is indicated by reference numeral 605 in FIG. After the determined context is read from the context memory 1004, a decoding instruction is sent to the arithmetic decoding unit 1001 along with the context. The arithmetic decoding unit 1001 performs decoding processing of the encoded sequence using the context, and decodes significant difference coefficient information (S416).

  When the peripheral significant coefficient sum countCoeff is large, there is a high possibility that all the significant coefficient information in the processing target sub-block is 1. For this reason, in the above procedure, when the marginal significant coefficient sum countCoeff is 3 or more, ctxIdx is set to 2 regardless of the value of pos. It is also possible to subdivide the judgment conditions for the marginal significant coefficient sum countCoeff. For example, when the marginal significant coefficient sum countCoeff is 3 or more, if the marginal significant coefficient sum countCoeff is 3, the context index definition of the reference numeral 604 in FIG. 11 and if the marginal significant coefficient sum countCoeff is 4 or more, the context index definition in FIG. You can also take When such a configuration is adopted, the correlation utilization efficiency of peripheral information is improved, and the encoding efficiency can be improved.

  In this procedure, the calculation of the context index ctxIdx for significant difference coefficient information refers to the sum of the numbers of significant coefficient information of decoded adjacent subblocks and the position within the subblock of the processing target difference coefficient. The reason for adopting such a configuration will be described below.

  In general, the orthogonal transform coefficients of an image tend to concentrate on low-frequency components, and there is a high possibility that the significant coefficient information becomes 1. Furthermore, since the high frequency component of the orthogonal transform coefficient is not easily affected by visual effects, it is often coarsely quantized. Therefore, the coefficient value of the high frequency component can be 0, and the significant coefficient information of the high frequency component can be 0. High nature. This property is not limited to the entire difference coefficient region, but is the same for each sub-block. The component on the low frequency side of the sub block has significant coefficient information compared to the component on the high frequency side of the same sub block. It can be said that the possibility of becoming 1 becomes high. Setting the value of the context index ctxIdx of the significant difference coefficient information in the low range within the sub-block to be larger than the value of the context index ctxIdx of the significant difference coefficient information in the high range is to estimate the probability of occurrence of the significant coefficient information Leads to improvement. In addition, in the high region where the probability that the significant difference coefficient is 0 is high, the marginal significant coefficient sum is also small, and in the low region where the probability that the significant difference coefficient is 1 is high, the marginal significant coefficient sum tends to be large. Using the marginal significant coefficient sum as an index of how much significant difference coefficient information the block contains leads to improvement in the estimation accuracy of the occurrence probability of significant coefficient information.

  In this procedure, it is possible to calculate the marginal significant difference coefficient sum only once for the sub-block, and calculate the context indexes of all the coefficient positions in the sub-block. Compared with the method of calculating the peripheral significant difference coefficient sum individually at each coefficient position, the calculation amount of the peripheral significant difference coefficient sum can be reduced. Further, in the configuration in which the decoding result of the significant difference coefficient immediately before in the scanning order is used for calculating the context index, it is necessary to sequentially process all the calculation of the context index and the significant difference coefficient decoding in the sub-block. In this embodiment, the peripheral significant difference coefficient sum and the processing target coefficient position are referred to in calculating the context index. However, since the peripheral significant difference coefficient sum does not target the difference coefficient belonging to the processing target sub-block, the context index There is no dependency in the sub-block in the calculation of. Since the context index for all significant difference coefficients can be calculated at the head of the sub-block, the calculation of the context index can be calculated in parallel with the decoding process of the significant difference coefficient information. It is possible to reduce a processing delay related to decoding significant coefficient information having a high occurrence frequency in the encoded sequence.

  It is also possible to perform context calculation with reference to significant sub-block information instead of referring to marginal significant coefficients. In other words, the context can also be calculated based on the sum of adjacent significant sub-block information instead of the sum of the peripheral significant coefficients. The configuration using the sum of the significant sub-block information of the sub-block adjacent to the right side of the processing target sub-block, for example, and the significant sub-block information of the sub-block adjacent to the lower side, for example, is more complex than the above-described configuration. The circuit scale can be reduced. It is also possible to reflect the sub-block position in the context calculation. As described above, the low frequency component has a characteristic that the probability of occurrence of a significant coefficient is higher than the high frequency component. By reflecting the sub-block position in the context calculation, more accurate context estimation can be realized. FIG. 20 shows an example in which the difference coefficient area is classified into two areas, a low-frequency area and a high-frequency area. Reference numerals 1101, 1102, 1103, 1104, 1105, and 1109 in FIG. 20 are low-frequency components, and reference numerals 1106, 1107, 1108, 1110, 1111, 1112, 1113, 1114, 1115, and 1116 are high-frequency areas. For the high frequency region, the context index ctxIdx is calculated by the procedure described above, and for the low frequency region, an offset corresponding to a predetermined sub-block position is added to the context index ctxIdx described above. In addition, for the low frequency region, it is possible to adopt a configuration in which a conditional branch depending on the sub-block position is added during the calculation of the context index ctxIdx described above. Further, after calculating the context index ctxIdx in the above-described procedure for the low-frequency region, generally, the high-frequency region is likely to have a significant difference coefficient of 0, and the number of peripheral significant difference coefficients is Since it is easy to include an error in probability estimation, it is possible to adopt a configuration in which context ctxIdx = 0 is always set.

  It is also possible to calculate the context index using the peripheral coefficient absolute value sum instead of the peripheral significant difference coefficient sum. Since the absolute value of the difference coefficient of the low frequency component is generally large, the coding efficiency can be improved by setting a context that increases the probability of occurrence of significant difference coefficient information when the sum of the absolute values of the peripheral difference coefficients is large. .

  Furthermore, the accuracy of context estimation can be improved by adding the prediction mode used when calculating the difference coefficient to the condition determination in the context index calculation procedure of the significant difference coefficient. This is because, in general, inter-screen prediction that can refer to a plurality of decoded images has a high prediction accuracy and is unlikely to generate a difference, compared to intra-screen prediction in which only a decoded region of a decoding target image is a reference target. .

[Difference coefficient value decoding process (S204)]
The significant coefficient information decoding control unit 1006 determines a difference coefficient to be processed (S501). It is assumed that the scanning order of the difference coefficients within the sub-block follows the rules shown in FIG. 7 in the same manner as the scanning order of the significant difference coefficients. If all the difference coefficients of the sub-block have been scanned, the decoding process of the difference coefficient value is completed, and the process proceeds to the next sub-block determination procedure (S101).

  The significant coefficient information decoding control unit 1006 determines whether or not the difference coefficient value at the processing target difference coefficient position is 0 (S502). When the difference coefficient value at the processing target difference coefficient position is 0, the decoding of the difference coefficient value at the processing target difference coefficient position is completed, and the process proceeds to step S501.

  When the difference coefficient value at the processing target difference coefficient position is 1, the absolute value of the difference coefficient at the processing target difference coefficient position is decoded (S503). When this procedure is executed, it is determined that the difference coefficient value is not 0, and a code word corresponding to a value obtained by subtracting 1 from the absolute value of the difference coefficient is decoded as an encoded sequence. Therefore, a value obtained by adding 1 to the value obtained by entropy decoding the code word is set as the absolute value of the difference coefficient.

  The sign of the difference coefficient at the processing target difference coefficient position is decoded (S504). The difference coefficient value is determined from the absolute value of the difference coefficient and the sign of the difference coefficient.

  In this embodiment, the context index for decoding significant difference coefficient information is calculated from the significant difference coefficient information of the decoded sub-block, but the same procedure is applied to the calculation of the context index of the difference coefficient value. Is also possible. Like the significant difference coefficient information, the difference coefficient value has a correlation with the peripheral coefficient value and the concentration to the low frequency component, so when the peripheral significant difference coefficient sum or the peripheral difference coefficient absolute value sum is large, a large difference coefficient Set a context index that indicates that the probability of occurrence of a numerical value is high, and if the peripheral significant difference coefficient sum or the peripheral difference coefficient absolute value sum is small, set a context index that indicates that the probability of occurrence of a small difference coefficient value is high. Thus, the difference coefficient value can be efficiently encoded.

  The image encoding device and image decoding device of the first embodiment described above have the following operational effects.

  (1) The context index of the processing target difference coefficient can be calculated from the difference coefficient belonging to the decoded sub-block adjacent to the sub-block to which the processing target difference coefficient belongs. When the peripheral significant difference coefficient sum is large, the occurrence probability of the significant difference coefficient information 1 is high, and when the peripheral significant difference coefficient sum is small, by setting a context for estimating the occurrence probability of the significant difference coefficient information 0 high, An appropriate probability model can be set based on the peripheral correlation of significant difference coefficient information. Therefore, significant difference coefficient information can be efficiently encoded.

  (2) A context index is calculated based on the position within the sub-block of the processing target difference coefficient. The difference coefficient in the low band within the sub-block sets a context for estimating the occurrence probability of the significant difference coefficient 1 higher than the difference coefficient in the high band within the sub-block. An appropriate probability model can be set based on the property of the significant difference coefficient information on the frequency domain, and the significant difference coefficient information can be efficiently encoded.

  (3) The calculation of the peripheral significant difference coefficient sum and the position of the processing target difference coefficient in the sub-block do not depend on the decoding result of the significant difference coefficient information in the sub-block. Therefore, since it is possible to adopt a configuration in which the calculation of the context index in the sub-block and the decoding of the significant difference coefficient information can be performed in parallel, the processing delay related to the decoding process of the significant difference coefficient information can be reduced. The difference coefficient has a high ratio of occupying the encoded sequence, and by reducing the processing delay of the significant difference coefficient information having a large number of processes, a decoding apparatus suitable for real-time processing can be realized. Similarly, the encoding apparatus can reduce the processing delay of the significant difference coefficient information encoding.

  (4) A peripheral significant difference coefficient sum related to context index calculation of significant difference coefficient information, that is, a calculation result based on significant sub-block information, significant difference coefficient information, or the sum of absolute values of difference coefficients is the position of the processing target difference coefficient Therefore, it is only necessary to calculate the sub block once. Compared to a configuration in which individual peripheral significant difference coefficient sums are calculated according to processing target difference coefficient positions, it is possible to reduce the amount of calculation related to context index calculation.

  The encoded stream of the image output from the image encoding device according to the embodiment described above has a specific data format so that it can be decoded according to the encoding method used in the embodiment. The image decoding device corresponding to the image encoding device can decode the encoded stream of this specific data format.

  When a wired or wireless network is used to exchange an encoded stream between an image encoding device and an image decoding device, the encoded stream is converted into a data format suitable for the transmission form of the communication path and transmitted. May be. In that case, an image transmission apparatus that converts the encoded stream output from the image encoding apparatus into encoded data in a data format suitable for the transmission form of the communication path and transmits the encoded data to the network, and receives the encoded data from the network There is provided an image receiving apparatus that restores the encoded stream and supplies the encoded stream to the image decoding apparatus.

  An image transmission device includes a memory that buffers an encoded stream output from the image encoding device, a packet processing unit that packetizes the encoded stream, and a transmission unit that transmits packetized encoded data via a network. including. The image reception device receives a packetized encoded data via a network, a memory for buffering the received encoded data, packetizes the encoded data to generate an encoded stream, And a packet processing unit provided to the image decoding device.

  The above processing relating to encoding and decoding can be realized as a transmission, storage, and reception device using hardware, and is stored in a ROM (Read Only Memory), a flash memory, or the like. It can also be realized by firmware or software such as a computer. The firmware program and software program can be recorded on a computer-readable recording medium, provided from a server through a wired or wireless network, or provided as a data broadcast of terrestrial or satellite digital broadcasting Is also possible.

  The present invention has been described based on the embodiments. The embodiments are exemplifications, and it will be understood by those skilled in the art that various modifications can be made to combinations of the respective constituent elements and processing processes, and such modifications are within the scope of the present invention. .

  501 subtraction unit, 502 orthogonal transformation / quantization unit, 503 inverse quantization / inverse transformation unit, 504 addition unit, 505 decoded image memory, 506 prediction unit, 507 difference information coding unit, 508 prediction information coding unit, 509 mode Determination unit, 701 arithmetic coding unit, 702 difference coefficient buffer, 703 coding control unit, 704 context memory, 705 scanning control unit, 706 significant coefficient information coding control unit, 707 difference coefficient value coding control unit, 708 significant sub Block information coding control unit, 801 Difference information decoding unit, 802 Inverse quantization / inverse transformation unit, 803 Prediction information decoding unit, 804 Addition unit, 805 Decoded image memory, 806 Prediction unit, 1001 Arithmetic decoding unit, 1002 Difference coefficient buffer , 1003 Decoding control unit, 1004 Context A memory, 1005 scanning control unit, 1006 significant coefficient information decoding control unit, 1007 difference coefficient value decoding control unit, and 1008 significant sub-block information decoding control unit.

Claims (6)

An image decoding device that decodes an encoded stream in which difference information between an image to be decoded and an image to be predicted is divided into a plurality of sub-blocks and the divided sub-blocks are encoded in a predetermined order. There,
A significant sub-block information decoding unit for decoding significant sub-block information indicating whether or not all the values of the difference coefficients belonging to the sub-block are zero;
A significant difference coefficient information decoding unit that decodes significant difference coefficient information indicating whether or not the value of the difference coefficient is zero;
A difference coefficient value decoding unit for decoding the value of the difference coefficient;
The significance that is not based on the significant difference coefficient information of the difference coefficient belonging to the sub-block to be decoded , belongs to the decoded sub-block adjacent to the sub-block to be decoded, and is adjacent to the sub-block to be decoded An image decoding apparatus comprising: a context deriving unit that derives a context for decoding the significant difference coefficient information of the sub-block to be decoded based on a sum of difference coefficient information. An image decoding method in which difference information between an image to be decoded and an image to be predicted is divided into a plurality of subblocks, and an encoded stream in which the divided subblocks are encoded in a predetermined order is decoded. There,
A significant sub-block information decoding step for decoding significant sub-block information indicating whether or not all the values of the difference coefficients belonging to the sub-block are zero;
Significant difference coefficient information decoding step for decoding significant difference coefficient information indicating whether or not the value of the difference coefficient is zero;
A difference coefficient value decoding step for decoding the value of the difference coefficient;
The significance that is not based on the significant difference coefficient information of the difference coefficient belonging to the sub-block to be decoded , belongs to the decoded sub-block adjacent to the sub-block to be decoded, and is adjacent to the sub-block to be decoded And a context deriving step for deriving a context for decoding the significant difference coefficient information of the sub-block to be decoded based on a sum of the difference coefficient information. An image decoding program for decoding an encoded stream in which difference information between an image to be decoded and an image to be predicted is divided into a plurality of sub-blocks and the divided sub-blocks are encoded in a predetermined order There,
A significant sub-block information decoding step for decoding significant sub-block information indicating whether or not all the values of the difference coefficients belonging to the sub-block are zero;
Significant difference coefficient information decoding step for decoding significant difference coefficient information indicating whether or not the value of the difference coefficient is zero;
A difference coefficient value decoding step for decoding the value of the difference coefficient;
The significance that is not based on the significant difference coefficient information of the difference coefficient belonging to the sub-block to be decoded , belongs to the decoded sub-block adjacent to the sub-block to be decoded, and is adjacent to the sub-block to be decoded An image decoding program that causes a computer to execute a context deriving step for deriving a context for decoding the significant difference coefficient information of the sub-block to be decoded based on a sum of difference coefficient information. A receiving device that receives an encoded stream in which a moving image is encoded and decodes the received encoded stream,
Encoding in which difference information between an image to be decoded and an image to be predicted is divided into a plurality of sub-blocks, and an encoded stream obtained by encoding the divided sub-blocks in a predetermined order is packetized A receiver for receiving data;
A restoration unit that packet-processes the received packetized encoded data to restore the encoded stream;
A significant sub-block information decoding unit that decodes significant sub-block information indicating whether or not all the values of the difference coefficients belonging to the sub-block are zero from the restored encoded stream;
A significant difference coefficient information decoding unit that decodes significant difference coefficient information indicating whether or not the value of the difference coefficient is zero from the restored encoded stream;
A difference coefficient value decoding unit for decoding the value of the difference coefficient from the restored encoded stream;
The significance that is not based on the significant difference coefficient information of the difference coefficient belonging to the sub-block to be decoded , belongs to the decoded sub-block adjacent to the sub-block to be decoded, and is adjacent to the sub-block to be decoded A receiving apparatus comprising: a context deriving unit that derives a context for decoding the significant difference coefficient information of the sub-block to be decoded based on a sum of difference coefficient information. A receiving method for receiving an encoded stream in which a moving image is encoded and decoding the received encoded stream,
Encoding in which difference information between an image to be decoded and an image to be predicted is divided into a plurality of sub-blocks, and an encoded stream obtained by encoding the divided sub-blocks in a predetermined order is packetized A receiving step for receiving data;
A restoration step of packetizing the received packetized encoded data to restore the encoded stream;
A significant sub-block information decoding step for decoding significant sub-block information indicating whether or not all the values of the difference coefficients belonging to the sub-block are all zero from the restored encoded stream;
A significant difference coefficient information decoding step of decoding significant difference coefficient information indicating whether or not the value of the difference coefficient is zero from the restored encoded stream;
A difference coefficient value decoding step of decoding the value of the difference coefficient from the restored encoded stream;
The significance that is not based on the significant difference coefficient information of the difference coefficient belonging to the sub-block to be decoded , belongs to the decoded sub-block adjacent to the sub-block to be decoded, and is adjacent to the sub-block to be decoded And a context deriving step for deriving a context for decoding the significant difference coefficient information of the sub-block to be decoded based on the sum of the difference coefficient information. A reception program that receives an encoded stream in which a moving image is encoded and decodes the received encoded stream,
Encoding in which difference information between an image to be decoded and an image to be predicted is divided into a plurality of sub-blocks, and an encoded stream obtained by encoding the divided sub-blocks in a predetermined order is packetized A receiving step for receiving data;
A restoration step of packetizing the received packetized encoded data to restore the encoded stream;
A significant sub-block information decoding step for decoding significant sub-block information indicating whether or not all the values of the difference coefficients belonging to the sub-block are all zero from the restored encoded stream;
A significant difference coefficient information decoding step of decoding significant difference coefficient information indicating whether or not the value of the difference coefficient is zero from the restored encoded stream;
A difference coefficient value decoding step of decoding the value of the difference coefficient from the restored encoded stream;
The significance that is not based on the significant difference coefficient information of the difference coefficient belonging to the sub-block to be decoded , belongs to the decoded sub-block adjacent to the sub-block to be decoded, and is adjacent to the sub-block to be decoded A receiving program that causes a computer to execute a context deriving step of deriving a context for decoding the significant difference coefficient information of the sub-block to be decoded based on a sum of difference coefficient information.

Images