JP6686095B2 - Decoding device, decoding method, program, and storage medium - Google Patents

Description

  The present invention relates to a coding device, a coding method and program, a decoding device, a decoding method and program, and more particularly to a predictive coding method of a quantization parameter in an image.

  As a compression recording method of a moving image, H.264 is used. H.264 / MPEG-4 AVC (hereinafter, H.264) is known (Non-Patent Document 1). H. H.264 is widely used in 1-segment digital terrestrial broadcasting.

H. In H.264, the quantization parameter can be changed for each macroblock (16 pixels × 16 pixels) using the mb_qp_delta code. According to the 7-23 formula described in Non-Patent Document 1, by adding the difference value mb_qp_delta to the quantization parameter QPY _PREV of the macroblock decoded immediately before, the macroblock (16 pixels × 16 pixels ) The quantization parameter is changed in units.

  In recent years, H. As a successor to H.264, activities to carry out international standardization of HEVC (High Efficiency Video Coding), which is a more efficient encoding method, have been started. In this activity, with the increase of the target screen size, division with a larger block size than the conventional macroblock (16 pixels × 16 pixels) is being considered. According to Non-Patent Document 2, this large-sized basic block is called LCTB (Large Coding Tree Block). The size is being examined as 64 pixels × 64 pixels (Non-Patent Document 2). The LCTB is further divided to form fine sub-blocks, and is divided into CTBs (Condition Tree Blocks) as sub-blocks for conversion and quantization. As a dividing method, an area quadtree structure is used in which the block is divided into two vertically and horizontally.

  FIG. 2A shows the state of the area quadtree structure. A thick frame 10000 represents a basic block, and has a structure of 64 pixels × 64 pixels for simplification of description. Sub-blocks 10001 and 10010 are 16-pixel × 16-pixel sub-blocks. 10002 to 10009 are also sub-blocks, which are sub-blocks of 8 pixels × 8 pixels. In this way, the data is finely divided and encoding processing such as conversion is performed.

ISO / IEC 14496-10; 2004 Information technology--Coding of audio-visual objects--Part 10: Advanced Video Coding JCT-VC Contribution JCTVC-A205. doc <http: // wftp3. itu. int / av-arch / jctvc-site / 2010_04_A_Dresden />

  In HEVC, the control of the quantization parameter is controlled by H.264. Like the H.264 macroblock, it is considered to be performed in basic block units. However, in practice, it is desirable to perform the quantization parameter control in sub-block units from the viewpoint of image quality. In this case, it can be expected that quantization can be performed in smaller units by controlling the quantization parameter in sub-block units.

  However, even if quantization can be performed in smaller units, the processing is performed according to the area quadtree structure, so parallel processing cannot be performed efficiently in sub-block units, and encoding / decoding processing can be speeded up. I couldn't. Specifically, in FIG. 2A, sub-block 10001 of 16 pixels × 16 pixels, sub-blocks 10002 to 10009 of 8 pixels × 8 pixels are processed in numerical order, and finally, sub-blocks of 16 pixels × 16 pixels are processed. Sub-block 10010 is processed. Since the quantization parameter of each sub-block is calculated by the difference using the quantization parameter of the previous sub-block as a prediction value, it is necessary to sequentially perform these. Therefore, parallel processing cannot be performed efficiently in sub-block units.

  Furthermore, when the quantization parameter is optimized in each sub-block, the difference value of the quantization parameter is acquired according to the area quadtree structure, so that the difference value varies. For example, the quantization parameter value is shown in the center of each sub-block in FIG. In this figure, it is assumed that the quantization parameter value changes from upper left to lower right. Such a case is a phenomenon that easily occurs in a normal natural image. The subblock 10001 has a quantization parameter of 12, and the subblock 10002 has a difference of +2. In order, +4, -6, +6, -6, ± 0, +2, +4, +2. In this way, when the area quadtree structure is followed, the values fluctuate so much that the generated code becomes large.

  Therefore, the present invention has been made in order to solve the above-mentioned problems, and enables encoding / decoding of each sub-block in parallel, realizes high-speed processing and achieves a highly efficient quantization parameter code. It is intended to realize encryption / decryption.

In order to solve the above problems, the decoding device of the present invention has the following configuration. That is, a decoding unit that decodes coded data relating to a difference value between a quantization parameter of a sub-block included in one block of a plurality of blocks represented by a quadtree structure and a predetermined parameter,
A second sub-block coded after the first sub-block by adding the quantization parameter of the first sub-block as the predetermined parameter and the difference value acquired by the decoding means. Get the quantization parameter of
As the predetermined parameter, a parameter based on a value obtained by rounding off the average of the quantization parameters of at least two sub-blocks encoded before the third sub-block and the difference value obtained by the decoding means are added. Acquisition means for acquiring the quantization parameter of the third sub-block,
Coefficient decoding means for decoding the quantized coefficients of the first sub-block, the second sub-block, and the third sub-block,
Dequantizing the quantized coefficient of the first sub-block using the first quantization parameter acquired by the acquisition means to reproduce the orthogonal transform coefficient of the first sub-block, Dequantizing the quantized coefficient of the second sub-block using the second quantization parameter acquired by the acquisition means to reproduce the orthogonal transform coefficient of the second sub-block, Inverse of reconstructing the orthogonal transform coefficient of the third sub-block by dequantizing the quantized coefficient of the third sub-block using the third quantization parameter obtained by the obtaining means. Quantizing means,
The prediction error of the first sub-block is reproduced by performing an inverse orthogonal transform on the orthogonal transform coefficient of the first sub-block reproduced by the inverse quantizer, and is reproduced by the inverse quantizer. The prediction error of the second sub-block is reproduced by performing the inverse orthogonal transform on the orthogonal transform coefficient of the second sub-block, and the third sub-block of the third sub-block reproduced by the inverse quantizer is reproduced. Inverse transform means for reproducing the prediction error of the third sub-block by performing inverse orthogonal transform on the orthogonal transform coefficient;
Reproducing the image of the first sub-block based on the prediction error of the first sub-block reproduced by the inverse transforming means and the decoded pixel data of the surroundings or the pixel data of the previous frame, Reproducing the image of the second sub-block based on the prediction error of the second sub-block reproduced by the inverse transforming means, and the decoded pixel data of the surroundings or the pixel data of the previous frame, and Reproducing means for reproducing the image of the third sub-block based on the prediction error of the third sub-block reproduced by the inverse transforming means and the decoded pixel data of the surroundings or the pixel data of the previous frame. ,
It is characterized by including.

  According to the present invention, it is possible to encode / decode a quantization parameter with high efficiency processing while suppressing the code amount.

FIG. 3 is a block diagram showing the configuration of the image encoding device according to the first embodiment. Diagram showing an example of block division Detailed block diagram of the quantization parameter coding unit in the image coding apparatus according to the first embodiment. The flowchart which shows the image coding process in the image coding apparatus which concerns on Embodiment 1. Diagram showing parallel processing during encoding 3 is a block diagram showing the configuration of an image decoding device according to Embodiment 2. FIG. Detailed block diagram of the quantization parameter decoding unit according to the second embodiment The flowchart which shows the image decoding process in the image decoding apparatus in Embodiment 2. Diagram showing parallel processing at the time of decoding FIG. 3 is a block diagram showing the configuration of an image encoding device according to a third embodiment. Detailed block diagram of the quantization parameter coding unit in the image coding apparatus according to the third embodiment. The flowchart which shows the image coding process in the image coding apparatus which concerns on Embodiment 3. A block diagram showing a configuration of an image decoding apparatus according to a fourth embodiment. Detailed block diagram of the quantization parameter decoding unit in the image decoding device according to the fourth embodiment. The flowchart which shows the image decoding process in the image decoding apparatus in Embodiment 2. Detailed block diagram of the quantization parameter coding unit in the image coding apparatus according to the fifth embodiment. The flowchart which shows the image coding process in the image coding apparatus which concerns on Embodiment 5. Detailed block diagram of the quantization parameter decoding unit in the image decoding device according to the sixth embodiment. The flowchart which shows the image decoding process in the image decoding apparatus in Embodiment 6. Block diagram showing a configuration example of computer hardware applicable to the image encoding device and the decoding device of the present invention

  Hereinafter, the present invention will be described in detail based on its preferred embodiments with reference to the accompanying drawings. The configurations shown in the following embodiments are merely examples, and the present invention is not limited to the illustrated configurations.

<Embodiment 1>
Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an image coding apparatus according to this embodiment. In FIG. 1, 1000 is a terminal for inputting image data.
Reference numeral 1001 denotes a block dividing unit that cuts out the input image into a plurality of basic block units and divides the blocks into sub-blocks as necessary. Quantization control will be performed in sub-block units. For ease of explanation, the input image has an 8-bit pixel value, but is not limited to this. The size of the basic block is 64 pixels × 64 pixels, and the minimum size of the sub block is 8 pixels × 8 pixels. In this case, the basic block will include four sub-blocks. Further, regarding division of blocks, a method of dividing into four vertically and horizontally ½ size will be described as an example, but the shape and size of the block are not limited to this. The basic block may include at least two sub-blocks. The sub-block division is not particularly limited. For example, the whole may be divided into sub-blocks, the edge amount or the like may be calculated, and clustering may be performed. That is, a subblock is set finely in a portion where many edges are gathered, and a large subblock is set in a flat portion. Reference numeral 1002 denotes a quantization parameter determination unit that determines the quantization parameter of the basic block and the quantization parameter of each sub block.

  Reference numeral 1003 denotes a block predicting unit that predicts each subblock divided by the block dividing unit 1001 in subblock units and calculates a prediction error of each subblock. Intra prediction is performed for intra frames in the case of still images or moving images, and motion compensation prediction is also performed for moving images. A block transform unit 1004 performs orthogonal transform on the prediction error of each sub-block to calculate orthogonal transform coefficients. The orthogonal transform is not particularly limited, but discrete cosine transform, Hadamard transform, or the like may be used. Reference numeral 1005 denotes a block quantizer that quantizes the orthogonal transform coefficient according to the quantization parameter for each sub-block determined by the quantization parameter determiner 1002. A quantization coefficient can be obtained by this quantization. Reference numeral 1006 denotes a block coding unit that performs variable-length coding on the thus obtained quantized coefficient in sub-block units to generate quantized coefficient code data. The encoding method is not particularly limited, but Huffman code, arithmetic code, or the like can be used. Reference numeral 1007 is a block reproduction image generation unit. The block quantization unit and the block conversion unit 1004 perform the opposite operations to reproduce the prediction error, and generate a decoded image of the basic block from the result of the block prediction unit 1003. The reproduced image data is held and used for prediction in the block prediction unit 1003.

  Reference numeral 1008 denotes a quantization parameter encoding unit that encodes the quantization parameter of the basic block and the quantization parameter of each sub block determined by the quantization parameter determination unit 1002 to generate quantization parameter code data.

  Reference numeral 1009 is an integrated encoding unit. Codes relating to header information and prediction are generated, and the quantization parameter code data generated by the quantization parameter coding unit 1008 and the quantization coefficient code data generated by the block coding unit 1006 are integrated. Reference numeral 1010 denotes a terminal, which is a terminal for outputting the bit stream integrated and generated by the integrated encoding unit 1009 to the outside.

  An image encoding operation in the image encoding device will be described below. In the present embodiment, moving image data is input in frame units, but still image data for one frame may be input.

  The image data for one frame input from the terminal 1000 is input to the block division unit 1001 and divided into basic blocks of 64 pixels × 64 pixels. Further, it is divided into subblocks of a minimum of 8 pixels × 8 pixels as needed. Information regarding sub-block division and the divided image data are input to the quantization parameter determination unit 1002 and the block prediction unit 1003.

  The block prediction unit 1003 performs prediction by referring to the reproduction image held in the block reproduction image generation unit 1007, generates a prediction error, and inputs the prediction error to the block conversion unit 1004 and the block reproduction image generation unit 1007. The block transform unit 1004 performs an orthogonal transform on the input prediction error to obtain an orthogonal transform coefficient, which is input to the block quantizer 1005.

  On the other hand, the quantization parameter determination unit 1002 determines the optimum quantization parameter for each sub-block unit from the balance between the image quality and the code amount, in consideration of the input code amount generated for each sub-block unit. For example, the method described in Japanese Patent Laid-Open No. 4-323961 can be used. The determined quantization parameter of each sub-block is input to the block quantization unit 1005, the block reproduction image generation unit 1007, and the quantization parameter coding unit 1008.

  The block quantization unit 1005 quantizes the orthogonal transform coefficient output from the block transform unit 1004 with the quantization parameter determined by the quantization parameter determination unit 1002 to generate a quantized coefficient. The generated quantized coefficient is input to the block encoding unit 1006 and the block reproduction image generation unit 1007. The block reproduction image generation unit 1007 reproduces the input quantized coefficient with the quantization parameter determined by the quantization parameter determination unit 1002. The orthogonal transform coefficient is inverse orthogonally transformed to reproduce the prediction error. The reproduced prediction error is generated and held as a reproduced image based on the pixel value or the like referred to during prediction. On the other hand, the block coding unit 1006 codes the quantized coefficient, generates quantized coefficient code data, and outputs the quantized coefficient code data to the integrated coding unit 1009.

  The quantization parameter determined by the quantization parameter determination unit 1002 is encoded by the quantization parameter encoding unit 1008 in basic block units.

  FIG. 3 shows a detailed block diagram of the quantization parameter coding unit 1008. In FIG. 3, reference numeral 1 is a terminal for inputting the quantization parameter of each sub-block from the quantization parameter determination unit 1002 of FIG. Reference numeral 2 is a quantization parameter holding unit that temporarily stores the input quantization parameter of each sub-block. Reference numeral 3 is a basic block quantization parameter determination unit that determines the quantization parameter of the basic block from the stored quantization parameter of each sub block. Reference numeral 4 denotes a basic block quantization parameter coding unit that codes a basic block quantization parameter to generate a basic block quantization parameter code. Reference numeral 5 is a terminal for outputting the generated basic block quantization parameter code to the integrated coding unit 1009 in FIG. Reference numeral 6 denotes a sub-block quantization parameter difference unit that obtains a difference from the quantization parameter of each sub-block using the basic block quantization parameter. Reference numeral 7 denotes a sub-block quantization parameter coding unit that codes the difference to generate a sub-block quantization parameter difference value code. Reference numeral 8 is a terminal for outputting the sub-block quantization parameter code to the integrated coding unit 1009 in FIG.

  In the above configuration, the quantization parameter holding unit 2 stores the sub-block quantization parameter input from the terminal 1 in basic block units. Once stored, the basic block quantization parameter determination unit 3 calculates a basic block quantization parameter. In this embodiment, the average value of the sub-block quantization parameters is calculated. In the case of FIG. 2B, the average value is 14.6. If the coding of the quantization parameter is an integer unit, it is rounded off and the basic block quantization parameter is set to 15. The determined basic block quantization parameter is input to the basic block quantization parameter coding unit 4 and the sub block quantization parameter difference unit 6. The basic block quantization parameter coding unit 4 codes the input basic block quantization parameter by Golomb coding, generates a basic block quantization parameter code, and outputs it via a terminal 5.

  On the other hand, the sub-block quantization parameter difference unit 6 calculates the difference between each sub-block quantization parameter and the basic block quantization parameter. In the case of FIG. 2B, the difference values are -3, -1, +3, -3, +3, -3, -3, -1, -1, +5 in the order of the area quadtree structure. These difference values are input to the sub-block quantization parameter coding unit 7. The sub-block quantization parameter coding unit 7 codes these values together with the presence or absence of changes. First, the quantization parameter of the first sub-block 10001 is different from the basic block quantization parameter of 15. Therefore, the 1-bit value "1" indicating that there has been a change and the difference value "-3" are encoded by Golomb encoding and output from the terminal 8 as sub-block quantization parameter difference value encoded data. Then, the difference value of the sub-block quantization parameter of the second sub-block 10002 is encoded. Since this difference value is different from the basic block quantization parameter, a code obtained by Golomb-coding 1-bit values “1” and −1 indicating that there is a change is output from the terminal 8. Thereafter, similarly, a 1-bit value "1" indicating that there is a change and the difference value between the sub-block quantization parameters are encoded to generate sub-block quantization parameter difference value encoded data.

  Returning to FIG. 1, the integrated encoding unit 1009 generates a code such as an image sequence and a frame header, and acquires information such as a prediction mode from the block prediction unit 1003 and encodes each basic block. Then, the basic block quantization parameter code is input from the quantization parameter coding unit 1008. After that, the sub-block quantization parameter difference value encoded data and each quantized coefficient encoded data are integrated for each sub-block and output from the terminal 1010 as a bit stream.

  FIG. 4 is a flowchart showing an image coding process in the image coding apparatus according to the first embodiment. First, in step S001, the integrated coding unit 1009 generates and outputs a code such as a sequence or a frame header.

  In step S002, the block division unit 1001 sequentially cuts out the basic blocks of the input image from the upper left of the image.

  In step S003, the block division unit 1001 further divides the basic block into sub-blocks.

  In step S004, the quantization parameter of each sub-block is determined. In step S005, the quantization parameter of the basic block is determined from the sub-block quantization parameter obtained in step S004. In this embodiment, for the sake of explanation, the average value of the quantization parameters of the sub-blocks in the basic block is calculated as the quantization parameter of the basic block.

  In step S006, the basic block quantization parameter determined in step S005 is Golomb encoded and output as a basic block quantization parameter code.

  In step S007, the sub-block quantization parameter is encoded in sub-block units. When the same quantization parameter as that of the basic block is used in the order of the area quadtree structure, the 1-bit code “0” is output as the code. If they are different, the difference between the sub-block quantization parameter and the basic block quantization parameter is encoded and output following the 1-bit code “1”.

  In step S008, prediction is performed on the image data of the sub block, orthogonal transformation and quantization are performed on the prediction error, the obtained quantized coefficient is encoded, and quantized coefficient code data is output. .

  In step S009, the obtained quantized coefficient is inversely quantized and inversely transformed to calculate a prediction error. A reproduced image of the sub-block is generated from the prediction error and the predicted value from the reproduced image.

  In step S010, the image coding apparatus determines whether or not coding of all sub-blocks in the basic block is completed. If completed, the process proceeds to step S011. The process returns to step S007 for the sub-block of.

  In step S011, the image coding apparatus determines whether or not coding of all basic blocks is completed, and if completed, stops all operations and terminates processing, and otherwise. The process returns to step S002 for the next basic block.

  With the above configuration and operation, the amount of code generated by encoding the difference value of each sub-block quantization parameter using the basic block quantization parameter can be suppressed, particularly by the processing of steps S005 to S009. .

  Although the average value is used as it is in the present embodiment, the present invention is not limited to this, and an actual sub-block quantization parameter closest to the average value may be selected. For example, in the example of FIG. 2B, the average value is 14.6 as described above, but the nearest real sub-block quantization parameter value, 14, may be used instead of rounding. By doing so, the code indicating the change can be set to “0”, and the sub-block quantization parameter difference value to be transmitted can be further reduced.

  Further, in this configuration, prediction, quantization, conversion, and encoding can be efficiently performed in parallel, and there is an effect that speedup can be achieved.

  FIG. 5 shows an example of quantizing / transforming / encoding the sub-block of the basic block shown in FIG. Here, for the sake of explanation, it is assumed that there are three processors used in the encoding process. These processors (A to C) calculate the quantization parameter (QP) of each sub-block, calculate and encode the quantization parameter difference value (ΔQP), orthogonally transform and quantize the prediction error, and code the quantization coefficient. Take the case of conversion as an example. At this time, these codes are integrated by another processor.

  FIG. 5A is a diagram showing an example of conventional parallel processing. First, the processor A is assigned the processing of the sub-block 10001, the processor B is assigned the processing of the sub-block 10002, and the processor C is assigned the processing of the sub-block 10003. Regarding the calculation of QP, the processing time changes depending on the block size, the image complexity, and the like. The calculation of the quantization parameter of the sub-block 10001 having a large block size tends to take more time than the sub-block 10002 and the sub-block 10003.

  Following the quantization parameter calculation, the quantization parameter difference value is calculated. The calculation of the sub-block quantization parameter difference value of the sub-block 10002 cannot be calculated unless the calculation of the sub-block quantization parameter of the sub-block 10001 is completed. Therefore, during this period, the processor B is in a standby state until the processor A calculates the sub-block quantization parameter of the sub-block 10001. Further, if the calculation of the quantization parameter of the sub-block 10002 takes longer than that of the sub-block 10003, the calculation of the difference value of the sub-block 10003 is not required until the calculation of the sub-block quantization parameter of the sub-block 10002 is completed. During this time, the processor C is in a standby state until the processor B calculates the sub-block quantization parameter of the sub-block 10002.

  On the other hand, FIG. 5B is a diagram showing an example of parallel processing in this embodiment. As in the conventional case, the process of the sub-block 10001 is assigned to the processor A, the process of the sub-block 10002 is assigned to the processor B, and the process of the sub-block 10003 is assigned to the processor C. Following the calculation of the sub-block quantization parameter, the sub-block quantization parameter difference value is calculated. The calculation of the sub-block quantization parameter difference value of the sub-block 10002 can be started immediately after the calculation of the sub-block quantization parameter because the basic block quantization parameter has already been calculated. As described above, parallelization can be efficiently performed by the present invention. In particular, when sub-blocks of a plurality of sizes are mixed, the effect of shortening the processing interval becomes large.

  In the present embodiment, the basic block quantization parameter value itself is encoded, but prediction may be performed using the previously processed basic block quantization parameter for encoding.

  Further, in the present embodiment, the basic block is set to 64 pixels × 64 pixels and the sub block is set to 8 pixels × 8 pixels for the sake of description, but the present invention is not limited to this. For example, the basic block can be changed to a block size such as 128 pixels × 128 pixels, and the shapes of the basic block and the sub-block are not limited to squares, and a rectangle of 8 pixels × 4 pixels can form the essence of the invention. does not change.

  Further, in the present embodiment, the basic block quantization parameter is the average value of each sub block quantization parameter, but the present invention is not limited to this. For example, the median value of the sub-blocks may be used, or the sub-block quantization parameter with the highest frequency may be used. As described above, it is of course possible to prepare a plurality of calculation methods and select the most efficient basic block quantization parameter.

  It should be noted that the sub-block quantization parameter difference value encoded data is coded by adding 1 bit of a code indicating the presence or absence of change, but the present invention is not limited to this, and the sub-block quantization parameter difference value is not changed. Of course, it does not matter if it is encoded.

  In addition, although Golomb coding is used for coding the basic block quantization parameter, the sub-block quantization parameter difference value, and the quantization coefficient in the present embodiment, the present invention is not limited to this. For example, Huffman coding or other arithmetic coding may be used, and the value may be output without being coded.

  In the present embodiment, the frame using intra prediction is described as an example, but it is obvious that a frame that can use inter prediction for performing motion compensation for prediction can also be applied.

<Embodiment 2>
In the present embodiment, an image decoding method for decoding coded data coded using the coding method of the first embodiment will be described. FIG. 6 is a block diagram showing the configuration of the image decoding device according to the second embodiment of the present invention.

  1100 is a terminal for inputting an encoded bit stream. Reference numeral 1101 denotes a decoding / separation unit that decodes the header information of the bitstream, separates the necessary code from the bitstream, and outputs the code to the subsequent stage. The decoding / separation unit 1101 performs the reverse operation of the integrated coding unit 1009 in FIG. A quantization parameter decoding unit 1102 decodes the encoded data of the quantization parameter. A block decoding unit 1103 decodes the quantized coefficient code of each sub-block and reproduces the quantized coefficient. Reference numeral 1104 denotes a block dequantization unit that dequantizes the quantized coefficient using the subblock quantization parameter regenerated by the quantization parameter decoding unit 1102 and regenerates the orthogonal transform coefficient. A block inverse transform unit 1105 performs inverse orthogonal transform of the block transform unit 1004 of FIG. 1 to reproduce a prediction error. A block reproduction unit 1106 reproduces sub-block image data from the prediction error and the decoded image data. Reference numeral 1107 denotes a block synthesizing unit for arranging reproduced sub-block image data at respective positions and reproducing the basic block image data.

  The image decoding operation in the image decoding apparatus will be described below. In this embodiment, the moving image bitstream generated in the first embodiment is input in frame units, but a still image bitstream for one frame may be input.

  In FIG. 6, the stream data for one frame input from the terminal 1100 is input to the decoding / separation unit 1101 and the header information necessary for reproducing the image is decoded. The subsequent basic block quantization parameter code is input to the quantization parameter decoding unit 1102. Further, the subsequent sub-block quantization parameter difference value code is also input to the quantization parameter decoding unit 1102.

  FIG. 7 is a block diagram showing details of the quantization parameter decoding unit 1102. Reference numeral 101 is a terminal for inputting a basic block quantization parameter code from the decoding / separation unit 1101 in FIG. Reference numeral 102 denotes a terminal for inputting sub-block quantization parameter difference encoded data from the decoding / separation unit 1101 in FIG. Reference numeral 103 denotes a basic block quantization parameter decoding unit which inputs and decodes a basic block quantization parameter code and reproduces a basic block quantization parameter. A sub-block quantization parameter decoding unit 104 decodes the sub-block quantization parameter difference value encoded data and reproduces each sub-block quantization parameter difference value. Reference numeral 105 denotes a sub-block quantization parameter adding unit that adds the reproduced basic block quantization parameter and each sub-block quantization parameter difference value to reproduce each sub-block quantization parameter. Reference numeral 106 denotes a terminal for outputting the reproduced sub-block quantization parameters to the block dequantization unit 1104 in FIG.

  The basic block quantization parameter code input from the terminal 101 is input to the basic block quantization parameter decoding unit 103, and by decoding the Golomb code, the basic block quantization parameter is reproduced and held.

  The sub-block quantization parameter difference value encoded data input from the terminal 102 is input to the sub-block quantization parameter decoding unit 104, and the Golomb code is decoded to reproduce the sub-block quantization parameter difference value. That is, 1 bit indicating whether or not there is a change in the basic block quantization parameter is decoded, and when there is no change, 0 is output as the sub block quantization parameter difference value. If there is a change, the sub-block quantization parameter difference value is subsequently decoded and output. The sub-block quantization parameter difference value is added to the basic block quantization parameter reproduced by the sub-block quantization parameter addition unit 105 to reproduce the sub-block quantization parameter, and the sub-block quantization parameter is output from the terminal 106.

  Returning to FIG. 6, the quantized coefficient code data of the sub-block separated by the decoding / separation unit 1101 is input to the block decoding unit 1103, and each quantized coefficient is reproduced by decoding the Golomb code. When the quantized coefficient of the sub-block is reproduced, it is input to the block inverse quantization unit 1104. The block dequantization unit 1104 dequantizes the input subblock quantized coefficient and subblock quantized parameter to regenerate the orthogonal transform coefficient. The reproduced orthogonal transform coefficient is inversely transformed by the block inverse transform unit 1105 to reproduce the prediction error. The reproduced prediction error is input to the block reproduction unit 1106, and prediction is performed from the decoded surrounding pixel data or the pixel data of the previous frame to reproduce the sub-block image data. The block synthesizing unit 1107 arranges the reproduced sub-block image data at respective positions, and reproduces the basic block image data. The reproduced basic block image data is output to the outside via the terminal 1108. Also, it is input to the block reproducing unit 1106 and used for calculation of a prediction value.

  FIG. 8 is a flowchart showing an image decoding process in the image decoding device according to the second embodiment. First, in step S101, the decoding / separation unit 1101 decodes the header information.

  In step S102, the basic block quantization parameter decoding unit 103 decodes the basic block quantization parameter code and reproduces the basic block quantization parameter.

  In step S103, the sub-block quantization parameter decoding unit 104 decodes the sub-block quantization parameter difference value encoded data and reproduces the sub-block quantization parameter difference value. Further, the basic block quantization parameter is added to reproduce the sub-block quantization parameter.

  In step S104, the quantized coefficient code data of the sub-block is decoded, the quantized coefficient is reproduced, inverse quantization and inverse orthogonal transform are performed, and the prediction error is reproduced. Further, prediction is performed from the decoded surrounding pixel data or the pixel data of the previous frame, and the decoded image of the sub-block is reproduced.

  In step S105, the decoded image of the sub-block is arranged in the decoded image of the basic block. In step S106, the image decoding device determines whether or not the decoding of all the sub-blocks in the basic block is completed. If the decoding is completed, the process proceeds to step S107, and if not completed, the next sub-block is decoded. The process returns to step S103 for the block.

  In step S107, the decoded image of the basic block is arranged in the decoded image of the frame. In step S108, the image decoding apparatus determines whether or not the decoding of all the basic blocks is completed, and if completed, stops all the operations and ends the processing. The process returns to step S102 for the basic block.

  With the above configuration and operation, it is possible to decode the bit stream generated in the first embodiment and having a reduced code amount, and obtain a reproduced image.

  Furthermore, if the code can be identified by a delimiter for each sub-block, it is possible to efficiently reproduce the sub-block quantization parameter, the sub-block inverse quantization, the inverse transform, and the image data in parallel. It is also possible to perform high-speed decoding.

  FIG. 9 shows an example of reproducing the prediction error by decoding / inverse-quantizing / inverse-transforming the sub-block of the basic block shown in FIG. Similar to FIG. 5 of the first embodiment, it is assumed here that there are three processors used for description. An example of the case where these processors perform decoding of the quantization parameter difference value (ΔQP) of each sub-block, reproduction of the quantization parameter (QP), decoding of the quantization coefficient, and inverse quantization / inverse orthogonal transform of the quantization coefficient Take At this time, these codes are separated by another processor.

  FIG. 9A is a diagram showing an example of conventional parallel processing. First, the processor A is assigned the processing of the sub-block 10001, the processor B is assigned the processing of the sub-block 10002, and the processor C is assigned the processing of the sub-block 10003. Since the processor A is the head, the sub-block quantization parameter itself is decoded, the processors B and C decode each sub-block quantization parameter difference value, and subsequently the sub-block quantization parameter is reproduced. This is achieved by adding the sub-block quantization parameter difference value of the sub-block before it becomes the predicted value of the sub-block quantization parameter and the sub-block quantization parameter difference value.

  The reproduction of the sub-block quantization parameter of the sub-block 10002 is not required until the decoding of the quantization parameter of the sub-block 10001 is completed. Therefore, during this period, the processor B is in a standby state until the processor A reproduces the quantization parameter of the sub-block 10001.

  The same applies to the reproduction of the quantization parameter of the sub block 10002, and the processor C is in the standby state until the reproduction of the quantization parameter of the sub block 10002 in the processor B. After that, the processors that have completed processing are sequentially processed in the order of the area quadtree structure, that is, in the order of subblock 10004, subblock 10005, and subblock 10006. The reproduction of the sub-block quantization parameter of the sub-block 10006 is not required until the reproduction of the quantization parameter of the sub-block 10005 is completed. Therefore, during this period, the processor C is in a standby state until the processor A reproduces the quantization parameter of the sub-block 10005.

  On the other hand, FIG. 9B is a diagram showing an example of parallel processing in this embodiment. First, the processor A decodes and holds the basic block quantization parameter. Hereinafter, as in the conventional case, the process of the sub-block 10001 is assigned to the processor A, the process of the sub-block 10002 is assigned to the processor B, and the process of the sub-block 10003 is assigned to the processor C. Following the decoding of the sub-block quantization parameter difference value, the sub-block quantization parameter is reproduced. The reproduction of the quantization parameter of the sub-block 10002 can be started immediately after the decoding of the sub-block quantization parameter difference value because the basic block quantization parameter has already been reproduced. The present invention enables efficient parallelization. In particular, when sub-blocks of a plurality of sizes are mixed, the effect of shortening the processing interval becomes large.

  Further, when considering an application such as editing for cutting out a part of the image data, for example, consider a case where only the sub-block 10008 in FIG. 2A is cut out. In the case of the conventional example, decoding of sub-blocks 10001 to 1007 is required. According to the present invention, it can be seen that if only subblock 10001 and subblock 10006 are decoded, decoding can be performed even if intra prediction is included. By skipping the decoding process in this way, the processing speed can be improved.

  Further, similarly to the first embodiment, the block size, the processing unit size, the processing unit to be referred to, the pixel arrangement thereof, and the code are not limited to these.

  It should be noted that although Golomb coding is used for decoding the basic block quantization parameter, the sub-block quantization parameter difference value, and the quantization coefficient in the present embodiment, the present invention is not limited to this. For example, Huffman coding or other arithmetic coding may be used, and the value may be output without being coded.

  In the present embodiment, the frame using intra prediction is described as an example, but it is obvious that a frame that can use inter prediction for performing motion compensation for prediction can also be applied.

<Embodiment 3>
FIG. 10 is a block diagram showing the image coding apparatus according to this embodiment. In the present embodiment, the quantization parameter of the first sub-block is the basic block quantization parameter, and the basic block quantization parameter is not individually encoded. The difference from the first embodiment is that there is no code indicating the presence or absence of change. However, as in the first embodiment, encoding may be performed using a code indicating the presence or absence of change. In FIG. 10, parts having the same functions as those in FIG. 1 of the first embodiment are designated by the same reference numerals, and description thereof will be omitted.

  Reference numeral 1208 denotes a quantization parameter encoding unit that encodes sub-block quantization parameters and generates quantization parameter code data. 1209 is an integrated encoding unit. Codes relating to header information and prediction are generated, and the quantization parameter code data generated by the quantization parameter coding unit 1208 and the quantization coefficient code data generated by the block coding unit 1006 are integrated.

  FIG. 11 shows a detailed block diagram of the quantization parameter coding unit 1208. In FIG. 11, the same numbers are given to the portions that perform the same functions as those in FIG. 3 of the first embodiment, and the description will be omitted.

  A selector 200 selects an output destination according to the position of the sub-block of the input sub-block quantization parameter. Reference numeral 203 denotes a basic block quantization parameter holding unit that holds the quantization parameter of the first sub block in the area quadtree structure order of the basic block as a basic block quantization parameter. A sub-block quantization parameter difference unit 206 calculates a difference value between the quantization parameter of each subsequent sub-block and the basic block quantization parameter. Reference numeral 207 denotes a sub-block quantization parameter coding unit that codes the quantization parameter of the head sub-block and the sub-block quantization parameter difference value of each sub-block.

  In the above configuration, as in the first embodiment, sub-block quantization parameters are input from terminal 1 in the order of the area quadtree structure. The selector 200 outputs the quantization parameter of the first subblock in the area quadtree structure order to the basic block quantization parameter holding unit 203. The subsequent sub-block quantization parameters are output to the sub-block quantization parameter difference unit 206.

  The sub-block of the first sub-block is held in the basic block quantization parameter holding unit 203 as a basic block quantization parameter. After that, the leading quantization parameter is also input to the sub-block quantization parameter difference unit 206. Since the sub-block quantization parameter is at the head of the basic block, the difference is not calculated and is output as it is to the sub-block quantization parameter coding unit 207 in the subsequent stage. The sub-block quantization parameter coding unit 207 codes the input sub-block quantization parameter by Golomb code and outputs it from the terminal 8 as sub-block quantization parameter coded data.

  Then, the sub-block quantization parameter is input from the terminal 1 in the order of the region quadtree structure, and is input to the sub-block quantization parameter difference unit 206 via the selector 200. The sub-block quantization parameter difference unit 206 calculates a difference value between the input sub-block quantization parameter and the basic block quantization parameter held in the basic block quantization parameter holding unit 203. The sub-block quantization parameter encoding unit 207 encodes the quantization parameter difference value by Golomb code to generate sub-block quantization parameter difference value encoded data, which is output from the terminal 8 as sub-block quantization parameter encoded data. . After that, the difference is obtained and encoded with respect to the sub-block quantization parameter of the sub-block in the basic block.

  FIG. 12 is a flowchart showing an image coding process in the image coding apparatus according to the third embodiment. In FIG. 12, parts having the same functions as those in FIG. 4 of the first embodiment are designated by the same reference numerals, and description thereof will be omitted.

  From step S001 to step S004, sub-block division and sub-block quantization parameter determination are performed as in the first embodiment.

  In step S205, the top sub-block quantization parameter is held as a basic block quantization parameter.

  In step S206, it is determined whether or not the input sub-block is the head sub-block of the basic block, and if it is the head, the process proceeds to step S208, and if not, the process proceeds to step S207. In step S207, the difference between the basic block quantization parameter held in step S205 and the input sub-block quantization parameter is calculated.

In step S208, the input sub-block quantization parameter or sub-block quantization parameter difference value is encoded by Golomb code and output as sub-block quantization parameter encoded data.
Steps S008 and S009 operate as in the first embodiment. In step S210, it is determined whether the processing has been completed for all sub blocks in the basic block. If not completed, the process proceeds to step S206 to process the next sub-block. If it is finished, the process proceeds to step S011. After that, the entire image is encoded as in the first embodiment. With the above configuration and operation, by setting the basic block quantization parameter as the leading sub-block quantization parameter, it is not necessary to send the basic block quantization parameter, which has the effect of improving the coding efficiency.

  Further, it is possible to effectively perform parallel processing as in the first embodiment. That is, in FIG. 5A, the processors B and C must wait for the calculation of the sub-block quantization parameter at the head of the processor A for the first time. However, thereafter, the calculation of the quantization parameter difference value of the sub block 10005 can be performed without waiting for the processing of the sub block 10004.

  It should be noted that a code that switches between the method of encoding the basic block quantization parameter of the first embodiment and the method of using the basic block quantization parameter of the present embodiment as the top sub-block quantization parameter is provided, and the one with higher encoding efficiency is selected. Of course it doesn't matter.

  In addition, although the same encoding is applied to the leading sub-block quantization parameter (basic block quantization parameter) and the subsequent sub-block quantization parameter difference value, the present invention is not limited to this. It goes without saying that individual coding methods may be applied.

  In addition, although Golomb coding is used for coding the basic block quantization parameter, the sub-block quantization parameter difference value, and the quantization coefficient in the present embodiment, the present invention is not limited to this. For example, Huffman coding or other arithmetic coding may be used.

  In the present embodiment, the frame using intra prediction is described as an example, but it is obvious that a frame that can use inter prediction for performing motion compensation for prediction can also be applied.

<Embodiment 4>
In this embodiment, an image decoding method for decoding coded data that has been coded using the coding method of the third embodiment will be described. FIG. 13 is a block diagram showing an image decoding apparatus that decodes coded data that has been coded using the coding method according to the third embodiment. In FIG. 13, parts having the same functions as those in FIG. 6 of the second embodiment are designated by the same reference numerals, and the description thereof will be omitted.

  In FIG. 13, a decoding / separating unit 1301 decodes header information of the bitstream, separates a necessary code from the bitstream, and outputs it to the subsequent stage. A quantization parameter decoding unit 1302 reproduces sub-block quantization parameters. The decoding / separating unit 301 and the quantization parameter decoding unit 302 in FIG. 10 of the second embodiment are different in the coded data of the quantization parameter.

  The image decoding operation in the image decoding apparatus will be described below. In the present embodiment, the moving image bitstream generated in the third embodiment is input in frame units, but a still image bitstream for one frame may be input.

  Similar to the second embodiment, the one-frame stream data input from the terminal 1100 is input to the decoding / separation unit 1301 and the header information necessary for reproducing the image is decoded. Subsequent sub-block quantization parameter coded data is input to the quantization parameter decoding unit 1302 in the order of the area quadtree structure.

  FIG. 14 is a detailed block diagram of the quantization parameter decoding unit 1302. In FIG. 14, parts that perform the same functions as those in FIG. 7 of the second embodiment are given the same numbers and description thereof is omitted.

  A sub-block quantization parameter decoding unit 304 decodes the sub-block quantization parameter and the sub-block quantization parameter difference value encoded data, and reproduces each sub-block quantization parameter difference value. A selector 300 selects an output destination according to the position of the sub-block of the input sub-block quantization parameter. Reference numeral 310 denotes a basic quantization parameter holding unit that holds the first decoded sub-block quantization parameter as a basic block quantization parameter. A sub-block quantization parameter addition unit 305 reproduces each sub-block quantization parameter by adding the held basic block quantization parameter and each sub-block quantization parameter difference value.

  In the above configuration, at the start of decoding the basic block, the selector 300 sets the output destination to the basic block quantization parameter holding unit 310. The sub-block quantization parameter coded data of the head sub-block of the basic block input from the terminal 102 is input to the sub-block quantization parameter decoding unit 104, and the sub-block quantization parameter is reproduced by decoding the Golomb code. The sub-block quantization parameter of the head sub-block is input to the basic block quantization parameter holding unit 310 via the selector 300 and held during the processing of the basic block. Further, the sub-block quantization parameter of the head sub-block is also input to the sub-block quantization parameter addition unit 305. The sub-block quantization parameter addition unit 305 outputs the reproduced sub-block quantization parameter from the terminal 106 because it is not the difference value for the head sub-block. When the head sub-block quantization parameter is recorded in the basic block quantization parameter holding unit 310, the selector 300 sets the output destination to the sub-block quantization parameter addition unit 305.

  Subsequently, the second and subsequent sub-block quantization parameter difference value encoded data are input to the sub-block quantization parameter decoding unit 304. The input sub-block quantization parameter difference value encoded data reproduces the sub-block quantization parameter difference value by decoding the Golomb code. The sub-block quantization parameter addition unit 305 adds the sub-block quantization parameter difference value input via the selector 300 to the held basic block quantization parameter. In this way, the sub-block quantization parameter is reproduced and output from the terminal 106. After that, the sub-block quantization parameter of the sub-block in the basic block is decoded, the sub-block quantization parameter difference value is calculated, and added to reproduce the sub-block quantization parameter.

  FIG. 15 is a flowchart showing the image decoding process in this embodiment. In FIG. 15, parts having the same functions as those in FIG. 8 of the second embodiment are designated by the same reference numerals, and description thereof will be omitted.

  In step S101, the header information is decoded as in the second embodiment. In step S310, it is determined whether or not the subblock to be decoded is the head subblock of the basic block. If the subblock is the head subblock, the process proceeds to step S311; otherwise, the process proceeds to step S303.

  In step S311, the code relating to the input sub-block quantization parameter is the encoded data of the sub-block quantization parameter. This is decoded by Golomb code. The decoding result is held as a basic block quantization parameter. After that, in order to generate a decoded image of the first sub block, the process proceeds to step S104.

  In step S303, the code relating to the input sub-block quantization parameter is the encoded data of the sub-block quantization parameter difference value. This is decoded by Golomb code to reproduce the sub-block quantization parameter difference value. The reproduced sub-block quantization parameter difference value is added to the basic block quantization parameter held in step S311. The added result becomes the sub-block quantization parameter. In order to generate the decoded image of the second and subsequent sub-blocks, the process proceeds to step S104.

  Hereinafter, as in the second embodiment, the sub-block decoded image is generated and the frame image is reproduced.

  With the above configuration and operation, it is possible to decode the coded data generated in the third embodiment and having a reduced code amount without individually coding the basic quantization parameters.

  Further, it is possible to effectively perform parallel processing as in the second embodiment. That is, in FIG. 9B, instead of decoding the quantization parameter of the processor A, decoding of the head sub-block quantization parameter of the basic block is performed by decoding the basic block quantization parameter and the head sub-block quantization parameter difference value. It will replace the decryption. As a result, B and C need to wait for the decoding of the sub-block quantization parameter at the beginning of the processor A at the beginning, but thereafter, the reproduction of the quantization parameter of all sub-blocks does not wait for the processing of other sub-blocks. You will be able to.

  It should be noted that although Golomb coding is used for decoding the basic block quantization parameter, the sub-block quantization parameter difference value, and the quantization coefficient in the present embodiment, the present invention is not limited to this. For example, Huffman coding or other arithmetic coding may be used.

  If a code for switching between the method of coding the basic block quantization parameter in the third embodiment and the method of setting the basic block quantization parameter of the present embodiment as the head sub-block quantization parameter is provided, the code is interpreted. , Step S102 of FIG. 8 is executed. Alternatively, it may be selected whether to execute steps S310, S311, and S303 in FIG.

  In the present embodiment, the frame using intra prediction is described as an example, but it is obvious that a frame that can use inter prediction for performing motion compensation for prediction can also be applied.

<Fifth Embodiment>
In the present embodiment, an embodiment will be described in which the basic block quantization parameter is determined using the sub-block quantization parameter of the immediately previous basic block.
In the present embodiment, the encoding device has the same configuration as that of FIG. 10 of the third embodiment. However, the configuration of the quantization parameter coding unit 1208 is different.

  FIG. 16 is a block diagram showing a detailed configuration of the quantization parameter coding unit 1208 of this embodiment.

  In FIG. 16, reference numeral 400 denotes a selector, which selects an input destination according to the position of the basic block of the input sub-block quantization parameter. A sub-block quantization parameter holding unit 410 holds the sub-block quantization parameter of the immediately preceding basic block. Reference numeral 403 is a basic block quantization parameter determination unit. The basic block quantization parameter of the basic block to be encoded is determined from the sub block quantization parameter held by the sub block quantization parameter holding unit 410. A sub-block quantization parameter difference unit 406 calculates a difference value between the basic block quantization parameter and each sub-block quantization parameter. Reference numeral 407 is a sub-block quantization parameter coding unit that codes the quantization parameter of the first sub-block and the sub-block quantization parameter difference value of each sub-block.

  In the above configuration, as in the third embodiment, image data is input from the terminal 1000, divided into sub-blocks by the block division unit 1001, and the quantization parameter of each sub-block is determined by the quantization parameter determination unit 1002. The determined sub-block quantization parameter is input to the quantization parameter coding unit 1208.

  16, if the input sub-block quantization parameter is the head sub-block of the head basic block of the image, the selector 400 selects the input from the terminal 1. The input sub block quantization parameter is input to the basic block quantization parameter determination unit 403 via the sub block quantization parameter holding unit 410, the sub block quantization parameter difference unit 406, and the selector 400. The sub-block quantization parameter holding unit 410 holds the sub-block quantization parameter for processing the next basic block. The basic block quantization parameter determination unit 403 holds the input sub block quantization parameter as a basic block quantization parameter, as in the basic block quantization parameter holding unit 203 of the third embodiment. The sub-block quantization parameter difference unit 406 outputs the sub-block quantization parameter as it is, similar to the sub-block quantization parameter difference unit 206 of the third embodiment. The sub-block quantization parameter encoding unit 407 encodes this head sub-block quantization parameter with Golomb code and outputs it from the terminal 8.

  After that, the other sub-block quantization parameters of the basic block at the beginning of the image are input from the terminal 1, and are input to the sub-block quantization parameter holding unit 410 and the sub-block quantization parameter difference unit 406. The sub-block quantization parameter difference unit 406 calculates the difference value between the basic block quantization parameter output from the basic block quantization parameter determination unit 403 and the input. The difference value is input to the sub-block quantization parameter encoding unit 407, encoded as in the third embodiment, and output from the terminal 8.

  On the other hand, the processing will be described for a basic block that is not the beginning of an image that is subsequently input. Prior to the encoding process of the basic block, the selector 400 selects the sub-block quantization parameter holding unit 410 as the input destination. The basic block quantization parameter determination unit 403 calculates the average value of the held sub block quantization parameters and sets this as the basic block quantization parameter. After that, the sub-block quantization parameter of the basic block is input from the terminal 1 to the sub-block quantization parameter difference unit 406. The sub-block quantization parameter difference unit 406 calculates the difference value between the basic block quantization parameter output from the basic block quantization parameter determination unit 403 and the input. The difference value is input to the sub-block quantization parameter encoding unit 407, encoded as in the third embodiment, and output from the terminal 8.

  FIG. 17 is a flowchart showing the image coding process in the image coding apparatus according to the fifth embodiment. In FIG. 17, parts having the same functions as those in FIG. 4 of the first embodiment are designated by the same reference numerals, and description thereof will be omitted.

  In steps S001 to S003, header information is encoded, basic blocks are divided, and sub-blocks are divided, as in the first embodiment. In step S401, it is determined whether the basic block is the head of the image. If it is the head, the process proceeds to step S402, and if not, the process proceeds to step S409. In step S402, it is determined whether the sub block is the head sub block of the head basic block. If it is the first sub-block, the process proceeds to step S403, and if not, the process proceeds to step S406.

  In step S403, the sub-block quantization parameter of the first sub-block of the first basic block is determined and held so that it can be referred to when the next basic block is processed. In step S404, the sub-block quantization parameter determined in step S403 is held as a basic block quantization parameter. In step S405, the sub-block quantization parameter determined in step S403 is encoded, and the process proceeds to step S008. In step S406, the sub-block quantization parameter of the sub-block is determined and held so that it can be referred to when the next basic block is processed.

  In step S407, the basic block quantization parameter held in step S404 is subtracted from the sub-block quantization parameter determined in step S406. Thereby, the sub-block quantization parameter difference value of the sub-block is calculated. In step S408, the quantization parameter difference value calculated in step S407 is encoded to generate quantization parameter difference value encoded data, and the process proceeds to step S008. In step S409, it is determined whether the sub-block is the head sub-block of the basic block that is not the head. If it is the first sub-block, the process proceeds to step S410, and if not, the process proceeds to step S406. In step S410, the basic block quantization parameter of the basic block is calculated with reference to the sub-block quantization parameter of the previous basic block held in steps S403 to S406. In this embodiment, the average value of the sub-block quantization parameters is calculated and used as the basic block quantization parameter. In step S411, the sub-block quantization parameter of the sub-block is determined and held so that it can be referred to when the next basic block is processed.

  In step S412, the basic block quantization parameter calculated in step S410 is subtracted from the sub-block quantization parameter determined in step S411. Thereby, the sub-block quantization parameter difference value of the sub-block is calculated.

  In step S413, the quantization parameter difference value calculated in step S412 is encoded to generate quantization parameter difference value encoded data, and the process proceeds to step S008. In step S414, the image coding apparatus determines whether or not coding of all the sub-blocks in the basic block is completed. If completed, the process proceeds to step S011. The process returns to step S401 for the sub-block of. Steps S008, S009, and S011 operate in the same manner as in the first embodiment, and the entire image is encoded.

  With the above configuration and operation, since the basic block quantization parameter is determined using the sub-block quantization parameter of the previous basic block, the basic block quantization parameter can be determined at the same time when the processing of the basic block is started. Therefore, there is an effect of minimizing the processing delay. Further, since the basic block quantization parameter is generated from the sub-block quantization parameter of the previous basic block, it is not necessary to send the basic block quantization parameter, which also has an effect of improving the coding efficiency.

  Further, it is possible to effectively perform parallel processing as in the first embodiment. That is, in FIG. 5B, the basic block quantization parameter is calculated before encoding by the sub-block quantization parameter of the previous basic block. As a result, the calculation of the quantization parameter difference values for all subblocks can be performed without waiting for the processing of each subblock.

  Note that, in the present embodiment, the sub-block quantization parameter of the head sub-block is directly encoded only when the head basic block of the image is used, but the present invention is not limited to this. In other words, a slice-like structure composed of a plurality of basic blocks may be used and the same processing may be performed on the first basic block.

  Further, in the present embodiment, the determination is made by referring to the sub-block quantization parameter of the previous basic block, but the present invention is not limited to this, and the last sub-block quantization parameter of the previous basic block is set to the basic block quantum of the basic block. It may be used as a conversion parameter. Further, it is of course possible to refer to sub-block quantization parameters or basic block quantization parameters of surrounding basic blocks.

  Further, in the present embodiment, the basic block quantization parameter is the average value of each sub-block quantization parameter of the previous basic block, but the present invention is not limited to this. For example, the median value of the sub-blocks may be used, or the sub-block quantization parameter with the highest frequency may be used. As described above, it is of course possible to prepare a plurality of calculation methods, select the most efficient basic block quantization parameter, and perform coding using a code representing the method.

  In the present embodiment, the frame using intra prediction is described as an example, but it is obvious that a frame that can use inter prediction for performing motion compensation for prediction can also be applied.

<Sixth Embodiment>
In the present embodiment, an image decoding method for decoding coded data coded using the coding method of the fifth embodiment will be described. In the present embodiment, the encoding device has the same configuration as that of FIG. 13 of the fourth embodiment. However, the configuration of the quantization parameter decoding unit 1302 is different.

  FIG. 18 is a block diagram showing the configuration of the quantization parameter decoding unit 1302 according to the sixth embodiment of the present invention. In FIG. 18, parts having the same functions as those in FIG. 14 of the fourth embodiment are designated by the same reference numerals, and description thereof will be omitted.

  A selector 500 selects an output destination according to the position of the sub-block of the input sub-block quantization parameter and the basic block of the sub-block. A sub-block quantization parameter decoding unit 501 decodes a code when the sub-block quantization parameter itself is encoded and reproduces the sub-block quantization parameter. A sub-block quantization parameter difference value decoding unit 502 decodes the code when the sub-block quantization parameter difference value is encoded and reproduces the sub-block quantization parameter difference value. A selector 503 selects an input destination according to the position of the input sub-block quantization parameter sub-block and the basic block of the sub-block. A basic block quantization parameter determination unit 504 determines a basic block quantization parameter. A sub-block quantization parameter addition unit 505 reproduces each sub-block quantization parameter by adding the determined basic block quantization parameter and each sub-block quantization parameter difference value. A selector 506 selects an input destination according to the position of the input sub-block quantization parameter sub-block and the basic block of the sub-block. A sub-block quantization parameter holding unit 507 holds the reproduced sub-block quantization parameter.

  The image decoding operation in the image decoding apparatus will be described below. In the present embodiment, the moving image bitstream is input in frame units, but a still image bitstream for one frame may be input.

  Prior to the decoding processing of the bit stream for one frame, the selector 500 sets the output destination to the sub-block quantization parameter decoding unit 501, and the selector 503 sets the input destination to the sub-block quantization parameter decoding unit 501. The selector 505 sets the input destination to the sub-block quantization parameter decoding unit 501.

  First, the encoded data of the sub-block quantization parameter of the first basic block is input to the sub-block quantization parameter decoding unit 501 via the selector 500. The sub-block quantization parameter decoding unit 501 decodes using the Golomb code to reproduce the sub-block quantization parameter. This sub-block quantization parameter is input to the basic block quantization parameter determination unit 504 via the selector 503. Since the sub-block of the sub-block quantization parameter is the head sub-block of the head basic block, the basic block quantization parameter determination unit 504 holds the input as it is and sets it as the basic block quantization parameter. In addition, the sub-block quantization parameter reproduced by the sub-block quantization parameter decoding unit 501 is output from the terminal 106 via the selector 506. The sub-block quantization parameter holding unit 507 holds the sub-block quantization parameter.

  After that, the selector 500 sets the output destination to the sub-block quantization parameter difference value decoding unit 502, and the selector 503 sets the input destination to the sub-block quantization parameter holding unit 507. The selector 506 sets the input destination to the sub-block quantization parameter addition unit 305.

  When the sub block quantization parameter difference value encoded data of the subsequent sub block is input, it is input to the sub block quantization parameter difference value decoding unit 502 via the selector 500. The sub-block quantization parameter difference value decoding unit 502 decodes the sub-block quantization parameter difference value encoded data to reproduce the sub-block quantization parameter difference value. The sub-block quantization parameter addition unit 305 adds the sub-block quantization parameter difference value to the basic block quantization parameter to reproduce the sub-block quantization parameter. The reproduced sub-block quantization parameter is output from the terminal 106. The sub-block quantization parameter holding unit 507 holds the sub-block quantization parameter.

  Sub-block quantization parameter difference value encoded data of the following basic block is input. At this time, the basic block quantization parameter determination unit 504 reads the sub block quantization parameter of the previous basic block from the sub block quantization parameter holding unit 507. The average value of the read sub-block quantization parameters is calculated and used as the basic block quantization parameter of the basic block.

The input sub-block quantization parameter difference value encoded data is decoded by the sub-block quantization parameter difference value decoding unit 502 to reproduce the sub-block quantization parameter difference value. The sub-block quantization parameter addition unit 305 reproduces the sub-block quantization parameter, outputs it from the terminal 106, and holds it in the sub-block quantization parameter holding unit 507.
Similarly, for the sub-block quantization parameter difference value encoded data that is subsequently input, the sub-block quantization parameter difference value is reproduced in the same manner, and then the sub-block quantization parameter is reproduced. The reproduced sub-block quantization parameter is output from the terminal 106 and held in the sub-block quantization parameter holding unit 507.

  FIG. 19 is a flowchart showing the image decoding process in the image decoding device according to the sixth embodiment.

  In step S101, the header information is decoded as in the second embodiment. In step S501, it is determined whether the basic block of the sub-block to be decoded is the head of the image. If it is the head basic block, the process proceeds to step S502, and if not, the process proceeds to step S504.

In step S502, it is determined whether or not the sub block to be decoded is the head sub block of the basic block. If the sub block is the head block, the process proceeds to step S503, and if not, the process proceeds to step S506. In step S503, the code relating to the input sub-block quantization parameter is the encoded data of the sub-block quantization parameter. This is decoded by Golomb code. The decoding result is held as a basic block quantization parameter. At the same time, it is held separately for reference when determining the basic block quantization parameter of the next basic block. After that, in order to generate a decoded image of the first sub block, the process proceeds to step S104.
In step S504, it is determined whether or not the subblock to be decoded is the head subblock of the basic block. If the subblock is the head block, the process proceeds to step S505, and if not, the process proceeds to step S506.

  In step S505, an average value is obtained from the sub-block quantization parameters of the previous basic block that have been held, and the average value is calculated as the basic block quantization parameter. After that, the process proceeds to step S506.

  In step S506, the code regarding the input sub-block quantization parameter is the sub-block quantization parameter difference value encoded data. This is decoded by Golomb code to reproduce the sub-block quantization parameter difference value. The reproduced sub-block quantization parameter difference value is added to the basic block quantization parameter held or calculated in step S503 or step S505. The added result becomes the sub-block quantization parameter. After that, the process proceeds to step S104 to generate a decoded image of the sub block. Hereinafter, similarly to the fourth embodiment, the sub-block decoded image is generated and the frame image is reproduced.

  With the above configuration and operation, it is possible to decode a bitstream that does not encode the basic block quantization parameter value generated in the fifth embodiment.

  Further, it is possible to effectively perform parallel processing as in the second embodiment. That is, instead of decoding the basic block quantization parameter of the processor A in FIG. 9B, it is replaced with calculation of the basic block quantization parameter by the sub-block quantization parameter of the previous basic block. As a result, the reproduction of the quantization parameters of all subblocks can be performed without waiting for the processing of other subblocks.

  In the present embodiment, the basic block quantization parameter is the average value of the sub-block quantization parameters of the previous basic block. However, if the same method as the basic block quantization parameter calculation method of the fifth embodiment is used, Not limited. For example, the median value of the sub-blocks may be used, or the sub-block quantization parameter with the highest frequency may be used. Such information can be derived from the sub-block quantization parameter held in the sub-block quantization parameter holding unit 507.

  In addition, even if a plurality of calculation methods are prepared on the encoding side and the most efficient basic block quantization parameter is selected and coded using a code representing the method, decoding is performed and sub-block quantum is similarly performed. The calculation parameter can be calculated.

In addition, in the present embodiment, a frame using intra prediction has been described as an example, but it is clear that a frame that can use inter prediction that performs motion compensation for prediction can also be supported.
In the above embodiment, it is assumed that each processing unit shown in FIGS. 1, 3, 6, 7, 10, 10, 11, 13, 14, 16 and 18 is configured by hardware. explained. However, the processing performed by each processing unit shown in these figures may be configured by a computer program.

  FIG. 20 is a block diagram showing an example of the hardware configuration of a computer applicable to the image display device according to each of the above embodiments.

  The CPU 1401 controls the entire computer using the computer programs and data stored in the RAM 1402 and the ROM 1403, and also executes the processes described above as the processes performed by the image processing apparatus according to each of the above embodiments. That is, the CPU 1401 functions as each processing unit shown in FIGS. 1, 3, 6, 7, 7, 10, 11, 13, 14, 16 and 18.

  The RAM 1402 has an area for temporarily storing a computer program or data loaded from the external storage device 1406, data acquired from the outside via an I / F (interface) 1409, and the like. Further, the RAM 1402 has a work area used when the CPU 1401 executes various processes. That is, the RAM 1402 can be allocated as, for example, a frame memory or can appropriately provide other various areas.

  The ROM 1403 stores setting data of the computer, a boot program, and the like. The operation unit 1404 is composed of a keyboard, a mouse and the like, and can be operated by the user of the computer to input various instructions to the CPU 1401. The display unit 1405 displays the processing result by the CPU 1401. Further, the display unit 1405 is configured by a hold type display device such as a liquid crystal display or an impulse type display device such as a field emission type display device.

  The external storage device 1406 is a large capacity information storage device represented by a hard disk drive device. The external storage device 1406 has an OS (Operating System) and functions of each unit shown in FIGS. 1, 3, 6, 7, 7, 10, 11, 14, 16, and 18. A computer program to be realized by the CPU 1401 is stored. Furthermore, each image data to be processed may be stored in the external storage device 1406.

  The computer programs and data stored in the external storage device 1406 are appropriately loaded into the RAM 1402 under the control of the CPU 1401 and are processed by the CPU 1401. The I / F 1407 can be connected to a network such as a LAN or the Internet, and other devices such as a projection device and a display device. The computer acquires and sends various information via the I / F 1407. You can 1408 is a bus connecting the above-mentioned units.

  The operation having the above-described configuration is controlled mainly by the CPU 1401 which is the operation described in the above flowchart.

  Further, when the CPU is a multi-core, parallel processing can be efficiently performed by assigning threads for each processing to each core.

<Other embodiments>
The object of the present invention is also achieved by supplying a storage medium storing a code of a computer program that realizes the above-described functions to a system, and the system reading and executing the code of the computer program. In this case, the computer program code itself read from the storage medium realizes the functions of the above-described embodiments, and the storage medium storing the computer program code constitutes the present invention. Further, it includes a case where an operating system (OS) running on a computer performs some or all of actual processing based on the instructions of the code of the program, and the processing realizes the above-described functions. .

  Further, it may be realized in the following forms. That is, the computer program code read from the storage medium is written in the memory provided in the function expansion card inserted in the computer or the function expansion unit connected to the computer. Then, based on the instructions of the code of the computer program, a case where the CPU or the like included in the function expansion card or the function expansion unit performs some or all of the actual processing to realize the above-described functions is also included.

  When the present invention is applied to the storage medium, the storage medium stores the code of the computer program corresponding to the above-described flowchart.

Claims (3)

A decoding device for decoding a quantization parameter for performing a quantization process on an input image,
Decoding means for decoding coded data relating to a difference value between a quantization parameter of a sub-block included in one block of a plurality of blocks represented by the quadtree structure and a predetermined parameter,
The second sub-block coded after the first sub-block by adding the quantization parameter of the first sub-block as the predetermined parameter and the difference value acquired by the decoding means. Get the quantization parameter of
As the predetermined parameter, a parameter based on a value obtained by rounding off the average of the quantization parameters of at least two sub-blocks encoded before the third sub-block and the difference value obtained by the decoding means are added. Acquisition means for acquiring the quantization parameter of the third sub-block,
Coefficient decoding means for decoding the quantized coefficients of the first sub-block, the second sub-block, and the third sub-block,
Dequantizing the quantized coefficient of the first sub-block using the first quantization parameter acquired by the acquisition means to reproduce the orthogonal transform coefficient of the first sub-block, Dequantizing the quantized coefficient of the second sub-block using the second quantization parameter acquired by the acquisition means to reproduce the orthogonal transform coefficient of the second sub-block, Inverse of reconstructing the orthogonal transform coefficient of the third sub-block by dequantizing the quantized coefficient of the third sub-block using the third quantization parameter obtained by the obtaining means. Quantizing means,
The prediction error of the first sub-block is reproduced by performing an inverse orthogonal transform on the orthogonal transform coefficient of the first sub-block reproduced by the inverse quantizer, and is reproduced by the inverse quantizer. The prediction error of the second sub-block is reproduced by performing the inverse orthogonal transform on the orthogonal transform coefficient of the second sub-block, and the third sub-block of the third sub-block reproduced by the inverse quantizer is reproduced. Inverse transform means for reproducing the prediction error of the third sub-block by performing inverse orthogonal transform on the orthogonal transform coefficient;
Reproducing the image of the first sub-block based on the prediction error of the first sub-block reproduced by the inverse transforming means and the decoded pixel data of the surroundings or the pixel data of the previous frame, Reproducing the image of the second sub-block based on the prediction error of the second sub-block reproduced by the inverse transforming means, and the decoded pixel data of the surroundings or the pixel data of the previous frame, and Reproducing means for reproducing the image of the third sub-block based on the prediction error of the third sub-block reproduced by the inverse transforming means and the decoded pixel data of the surroundings or the pixel data of the previous frame. ,
A decoding device comprising: A decoding method for decoding a quantization parameter for performing a quantization process on an input image,
A decoding step of decoding coded data relating to a difference value between a quantization parameter of a sub-block included in one block of a plurality of blocks represented by a quadtree structure and a predetermined parameter;
The second sub-block coded after the first sub-block by adding the quantization parameter of the first sub-block as the predetermined parameter and the difference value acquired in the decoding step. Get the quantization parameter of
As the predetermined parameter, a parameter based on a value obtained by rounding off the average of the quantization parameters of at least two sub-blocks encoded before the third sub-block and the difference value obtained in the decoding step are added. An acquisition step of acquiring the quantization parameter of the third sub-block,
A coefficient decoding step of decoding quantized coefficients of the first sub-block, the second sub-block, and the third sub-block,
Dequantizing the quantized coefficient of the first sub-block using the first quantization parameter acquired in the acquisition step to reproduce the orthogonal transform coefficient of the first sub-block, Dequantizing the quantized coefficient of the second sub-block using the second quantization parameter acquired in the acquisition step to reproduce the orthogonal transform coefficient of the second sub-block, Inverse of reconstructing the orthogonal transform coefficient of the third sub-block by dequantizing the quantized coefficient of the third sub-block using the third quantization parameter acquired in the acquisition step. The quantization process,
The prediction error of the first sub-block is reproduced by performing an inverse orthogonal transform on the orthogonal transform coefficient of the first sub-block reproduced in the inverse quantization step, and is reproduced in the inverse quantization step. The prediction error of the second sub-block is reproduced by performing the inverse orthogonal transform on the orthogonal transform coefficient of the second sub-block, and the prediction error of the third sub-block reproduced in the inverse quantization step is reproduced. An inverse transform step of reproducing the prediction error of the third sub-block by performing an inverse orthogonal transform on the orthogonal transform coefficient,
Reproducing the image of the first sub-block based on the prediction error of the first sub-block reproduced in the inverse transforming step and the decoded pixel data of the surroundings or the pixel data of the previous frame, Reconstructing the image of the second sub-block based on the prediction error of the second sub-block reconstructed in the inverse transforming step and surrounding decoded pixel data or pixel data of the previous frame, and A reproducing step of reproducing the image of the third sub-block based on the prediction error of the third sub-block reproduced in the inverse transforming step and the decoded pixel data of the surroundings or the pixel data of the previous frame; ,
A decoding method comprising:   A program that causes the computer to function as the decoding device according to claim 1, when the program is read and executed by the computer.

Images