JP2013034037A - Image encoder, image encoding method and program, image decoder, and image decoding method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve the problem in which, when controlling a quantization parameter in a block unit, there has conventionally been one calculation method of a predictive quantization parameter, and depending on an encoding method of an image or image characteristics, an absolute value of the differential value becomes large and a code amount of a quantization parameter code unnecessary increases.SOLUTION: The image encoding method, which divides an input image into a plurality of blocks of different sizes and encodes the image in the divided block unit, comprises processes of: acquiring attribute information of a block to be processed; calculating a parameter which controls image quality of the block to be processed; determining a predictive control parameter on the basis of the attribute information; calculating a differential value between the calculated control parameter and the predictive control parameter; and generating control parameter differential value encoded data by encoding the calculated differential value.

Description

  The present invention relates to an image encoding device, an image encoding method and program, an image decoding device, an image decoding method and a program, and more particularly to an image quality control parameter encoding method and an image quality control parameter decoding method.

  MPEG-2 Video (hereinafter abbreviated as MPEG-2, Non-Patent Document 1), H. A method such as H.264 (Non-patent Document 2) is known. Furthermore, JCT-VC (Joint Collaborative Team on Video Coding) was recently established as a joint organization of ITU-T and ISO / IEC. In this organization, standardization activities of HEVC (High Efficiency Video Coding), which is a new video standard, are being promoted. For example, as in Non-Patent Document 3, HEVC includes H.264. Improvement techniques based on H.264 have been proposed.

  MPEG-2, H.264. In an encoding method using orthogonal transformation and quantization represented by H.264, HEVC, quantization coefficient data is generated by orthogonally transforming and quantizing a predetermined block image on the encoding side. At this time, the image quality is controlled by performing quantization using an image quality control parameter called a quantization parameter. Specifically, when quantization is performed using a small quantization parameter value, the image quality is improved, but the code amount is large.When quantization is performed using a large quantization parameter value, the image quality is decreased, but the code amount is Get smaller. In the encoding process, encoding is performed while selecting an optimal quantization parameter value according to the target code amount. This is called rate control, and various methods such as TM5 (Non-Patent Document 4) have been proposed. There is also a method of determining whether the image area is visually important and controlling the image quality control parameter according to the importance (Patent Document 1).

  The quantized quantized coefficient data is subjected to variable length coding, and variable length coded coefficient data is generated. The quantization parameter is also encoded, and a quantization parameter code is generated. As a quantization parameter generation method, for example, a quantization parameter used for quantization of a block quantized before a quantization target block is set as a prediction quantization parameter, and the prediction quantization parameter and the target block are quantized. A difference value from the used quantization parameter is calculated. This difference value is called (QP_DELTA: Quantization Parameter Delta). This difference value is embedded in the bitstream as a quantization parameter code. The variable length coding coefficient data and the quantization parameter code generated in this way are transmitted as a bit stream to the decoder via an optical disk medium or a network. The decoding side decodes variable length coding coefficient data and quantization parameter code to generate quantization coefficient data and quantization parameter, and performs inverse quantization / inverse orthogonal transformation of the quantization coefficient data using the quantization parameter. To generate a decoded image.

  MPEG-2 and H.264 In H.264, a 16 × 16 lattice block obtained by dividing an image called a macroblock in a lattice form is used as one processing unit. If the size of the block to be orthogonally transformed is expressed in units of pixels, MPEG-2 is 8 × 8, In H.264, it is 8 × 8 or 4 × 4. That is, a plurality of orthogonal transform blocks exist in one macro block. MPEG-2 and H.264 In H.264, the quantization parameter can be controlled (rate control) in units of macroblocks, so that orthogonal transform blocks included in the same macroblock are quantized with the same quantization parameter.

  On the other hand, in HEVC, a lattice block obtained by dividing an image into a lattice is called an LCU (Large Coding Unit), and its size is 64 × 64. The LCU is divided into smaller sized blocks called CU (Cording Unit) using a region quadtree structure. Further, the CU includes an orthogonal transform block called a TU (Transform Unit), and this TU is also divided into smaller sizes using a region quadtree structure. Each unit of the quantization parameter has a division flag. A block whose division flag is True has a structure including four divided blocks each having a vertical and horizontal size of 1/2, and a block whose division flag is False has actual data of a block instead of including a divided block ( Non-patent document 3). Although various methods can be used to determine whether to divide a block, as one determination method, a block cost is calculated using a Lagrange multiplier, and a block dividing method with a lower cost is selected. There is a thing (patent document 2).

JP 2001-45494 A JP 2005-191706 A

ISO / IEC 13818-2: 2000 Information technology-Generic coding of moving pictures and associated audio information: Video 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 /> MPEG-2 Test Model 5 (TM5), Doc. ISO / IEC JTC1 / SC29 / WG11 / N0400, Test Model Editing Committee, April 1993

In the quantization parameter encoding method that encodes the difference value between the quantization parameter of the encoding target block and the predicted quantization parameter as a quantization parameter code, the larger the absolute value of the difference value, the larger the code of the quantization parameter code. The amount gets bigger. When embedding quantization parameters in units of blocks, the prediction quantization parameter is conventionally calculated in a single method. Depending on the image encoding method and the image characteristics, the absolute value of the difference value becomes large, and the quantum is unnecessarily increased. The code amount of the conversion parameter has increased.
Therefore, the present invention has been made to solve the above-described problem, and an object of the present invention is to prevent an unnecessary increase in the amount of quantization parameter code and to reduce the amount of quantization parameter code.

  In order to solve the above-described problems, the image coding apparatus of the present invention has the following configuration. That is, in an image encoding method that divides an input image into a plurality of blocks having different sizes and encodes an image in units of the divided blocks, an acquisition step of acquiring attribute information of a processing target block, and the processing target A setting step for setting a control parameter for controlling the image quality of the block; a determination step for determining a prediction control parameter based on the attribute information; and a difference value between the control parameter and the prediction control parameter calculated in the setting step. And a coding step for coding the difference value calculated in the calculation step.

  According to the present invention, the code amount of the quantization parameter code can be reduced.

Flowchart illustrating a divided block coding method described as the first embodiment Flowchart illustrating a divided block coding method described as the third embodiment Flowchart illustrating a divided block coding method described as the fifth embodiment Flowchart illustrating a divided block coding method described as the seventh embodiment Flowchart illustrating a divided block coding method described as the ninth embodiment Flowchart illustrating a divided block decoding method described as the second embodiment Flowchart illustrating a divided block decoding method described as the fourth embodiment Flowchart illustrating a divided block decoding method described as the sixth embodiment Flowchart illustrating a divided block decoding method described as the eighth embodiment Flowchart illustrating a divided block decoding method described as the tenth embodiment (A) (b) (c) The figure which shows the hardware constitutions by which this invention is performed (A) (b) The figure which shows the reference positional relationship of a surrounding block. (A) (b) The figure which shows the example of a division | segmentation of a block Diagram showing conversion table of block positional relationship and identification value Flow chart showing frame encoding method Flow chart showing frame decoding method A diagram showing a configuration example of a frame bitstream (A) (b) The figure which shows the structural example of the bit stream of a lattice block The figure which shows the example of the division | segmentation state of a lattice block, and a division | segmentation flag Flowchart explaining a method for obtaining a predicted image quality control parameter for each block size of a divided block Flowchart for determining predictive image quality control parameter from intra prediction method The figure which shows the reference relationship of the image quality control parameter in each block The figure which shows the operation example of the division | segmentation block encoding method demonstrated as Example 5. FIG. (A) (b) An example of a method for obtaining a predicted image quality control parameter Index indicating the prediction direction of intra prediction First flowchart for determining a predicted image quality control parameter from a plurality of attributes Second flowchart for determining a predicted image quality control parameter from a plurality of attributes Third flowchart for determining a predictive image quality control parameter from a plurality of attributes Fourth flowchart for determining a predicted image quality control parameter from a plurality of attributes

  Hereinafter, the present invention will be described in detail based on the 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.

<Example 1>
The encoding method of the encoding object block in a present Example is demonstrated using FIG.1, FIG.11, FIG.15 and FIG.

  FIG. 11A shows a hardware configuration for executing an encoding program including the encoding method according to the present invention. This encoding program is recorded in a hard disk device (hereinafter abbreviated as HDD) 1104, and is loaded into the RAM 1102 when the encoding program is started. The CPU 1101 executes the following steps to perform encoding processing. . In this embodiment, the input image data is stored in the HDD 1104, and the encoded bit stream is also recorded in the HDD 1104. An operation of performing encoding in the above configuration will be described. In the present embodiment, the frame is divided into lattice blocks obtained by dividing a frame into lattices of 64 × 64 pixels (hereinafter, a block composed of n × n pixels in the vertical and horizontal directions is expressed as n × n), and the lattice block is further divided. In the following description, it is assumed that the block is divided into divided blocks using the area quadtree structure, and the divided block is used as an encoding target block. The minimum divided block is 8 × 8. FIG. 13A shows an example of a state where the frame is divided into lattice blocks, and FIG. 13B shows an example of a state where the lattice block is further divided into divided blocks. Even when the lattice block is not divided, the processing is performed assuming that one 64 × 64 divided block exists in the lattice block.

  FIG. 15 is a flowchart illustrating a flow for encoding the entire frame. In S15010, frame header information is encoded. The frame header information includes a frame synchronization code, frame type information indicating whether the frame is an intra frame or an inter frame, and the like.

  In S15020, a grid block to be processed (hereinafter, target grid block) is set. The first target lattice block is the upper left lattice block in the frame. Thereafter, each time S15020 is executed, lattice blocks are set in raster order. For example, the lattice block numbers (B1, B2,...) Shown in FIG.

  In S15030, a prediction mode to be described later is determined for each division state and division block of the target lattice block. The target lattice block is divided using a region quadtree structure as shown in FIG. Further, in this embodiment, the prediction mode indicates an intra prediction mode and an inter prediction mode. The intra prediction mode is H.264. The intra prediction image block is generated by referring to the divided blocks around the divided block to be processed as in H.264, and the inter prediction mode is a mode in which a motion vector is determined to generate a motion compensated prediction image block. is there.

In S15031, the prediction mode information and divided block information of the divided blocks in the target lattice block are encoded. The prediction mode information is a flag for identifying the prediction mode and additional information related to the prediction mode. For example, in the intra prediction mode, the additional information is intra prediction method information, and in the inter prediction, it is motion vector information or motion vector derivation block identification information for identifying a block having the same motion vector from neighboring blocks. As an example of the motion vector derivation block identification information, the motion vector derivation block identification information includes a 1-bit motion vector identical flag and a motion vector prediction direction flag, respectively. When the motion vector identical flag is 1, the motion vector is not embedded and the motion vector prediction direction flag is embedded. When the motion vector identical flag is 1, the motion vector is embedded, and the motion vector prediction direction flag is not embedded. When the motion vector prediction direction flag is 1, the motion vector of the left block is set as the motion vector of the target block,
In the case of 0, the motion vector of the upper block is the motion vector of the target block. When embedding the motion vector derivation block identification information, the motion vector can be derived from the motion vector derivation block identification information in the decoding process without embedding the motion vector information in the bit stream. In this embodiment, the motion vector prediction direction flag is handled as a 1-bit flag, but the present invention is not limited to this. In addition to the motion vector prediction direction flag upper block and left block, multi-value data may be used to select from a plurality of block candidates such as an upper left block and an upper right block. Intra prediction method information is described in H.264. Similar to the prediction mode with the H.264 16 × 16 intra prediction block, for example, the information identifies horizontal prediction, vertical prediction, DC prediction, and planar prediction.

  A prediction mode information code obtained by variable-length encoding this prediction mode information is embedded in the bitstream. The division block information is information indicating a division state as to whether or not each node (block) in the area quadtree is further divided, and the division block information for one node (block) is used as a division flag. For example, if the lattice block is divided as shown in FIG. 13B, the division state of the hierarchy for each block size is expressed as shown in FIG. (A), (b), (c), and (d) of FIG. 19 show the division states of 64 × 64, 32 × 32, 16 × 16, and 8 × 8 blocks, respectively. Further, the case where the block is divided is expressed as (1), and the case where the block is not divided is expressed as (0). N1, N2,. . , N21 represent node (block) numbers of the area quadtree structure. The number is the order in which the area quadtree structure is scanned by depth-first search for each hierarchy, and the order in which scanning is performed in the order of upper left, upper right, lower left, lower right within the same block (hereinafter referred to as area quadtree structure). Are expressed in order). “-” Indicates that there is no node (block). N1, N2,. . . , N21 split flag information is transmitted to the decoder as split block information, so that the decoder can reconstruct the area quadtree structure. The prediction mode information code and the division flag are embedded in the bit stream for each divided block, for example, in a structure as shown in FIG. In addition, N1, N2,. . . , N21 correspond to the node (block) numbers in FIG. 19, and b1, b2,. . . , B16 correspond to the divided block numbers in FIG. N1, N2,. . , N21, the prediction mode information is the smallest divided block that is not further divided, that is, b1, b2,. . . , B16 are added only to the divided blocks.

  In S15040, a target divided block to be processed is set. At the first time, the upper left block in the lattice block is set. Thereafter, each time S15040 is executed, a divided block is set according to the order of the area quadtree structure. For example, the division block numbers (b1, b2,...) Shown in FIG.

  In S15050, the target divided block set in S15040 is encoded. Details will be described later. In S15060, it is determined whether all the divided blocks in the target lattice block have been processed. If all the divided blocks have been processed, the process proceeds to S15070. If all the divided blocks have not been processed, the process proceeds to S15070. The process returns to S15040. When processing of all the divided blocks for one lattice block is completed, a bit stream of the lattice block having a structure as shown in FIG. 18A is generated.

  In step S15070, it is determined whether the processing of all the lattice blocks in the frame has been completed. If the processing of all the lattice blocks has been completed, this flow ends. If the processing of all the lattice blocks has not been completed, Returns to S15020. When all the lattice blocks have been processed, a bit stream of a frame having a structure as shown in FIG. 17A is generated.

  FIG. 1 is a detailed flowchart of the divided block encoding method executed in S15050 of FIG.

  In S1010, the attribute information of the target divided block is acquired. In this embodiment, the block size is attribute information, but the present invention is not limited to this. For example, the attribute information includes block mode information for identifying intra / inter prediction, a motion vector, an intra prediction method in intra prediction, and the like. These attributes may be used alone or in combination. The image quality control parameters of the divided blocks that have the same attribute and are encoded immediately before may be used as the predicted image quality control parameters. The motion vectors do not have to be exactly the same, and may be regarded as the same if the absolute value of the vector difference is equal to or less than a certain value.

  In S1020, the image quality control parameter of the target divided block is calculated. The image quality control parameter is calculated, for example, so that the code amount of the frame becomes a preset target code amount. At this time, for the purpose of improving the image quality of a region having a specific image component such as a human face or the edge of an object, the image quality control parameter may be adjusted so that the image quality parameter is different for each region.

  In S1030, the image quality control parameter of the divided block having the same attribute as that of the target divided block and encoded immediately before is acquired as the prediction control parameter. In this embodiment, since the block size is used as attribute information, an image quality control parameter is recorded and updated for each block size, and a format in which the block size is referred to is used. Details will be described later.

  In S1040, the difference value between the image quality control parameter of the target divided block and the predicted image quality control parameter is calculated, and the difference value of the image quality control parameter is calculated.

  In step S1050, the image quality control parameter difference value is encoded to generate an image quality control parameter difference value code. The image quality control parameter difference value code is embedded in the bitstream with the structure shown in FIG.

  In S1060, the image data of the target divided block is encoded. In this embodiment, residual data of the target divided block is generated according to the prediction mode, and orthogonal transformation, quantization, and variable length coding are executed to encode the residual data and generate variable length coding coefficient data. To do. At this time, the variable length coding coefficient data is embedded in the bitstream with the structure shown in FIG.

  FIG. 20 is a flowchart illustrating a method for obtaining a predicted image quality control parameter for each block size shown in S1030 of FIG. In this flowchart, the final image quality control parameter 64, the final image quality control parameter 32, the final image quality control parameter 16, and the final image quality control parameter 8 are used. These parameters are common parameters for processing within a frame, and are initialized at the start of decoding or for each processing unit such as a frame or a slice.

In step S20001, the block size of the processing target block is determined, and processing for each block size is executed. Specifically, when the block size is 64 × 64, 32 × 32, 16 × 16, and 8 × 8, S20002 and S20003, S20004 and S20005, S20006 and S20007, and S20008 and S20009 are executed.
In S20002, the final image quality control parameter 64 is set as a predicted image quality control parameter.

  In S20003, the image quality control parameter of the target divided block is recorded in the final image quality control parameter 64 × 64. By recording values in this way, it is possible to easily refer to the next 64 × 64 block processing.

  Hereinafter, the same processing is performed for S20004 and S20005, S20006 and S20007, and S20008 and S20009 for each of 32 × 32, 16 × 16, and 8 × 8. Here, the acquired predicted image quality control parameter is used in S1040.

  FIG. 22 shows the relationship of blocks referred to as the predicted image quality control parameter. The arrows in the figure indicate that, for example, when B2 is a processing target divided block, the image quality control parameter of B1 last encoded with a block of the same size is used as the predicted image quality control parameter of B2. If C5 is a processing target divided block, the C4 image quality control parameter last encoded with the same size block is set as the C5 predicted image quality control parameter. In this embodiment, the image quality control parameters of the blocks having different sizes are referred to as the predicted image quality control parameters in an independent form for each size.

  Blocks having the same attribute are often encoded with the same or similar image quality control parameters. For example, an image block with many low-frequency components has a large block size, and an image block with many high-frequency components such as edges has a small block size, and coding tends to improve coding efficiency. Therefore, the block size at the time of block division is easily determined according to the tendency. In addition, when image quality control parameters are controlled to improve the image quality of a region having a specific image component, the image quality control parameters are likely to have similar values for each block size. Therefore, by employing the method described in the present embodiment, the absolute value of the difference value is simply smaller than when the difference value is calculated using the image quality control parameter of the block encoded immediately before as the predicted image quality control parameter. That is, the code amount when the image quality control parameter is encoded becomes small.

  In this embodiment, the lattice block size is 64 × 64 and the minimum divided block size is 8 × 8. However, the present invention is not limited to this. Further, the encoding method in the present invention may be performed with a configuration having dedicated hardware such as an orthogonal transformer 1003, a quantizer 1004, and a variable length encoder 1005, as shown in FIG. The calculation of the image quality control parameter according to the present invention may be configured by dedicated hardware such as a code amount controller (not shown). At this time, the flow shown in FIG. 15 and FIG. 1 is performed except that the CPU does not execute orthogonal transform, quantization, variable length coding, and control processing, but the CPU causes each hardware to execute each processing. The same operation is performed.

<Example 2>
An image decoding method for decoding encoded data encoded by the image encoding method described in the first embodiment will be described with reference to FIGS. 6, 11, 16, and 20.

  FIG. 11A shows a configuration for executing a decoding program including the decoding method according to the present invention. The decryption program recorded in the HDD 1104 is loaded into the RAM 1102, and the decryption process is performed by the CPU 1101 executing each step of the flow described later. The input code data is read from the HDD 1104, and the decoded image is output to the display via the VRAM 1105. Although the decoding process is performed for each block, the decoded image of the block obtained by decoding each block is copied to an appropriate position in the RAM 1102, and the decoding process of the block for one frame is completed. Thus, a decoded image of one frame is completed. The generated frame is then output to the VRAM 1105.

  The input bit stream will be described as the bit stream generated in the first embodiment.

  FIG. 16 shows a flow for decoding an input bit stream. Each lattice block is decoded in raster scan order.

  In S16010, the frame header information shown in FIG. 17A is decoded. In S16020, a target grid block to be processed is set. The target grid block to be processed for the first time is the grid block at the upper left of the screen. Thereafter, each time S16020 is executed, target grid blocks are set in raster order. For example, the lattice block numbers (B1, B2,...) Shown in FIG.

  In S16030, the division state of the target lattice block and the prediction mode for each division block are decoded. Here, the area quadtree structure of the divided block is restored based on the division flag. Further, prediction mode information for each divided block is also decoded, and thereafter, attribute information of the divided block can be acquired.

  In S16040, a target divided block to be processed is set. The divided blocks are set based on the order of the area quadtree structure restored in S16030. For example, the division block numbers (b1, b2,...) Shown in FIG.

  In S16050, the image data of the target divided block is decoded. Details will be described later with reference to FIG.

  In S16060, it is determined whether all the divided blocks in the lattice block have been processed. If all the divided blocks have been processed, S16070 is executed, and if all the divided blocks have not been processed. Returns to S16040.

  In step S16070, it is determined whether processing of all lattice blocks in the frame has been completed. If processing of all lattice blocks has been completed, this flow ends, and processing of all lattice blocks has not been completed. In this case, the processing of S16020 to S16070 is executed again.

  FIG. 6 is a detailed flowchart of the process executed in S16050 of FIG. In S6010, the attribute information of the target divided block is acquired. In this embodiment, the block size is used as attribute information, but the present invention is not limited to this. In S6020, the image quality parameter difference value code is decoded to generate an image quality parameter difference value.

  In S6030, the divided block has the same attribute information as the attribute information of the target divided block acquired in S6010, and the image quality control quality parameter of the divided block decoded immediately before is set as the predicted image quality control parameter. In this embodiment, since the block size is used as attribute information, the final image quality control parameter for each block size is recorded and referred to.

  In S6040, the image quality control parameter of the processing target block is calculated by adding the image quality control parameter difference value to the predicted image quality control parameter.

  In S6050, the image data of the target divided block is decoded. In the decoding process of the processing target block according to the present embodiment, the variable-length coding coefficient data is subjected to variable-length decoding, inverse quantization, and inverse orthogonal transform to generate residual data. Further, a predicted image is generated according to the prediction mode, and the decoded image of the divided block is obtained by adding the predicted image to the residual data. At the time of inverse quantization, the image quality control parameter calculated in S6040 is used as a quantization parameter.

  By adopting the method described in the present embodiment, it is possible to decode the stream encoded by the image encoding method described in the first embodiment.

  The decoding method according to the present invention has dedicated hardware such as an inverse orthogonal transformer 1304, an inverse quantizer 1305, a variable length decoder 1306, and a code amount controller (not shown) as shown in FIG. It may be done in a configuration. At this time, the inverse orthogonal transform, inverse quantization, and variable length decoding processing are not executed by the CPU 1101, but the CPU 1101 causes each hardware to execute each processing, and the flow shown in FIG. 6 and FIG. Similar processing is performed.

<Example 3>
An image encoding method in this embodiment will be described with reference to FIG. Note that, unless otherwise specified, the configuration and operation of each step are the same as those in the first embodiment.

  FIG. 2 is a detailed flowchart of the divided block encoding method, and S2021 is added after S1020 to FIG. In the present embodiment, description will be made assuming that the target divided block is an inter block.

  In S2021, one block having attribute information closest to the attribute information of the target divided block is selected from the peripheral blocks. FIG. 12B shows the relationship between the target divided blocks and the motion vectors of the peripheral blocks. In this embodiment, the peripheral blocks are left, upper, and upper right blocks as shown in FIG. 12B, but are not limited to this combination. In addition, selecting the block having the closest attribute information selects the block having the motion vector that minimizes the difference between the motion vector of the target divided block and the motion vector of the peripheral block, but is not limited thereto. When there is motion vector derivation block identification information for identifying a block having the same motion vector from among neighboring blocks, the block may be selected using this information. Referring to FIG. 12B as an example, the motion vector of the target divided block is VCurr, and the motion vectors of the left, upper, and upper right blocks are VLeft, VTTop, or VTTopRight, respectively. In FIG. 12B, VCurr and VTTop have the same vector value, so the upper block is selected. In the present embodiment, when there are a plurality of blocks having the closest attribute information or the same attribute information, the block having the closest spatial distance is selected. However, the present invention is not limited to this.

  In S2030, the image quality control parameter of the block selected in S2021 is set as the predicted image quality control parameter.

  Blocks having the same attribute existing around the processing target block are often encoded with the same or close values of image quality control parameters. For example, blocks with the same motion vector are likely to be included in the same object, and when performing processing to improve the image quality of a specific object, blocks with the same motion vector have similar image quality control. There is a high possibility of encoding with parameters. Therefore, compared to the case where the difference value is calculated using the image quality control parameter of the block encoded immediately before as the predicted image quality control parameter, the absolute value of the difference value is smaller in this method and the image quality control parameter is encoded. The code amount at that time becomes small. Therefore, it is possible to encode an image with the same image quality with a smaller bit rate.

<Example 4>
An image decoding method for decoding encoded data encoded by the image encoding method described in the third embodiment will be described with reference to FIG. Note that, unless otherwise specified, the configuration and operation of each step are the same as those in the second embodiment. In the following description, it is assumed that the input bit stream is the bit stream generated in the third embodiment.

  FIG. 7 is a detailed flowchart of the divided block decoding process shown in S16050 of FIG. Moreover, S7021 is added after S6020 with respect to FIG. In this embodiment, description will be made assuming that the processing target block is an inter block.

  In step S7021, one block having attribute information closest to the attribute information of the target divided block is selected from the peripheral blocks. In FIG. 12 (b), VCurr and VTTop are the same vector as in the third embodiment, so the upper block is selected. At this time, a block may be selected by directly comparing the motion vectors, and when there is information for identifying a block having the same motion vector from the neighboring blocks, the block is selected using the information. May be. In this embodiment, when there are a plurality of blocks having the closest attribute information or the same attribute information, the block having the closest spatial distance is selected. However, the present invention is not limited to this.

  In S7030, the image quality control parameter of the block selected in S6021 is set as the predicted image quality control parameter. The predicted image quality control parameter determined here is used in S6040. By adopting the method described in the present embodiment, it is possible to decode the stream encoded by the image encoding method described in the third embodiment.

<Example 5>
The image encoding method in the present embodiment will be described with reference to FIG. Note that, unless otherwise specified, the configuration and operation of each step are the same as those in the first embodiment.

  FIG. 3 is a detailed flowchart of the divided block encoding method shown in S15050 of FIG. Further, S3021 is added after S1020 with respect to FIG.

  In S3021, one or more blocks having the same attribute information are selected from the peripheral blocks. In this embodiment, the peripheral blocks are the left, upper, and upper right blocks with respect to the processing target block as shown in FIG. 12A, but are not limited to this combination. In this embodiment, block mode information is attribute information. That is, if the target divided block is an intra block, a surrounding intra block is selected, and if the target divided block is an inter block, a surrounding inter block is selected.

  In step S3030, an average value of the image quality control parameters of the selected one or more blocks is calculated and set as a predicted image quality control parameter. In this embodiment, the average value is used as the predicted image quality control parameter, but the present invention is not limited to this, and for example, a median value may be used.

  FIG. 23 shows an operation example of this embodiment. In FIG. 23A, since the left block of the target divided block is an inter block, the left block is selected. In FIG. 23B, since the block above the target divided block is an inter block, the upper block is selected. In FIG. 23C, since the left block and the upper block of the target divided block are inter blocks, the average value of the left block and the upper block is selected. In FIG. 23D, since the left block and the upper block of the target divided block are not inter blocks, the left block is selected.

  Blocks having the same attribute existing around the target divided block are often encoded with the same or close values of image quality control parameters. For example, a block at an edge portion of a moving subject is easily encoded by intra coding, and a block in the background or the subject is easily encoded by inter coding. When performing code amount control to increase the image quality by giving priority to the edge portion, it is an effective means to set the image quality control parameter so that the image quality is high for the intra-coded block. That is, the image quality control parameters of the blocks having the same block mode information tend to have similar values. Therefore, by using the method described in the present embodiment, the absolute value of the difference value becomes smaller than when the difference value is simply calculated using the image quality control parameter of the block encoded immediately before as the predicted image quality control parameter. The amount of code when the image quality control parameter is encoded becomes small.

<Example 6>
An image decoding method for decoding encoded data encoded by the image encoding method described in the fifth embodiment will be described with reference to FIG. Note that, unless otherwise specified, the configuration and operation of each step are the same as those in the second embodiment. In the following description, it is assumed that the input bit stream is the bit stream generated in the fifth embodiment.

  FIG. 8 is a detailed flowchart of the divided block decoding process shown in S16050 of FIG. Moreover, S8021 is added after S6020 with respect to FIG.

  In step S8021, one or more blocks having the same attribute information are selected from the neighboring blocks. In this embodiment, the peripheral blocks are the left, upper, and upper right blocks with respect to the processing target block as shown in FIG. 12A, but are not limited to this combination. If the target divided block is an intra block, a peripheral intra block is selected. If the target divided block is an inter block, a peripheral inter block is selected.

  In step S8030, the average of the image quality control parameters of the selected one or more blocks is calculated and set as the predicted image quality control parameter. In this embodiment, the average value is used as the predicted image quality control parameter, but the present invention is not limited to this, and for example, a median value may be used. The predicted image quality control parameter determined here is used in S6040.

  By adopting the method described in the present embodiment, it is possible to decode the stream encoded by the image encoding method described in the fifth embodiment.

<Example 7>
The image encoding method of the present embodiment will be described with reference to FIG. Note that, unless otherwise specified, the configuration and operation of each step are the same as those in the first embodiment.

  FIG. 4 is a flowchart showing the divided block coding method shown in S15050 of FIG. Further, with respect to FIG. 1, S4021 is added after S1020, and S4022 and S4023 are added.

  In step S4021, a predicted image quality control parameter determination method is selected from a plurality of candidates. As a predictive image quality control parameter determination method in the present embodiment, (0) the image quality control parameter of the divided block existing spatially to the left of the target divided block is set as the predictive image quality control parameter. (1) The average of the image quality control parameters of the divided blocks having the same attribute information as the target divided block among the peripheral divided blocks is set as the predicted image quality control parameter. (2) The median value of the image quality control parameters of the divided blocks having the same attribute information as the target divided block among the peripheral divided blocks is set as the predicted image quality control parameter. (3) The image quality control parameter of the divided block having the same attribute information as that of the target divided block and encoded immediately before is set as the predicted image quality control parameter. However, the present invention is not limited to this combination, and other determination methods may be used, or only some of them may be used. As specific processing in S4021, these prediction image quality control parameter determination methods are tried, and a prediction image quality control parameter determination method that minimizes the difference value between the image quality control parameter of the target divided block and the prediction image quality control parameter is selected.

  In step S4022, an identification value that identifies a method for determining the predicted image quality control parameter is determined. In this embodiment, the index value corresponding to the predicted image quality control parameter determination method determined in S4021 is used as the identification value.

  In step S4023, the identification value is encoded to generate an identification value code.

  In step S4030, the predicted image quality control parameter calculated by the predicted image quality control parameter determination method selected in step S4021 is acquired.

  According to the present embodiment, the bit stream structure of the lattice block is as shown in FIG. By transmitting this bit stream to the decoder, the decoder can identify the prediction image quality control parameter determination method.

  According to the present embodiment, it is possible to select an optimal predictive image quality control parameter determination method. In the present embodiment, the optimum prediction image quality control parameter determination method is selected for each divided block, and the identification value is encoded. However, the present invention is not limited to this, and an optimum prediction image quality control parameter determination method may be used in units of lattice blocks, slices, frames, and sequences, and an identification value may be embedded in the header unit of the encoding unit. For example, the identification value is embedded in the frame header of FIG.

<Example 8>
An image decoding method for decoding encoded data encoded by the image encoding method described in the seventh embodiment will be described with reference to FIG. Note that, unless otherwise specified, the configuration and operation of each step are the same as those in the second embodiment. In the following description, it is assumed that the input bit stream is the bit stream generated in the seventh embodiment.

  FIG. 9 is a detailed flowchart of the divided block decoding process shown in S16050 of FIG. 9 is different from FIG. 2 in that S9021 is added after S6020. In the present embodiment, it is assumed that the bit stream shown in FIG. 18B generated in the seventh embodiment is decoded.

  In S6010, the attribute information of the target divided block is acquired. In this embodiment, the block mode information for identifying intra / inter prediction is attribute information. However, the present invention is not limited to this. For example, the block size, motion vector, intra prediction method in intra prediction, simple substance, or combination thereof is used. It may be used.

  S9021 decodes the identification value code for identifying the prediction image quality control parameter determination method, and generates an identification value. In the present embodiment, this identification value corresponds to (0), (1), (2), and (3) described in the seventh embodiment, but is not limited thereto.

  In step S9030, the predicted image quality parameter is determined using the predicted image quality control parameter determination method identified by the identification value. The predicted image quality control parameter determined here is used in S6040.

  With the decoding method described in the present embodiment, it is possible to decode the stream encoded by the image encoding method described in the seventh embodiment.

<Example 9>
The image encoding method of the present embodiment will be described with reference to FIG. Note that, unless otherwise specified, the configuration and operation of each step are the same as those in the first embodiment.

  FIG. 5 is a detailed flowchart of the divided block encoding process shown in S15050 of FIG. Further, with respect to FIG. 1, S5021 is added after S1020, and S5041 and S5042 are added after S5040.

  In S5021, a block having the same attribute information as the target divided block is selected from the peripheral divided blocks. In this embodiment, as shown in FIG. 12A, the blocks to be processed are left, upper, and upper right blocks, but the present invention is not limited to this. Further, when the left, upper, and upper right blocks are all in different modes, all are selected in this embodiment, but the present invention is not limited to this.

  In S5030, all the difference values between the image quality control parameter of the target divided block and the image quality control parameter of the selected block are calculated, and the block having the smallest difference value is selected.

  In S5040, the difference value calculated using the image quality control parameter of the block selected in S5030 is acquired. This difference value is encoded in S1050.

  In S5041, an identification value for identifying an image quality control parameter prediction block from a plurality of blocks is calculated. For example, this identification value can be calculated using the table in FIG. “Presence / absence of selection of identical attribute block” in FIG. 14 indicates whether or not the divided blocks existing on the left, upper, and upper right have been selected in S5021. The identification code in the table is assigned according to the “presence / absence of selection of the same attribute block”, and the identification code corresponding to the left, upper, and upper right divided blocks selected in S5021 is used as the identification value. “-” In the table indicates a state in which no identification code is allocated because no block is selected. “X” indicates that the block can be uniquely determined without coding the identification value.

  In S5042, the identification value is encoded to generate an identification value code. In the present embodiment, the identification value is encoded by the encoding method described in the column “Encoding Method” in FIG. 14 according to “selection / non-selection of same attribute block”. In cases 1 and 8 of FIG. 14, a variable-length Gombb code is generated and embedded in the bitstream. In cases 2, 3, and 5 in FIG. 14, the value is embedded in the bitstream as 1-bit fixed length data. In cases 4, 6, and 7, since the block can be uniquely specified, the identification value code is not embedded in the bit stream. Note that the encoding method is not limited to this.

  According to the present embodiment, for each divided block, it is possible to select a peripheral divided block having a predictive image quality control parameter that can minimize the code amount of the image quality control parameter of the target divided block.

<Example 10>
An image decoding method for decoding encoded data encoded by the image encoding method described in the ninth embodiment will be described with reference to FIG. Note that, unless otherwise specified, the configuration and operation of each step are the same as those in the second embodiment. In the following description, it is assumed that the input bit stream is the bit stream generated in the ninth embodiment.

  FIG. 10 is a detailed flowchart of the divided block encoding process shown in S16050 of FIG. Further, in FIG. 6, S10021, S10022, and S10023 are added after S6020.

  In S10021, a block having the same attribute information as the processing target block is selected from the peripheral divided blocks. In the present embodiment, as shown in FIG. 12A, the block to be processed is the left, upper, and upper right blocks, but it is not limited to this as long as it is the same as the encoding method. Also, when the left, upper, and upper right blocks are all in different modes, all are selected in the present embodiment, but the encoding method may be the same, and the present invention is not limited to this.

  In S10022, the identification value is decoded from the identification value code. In the present embodiment, the identification value is decoded by a decoding method corresponding to the encoding method described in the “encoding method” column according to “selection / non-selection of same attribute block” in FIG. 14, but the present invention is not limited to this. Absent. In cases 2, 3, and 5 in FIG. 14, since the block can be uniquely identified and the identification value code is not embedded, the decoding process of the identification value code is not performed.

  In S10023, a divided block used for calculation of the predicted image quality control parameter is selected based on the identification value. Specifically, in the present embodiment, the left, upper, and upper right blocks are identified from the “identical attribute block selection presence / absence” and the identification value using the table of FIG. Even in the case where the identification value code is not decoded in S10022, one prediction block can be selected from the table of FIG. 14 and “whether or not the same attribute block is selected”. The block selected here is used in S6030.

  With the decoding method described in the present embodiment, the stream encoded by the image encoding method described in the ninth embodiment can be decoded.

<Example 11>
The image encoding method of the present embodiment will be described with reference to FIG. Note that, unless otherwise specified, the configuration and operation of each step are the same as those in the first embodiment.

  In the present embodiment, description will be made assuming that the target divided block is an intra block. The prediction method for intra prediction is described in H.264. Although it is described as DC prediction, horizontal direction prediction, vertical direction prediction, and plane prediction as in the case of H.264 16 × 16 intra prediction block, the present invention is not limited to this.

FIG. 21 is a flowchart for determining the prediction image quality control parameter from the intra prediction method in S1030 of FIG. 1. In S21001, if the intra prediction method of the target divided block is horizontal prediction, S21002 is performed by DC prediction or plane prediction. In the case of, S21003 is executed, and in the case of vertical direction prediction, S21004 is executed.

  In S21002, the image quality control parameter of the divided block existing on the left side is set as the predicted image quality control parameter.

  In S21003, the average value of the image quality control parameters of the left divided block and the upper divided block is set as the predicted image quality control parameter.

  In S21004, the image quality control parameter of the divided block existing above is set as the predicted image quality control parameter.

  However, the combination of the divided blocks used for the calculation of the intra prediction method and the predicted image quality control parameter is not limited to this. For example, when the intra prediction method has 34 directions as shown in FIG. 25, the predicted image quality control parameter is calculated as shown in FIG. In this case, for example, when the index indicating the prediction direction is 18, 10, or 19, the average value of the image quality control parameters of the divided block existing on the left and the divided block existing above is used as the predicted image quality control parameter of the target divided block. Is calculated (top_left and top shown in FIG. 24B).

  Blocks in the prediction direction of intra prediction are likely to be included in the same object, and can be encoded with similar image quality control parameters when performing processing to improve the image quality of a specific object. High nature. Therefore, compared with the case where the difference value is calculated using the image quality control parameter of the block encoded immediately before as the predicted image quality control parameter, the absolute value of the difference value is smaller in the method described in the present embodiment, and the image quality control is performed. The code amount when the parameter is encoded becomes small.

<Example 12>
An image decoding method for decoding encoded data encoded by the image encoding method described in the eleventh embodiment will be described with reference to FIG. Note that, unless otherwise specified, the configuration and operation of each step are the same as those in the second embodiment. Further, the description will be made assuming that the input bit stream is the bit stream generated in the eleventh embodiment.

  In the present embodiment, description will be made assuming that the target divided block is an intra block. The prediction method for intra prediction is described in H.264. Similarly to the H.264 16 × 16 intra prediction block, DC prediction, horizontal direction prediction, vertical direction prediction, and plane prediction are assumed, but the present invention is not limited thereto.

  In S6030 of FIG. 6, the prediction image quality control parameter is determined from the intra prediction method according to the flowchart of FIG.

  In S21001, if the intra prediction method of the target divided block is horizontal prediction, S21002 is executed, S21003 is executed in the case of DC prediction or plane prediction, and S21004 is executed in the case of vertical prediction.

  In S21002, the image quality control parameter of the divided block existing on the left side is set as the predicted image quality control parameter.

  In S21003, the average value of the image quality control parameters of the left divided block and the upper divided block is set as the predicted image quality control parameter.

  In S21004, the image quality control parameter of the divided block existing above is set as the predicted image quality control parameter.

However, the combination of the divided blocks used for calculating the intra prediction method and the predicted image quality control parameter is not limited to this as long as it is the same as the encoding method.
By adopting the method described in the present embodiment, it is possible to decode the stream encoded by the image encoding method described in the eleventh embodiment.

<Example 13>
An image encoding method in this embodiment will be described with reference to FIG. Unless otherwise specified, the configuration and the operation of each step are the same as those in the first and third embodiments.

  In S2021, a block having the same attribute information of the target divided block is selected from the peripheral blocks. In the present embodiment, it is a motion vector. In the example of FIG. 12B, VCurr and VTTop have the same vector value, so the upper block is selected.

  In S2030, the image quality control parameter of the block selected in S2021 is set as the predicted image quality control parameter.

  In the present embodiment, when the difference between the motion vector of the target divided block calculated in S2021 and the motion vector of the peripheral block is the same, the image quality control parameter difference value code is not embedded. When the image quality control parameter difference value code is not embedded, S1040 and S1050 are skipped, and the image quality control parameter difference value code is not embedded. In the image block encoding in S1060, the image quality control parameter calculated in S1020 is not used, and the image block is encoded using the predicted image quality control parameter determined in S2030 as a parameter for image quality control. Become.

  Blocks having the same attribute existing around the processing target block should be encoded with the same image quality control parameter. For example, blocks having the same motion vector are more likely to be included in the same object, and image quality differences are not generated in the same object when encoded with the same image quality control parameters. By not embedding the difference value of the image quality control parameter in the stream, it is possible to encode an image with the same image quality with a smaller bit rate.

  Furthermore, there may be a mode in which a block having the same motion vector exists in the peripheral blocks, and a code for identifying the position of the block and the motion vector is inserted in the bit stream. When there is a block having the same motion vector, processing in a form in which the above-described image quality control parameter difference value code is not embedded is performed. At this time, no motion vector is embedded in the target divided block. By adopting this embodiment, it is possible to further reduce the amount of code when the image quality control parameter is encoded.

<Example 14>
An image decoding method for decoding a stream encoded by the image encoding method described in the thirteenth embodiment will be described with reference to FIG. Unless otherwise specified, the configuration and operation of each step are the same as those in the second and fourth embodiments.

  In S7021, a block having the same attribute information of the target divided block is selected from the peripheral blocks. In the present embodiment, it is a motion vector. In the example of FIG. 12B, VCurr and VTTop have the same vector value, so the upper block is selected.

  In S7030, the image quality control parameter of the block selected in S7021 is set as the predicted image quality control parameter. The predicted image quality control parameter determined here is used in S6040.

  In the present embodiment, when the difference between the motion vector of the target divided block calculated in S7021 and the motion vector of the peripheral block is the same, the image quality control parameter difference value code is not decoded. When the image quality control parameter difference value code is not decoded, S6020 and S6040 are skipped, and the decoding of the image block in S7040 uses the predicted image quality control parameter determined in S7030 as a parameter for image quality control. Become.

  By adopting the method described in the present embodiment, it is possible to decode the stream encoded by the image encoding method described in the third embodiment.

  Furthermore, even when there is a block having the same motion vector among the peripheral blocks, and a code identifying which block has the same motion vector is inserted in the bitstream, the above-described case is also possible. The image quality control parameter difference value code may not be decoded. In this case, the predicted image quality control parameter is an image quality control parameter for blocks having the same motion vector identified by the above-described code. These processes need to correspond to the process of the encoding method.

<Example 15>
An image encoding method in this embodiment will be described with reference to FIG. Unless otherwise specified, the configuration and the operation of each step are the same as those in the first and third embodiments.

  In S2021, the screen control parameter of the block having the same attribute information of the target divided block is selected from the peripheral divided blocks. In this embodiment, the selection is performed by a combination of a plurality of attributes, that is, a prediction mode, an intra prediction method, and presence / absence of a divided block to be referred to. The plurality of attributes are not limited to this, and the above-described motion vector and block size may be used, or other attributes may be used.

FIG. 26 shows a flowchart of the selection. For selection, first, it is determined whether or not the prediction mode of the target divided block is intra, and if it is intra coding, screen control parameters are selected using the intra prediction method as an attribute. For other prediction modes, prediction screen control parameters are selected using the prediction mode itself as an attribute. In addition to the intra prediction mode, a prediction mode such as an inter prediction mode or a skip mode corresponds to this, but is not limited to this. Screen control parameters are selected using the presence or absence of a block that can be referred to last as an attribute.
Hereinafter, it demonstrates in detail according to a flowchart.

  In S26001, it is determined whether the prediction mode of the target divided block is the intra prediction mode or not. If the prediction mode is the intra prediction mode, the process proceeds to S26002, and if not, the process proceeds to S26003.

  In S26002, if the intra prediction method is horizontal prediction, the left image quality control parameter is set as the predicted image quality control parameter in S26009. For vertical prediction, the upper image quality control parameter is set as the predicted image quality control parameter in S26010. Further, in the case of DC prediction or planar prediction, the average value of the left image quality control parameter and the upper image quality control parameter is set as the predicted image quality control parameter in S26011. Thereafter, the predicted image quality control parameter selection process is terminated.

  In S26003, it is determined whether or not the prediction mode of the target divided block matches the prediction mode of the upper divided block. If they match, the process proceeds to S26004, and if not, the process proceeds to S26005.

  In step S26004, it is determined whether the prediction mode of the target divided block matches the prediction mode of the left divided block. If the prediction mode matches, the process proceeds to S26010. Otherwise, the process proceeds to S26011.

  In step S26005, it is determined whether or not the prediction mode of the target divided block matches the prediction mode of the left divided block. If they match, the process proceeds to S26009, and if not, the process proceeds to S26006.

  In step S26006, it is determined whether there is a divided block that can be referred to above the target divided block. If there is a divided block, the process proceeds to S26007. Otherwise, the process proceeds to S26008. In S26007, it is determined whether there is a divided block that can be referred to on the left side of the target divided block. If there is a divided block, the process proceeds to S26010. Otherwise, the process proceeds to S26011.

  In step S26008, it is determined whether there is a divided block that can be referred to on the left side of the target divided block. If there is a divided block, the process proceeds to S22009, and if not, the process proceeds to S22012. In S26012, there is no divided block to be referenced, so the image quality control parameter of the divided block encoded immediately before is set as the predicted image quality control parameter. Thereafter, the predicted image quality control parameter selection process is terminated.

  As shown in the present embodiment, it is possible to finely control the image quality by selecting the predicted image quality control parameter using a plurality of attributes.

  In the present embodiment, the order of attribute determination and the processing method after determination are not limited to this. In the present embodiment, the selection based on the prediction mode is performed when the process division block is other than the intra prediction. However, as shown in FIG. 27, a configuration in which S26003 to S26005 are omitted may be employed. Moreover, although the intra prediction method was divided roughly into three by S26002, it is not limited to this, You may set an intra prediction method more finely. Further, a method may be used in which the prediction mode of the divided block to be referred to after S26002 is selected from those that further match. That is, if horizontal prediction is performed in S26002, if the left divided block is the intra prediction mode, the process proceeds to S26009, but if not, the condition is selected to be a logical sum so that the process proceeds to S26003. It doesn't matter.

  Further, conditions may be further added for the divided blocks that can be referred to. For example, the prediction mode of the divided block that can be referred to in S26008 may be referred to, or the presence / absence of a significant coefficient in the divided block and the presence / absence of skip may be added as conditions.

  In this embodiment, when there is no divided block to be referred to in S26012, the image quality control parameter of the divided block encoded immediately before is set as the predicted image quality control parameter. However, the present invention is not limited to this, and S28001 in FIG. The image quality control parameter given to the slice may be used as shown in S29001 of FIG. This value is H.264. If it is H.264, it can be obtained from pic_init_qp_minus 26 set by PPS and slice_qp_delta determined in units of slices. Further, it may be a predetermined value determined in advance or an image quality control parameter given to a picture. Of course, image quality control parameters given in units of LCUs may be used when dividing an LCU as in HEVC.

  In the present embodiment, the upper and left blocks are referred to as the information on the surrounding divided blocks. However, the present invention is not limited to this, and the divided blocks on the upper left and the upper right may be referred to. Of course, the predicted image quality control parameter may be obtained by referring to the divided blocks at the same position in the frame.

<Example 16>
An image decoding method for decoding a stream encoded by the image encoding method described in the fifteenth embodiment will be described with reference to FIG. Unless otherwise specified, the configuration and operation of each step are the same as those in the second and fourth embodiments.

  In S7021, a block having the same attribute information of the target divided block is selected from the peripheral blocks. In this embodiment, the selection is performed by a combination of a plurality of attributes, that is, a prediction mode, an intra prediction method, and presence / absence of a divided block to be referred to. The plurality of attributes are not limited to this, and the above-described motion vector and block size may be used, or other attributes may be used. For details of S7021, according to the flowcharts shown in FIG. 26, FIG. 27, FIG. 28, and FIG. 29, the prediction mode, intra prediction method, and presence / absence of the reference divided block are determined as attributes. Accordingly, the predicted image quality control parameter can be correctly selected by performing the same operation as that on the encoding side.

  In the present embodiment, the upper and left blocks are referred to as the information on the surrounding divided blocks. However, the present invention is not limited to this, and the divided blocks on the upper left and the upper right may be referred to. Of course, the predicted image quality control parameter may be obtained by referring to the divided blocks at the same position in the frame.

<Other examples>
In the present invention, the encoded data has been described as an example of being recorded in the HDD 903. However, the present invention is not limited to this, and it may be recorded on a recording medium other than the hard disk device. In addition, the communication interface 905 may be transmitted to the communication circuit, or may be connected to an external recording device and recorded on a portable medium.

  Furthermore, although the present invention has been described with reference to an example in which software is recorded in a hard disk device, the present invention is not limited to this, and may be recorded in hardware such as a ROM, or a portable medium such as a memory card or a disk. Of course, it does not matter if it is recorded in

Claims (4)

In an image encoding method for dividing an input image into a plurality of blocks having different sizes and encoding the image in units of the divided blocks,
An acquisition step of acquiring attribute information of the processing target block;
A setting step for setting a control parameter for controlling the image quality of the processing target block;
A determination step of determining a predictive control parameter based on the attribute information;
A calculation step of calculating a difference value between the control parameter calculated in the setting step and the predicted control parameter;
An encoding step for encoding the difference value calculated in the calculation step;
An image encoding method comprising: An image decoding method for decoding input encoded data,
A decoding step of decoding the attribute information encoded from the input encoded data and the difference value of the encoded control parameter;
A determination step of determining a predictive control parameter based on the attribute information;
A calculation step of calculating a control parameter of the processing target block from the prediction control parameter and the difference value generated in the decoding step;
An image decoding method characterized by comprising: In an image encoding device that divides an input image into a plurality of blocks having different sizes and encodes the image in units of the divided blocks,
An acquisition means for acquiring attribute information of the processing target block;
Setting means for setting a control parameter for controlling the image quality of the processing target block;
Determining means for determining a predictive control parameter based on the attribute information;
Calculating means for calculating a difference value between the control parameter calculated by the setting means and the predicted control parameter;
Encoding means for encoding the difference value calculated by the calculating means;
An image encoding device comprising: An image decoding device for decoding input encoded data,
Decoding means for decoding attribute information encoded from input encoded data and a difference value between encoded control parameters;
Determining means for determining a predictive control parameter based on the attribute information;
Calculating means for calculating a control parameter of the processing target block from the prediction control parameter and the difference value generated by the decoding means;
An image decoding apparatus comprising: