JP2011259230A - Moving image decoder, moving image decoding method and moving image decoding program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique for improving, while suppressing increase of an operation amount for a direct mode, efficiency of motion compensation prediction by enhancing quality of a prediction image.SOLUTION: A moving image decoder includes a standard reference image determination part 214 for selecting, when information indicating not to transmit designation information of a reference image in encoding and designation information of a motion vector in encoding is detected, a first reference image and a first motion vector value used for motion compensation processing of a decoding object block based on information used for motion compensation prediction in a block in the periphery of the decoding object block. The moving image decoder includes a reference image synthesis part for synthesizing a first reference block extracted from the first reference image and a predetermined region of at least another reference image using the first motion vector value to create a synthesized reference block. The moving image decoder includes an entropy decoding part 202 creating a decoded image by adding a prediction block to a prediction difference block decoded from the decoding object block, where the synthesized reference block is used as the prediction block.

Description

  The present invention relates to a moving image signal decoding technique.

  In recent years, services that deliver digital image and sound content via broadcast waves such as satellite and terrestrial waves and networks have been put into practical use, and content with a huge amount of information can be efficiently recorded and transmitted. In order to do so, a high-efficiency encoding technique is required. As high-efficiency encoding of moving images, the correlation between pixels that are spatially adjacent in the same frame of a moving image signal, and the correlation between temporally adjacent frames and fields, represented by MPEG4-AVC, are used. A method of compressing information using it is used.

  In MPEG4-AVC, as a compression using temporal correlation, a local decoded image of a frame that has already been encoded is used as a reference image for a target image that is an encoding target frame, and a two-dimensional block of a predetermined size is used. A motion amount (hereinafter referred to as “motion vector”) between the target image and the reference image is detected in units (hereinafter referred to as “target block”), and a predicted image based on the target block and the motion vector is detected. The generated motion compensated prediction is used.

  In MPEG4-AVC, the size of a target block in a 16 × 16 pixel two-dimensional block (hereinafter referred to as “macroblock”), which is a unit of encoding processing, is changed, and a motion vector for each target block is obtained. By using a method of predicting using, a method of storing a plurality of reference images and selecting a reference image used for prediction, and a method of generating a motion predicted image by obtaining a motion vector between two reference images and a target block, It is possible to improve the prediction accuracy of motion compensated prediction, thereby reducing the amount of information.

  Also, in motion compensated prediction, it is necessary to encode and transmit the generated motion vector, and in order to prevent an increase in the amount of information due to the motion vector, a predicted motion vector predicted from the motion vector for a decoded block around the target block By encoding using values, it is possible to use motion compensated prediction called direct mode in which no motion vector is transmitted.

  However, since the prediction of the motion vector cannot always be obtained with high accuracy, as shown in Patent Document 1, both the encoding side and the decoding side detect a motion vector between reference images, and the motion vector is A method of generating a predicted motion vector of a target block and assuming a direct mode on the assumption that it is continuous in time is also presented.

JP 2008-154015 A

  In motion compensated prediction in conventional moving picture coding represented by MPEG4-AVC, the following problems cannot be solved, and hence improvement in coding efficiency is hindered.

  The first problem is the degradation of the quality of the motion compensated predicted image due to the degradation of the quality of the decoded image used as the reference image, especially the degradation mixed in the motion compensated predicted image when high-compression encoding is performed. While the component deteriorates the prediction accuracy, it is necessary to encode information for restoring the deteriorated component as a prediction difference, and the amount of information is increasing.

  The second problem is that the motion vector prediction is not accurate enough for image signals with little temporal and spatial motion continuity, and the predicted image quality when using the direct mode is effective. It is a point not to do. This degradation occurs when adjacent blocks have different motions across the target object, and the motion vector used for prediction when the motion is large in time has moved corresponding to the motion of the original target block This degradation occurs because a block of positions is assumed. Similarly, when the motion changes with time, the prediction is not successful and deterioration occurs.

  In MPEG4-AVC, a reference image used for motion compensation prediction can be switched in units of a target block from a plurality of reference images, and intra prediction that performs prediction within a screen without using motion compensation prediction is selected. Is also possible. For this reason, there are many cases where there is no adjacent motion vector for the corresponding reference image used for motion vector prediction in the adjacent block, and the motion vector prediction accuracy in the direct mode further decreases.

  The third problem is an increase in the amount of code required for motion vector transmission when using prediction using two reference images or motion compensated prediction in units of fine blocks. When two reference images are used, prediction deterioration is smoothed by adding the reference images, and the influence of the deterioration component can be reduced. To increase. Also, in motion compensation in fine block units, it is possible to obtain appropriate motion according to the boundary of the object, and the accuracy of the predicted image is improved, but it is necessary to transmit motion vectors in fine units. The amount of code increases.

  Patent Document 1 is a technique presented to solve the second problem described above. When a spatially uniform motion is present, a motion vector obtained between reference images is a target block. The motion vector prediction accuracy is improved because the motion passes through the position, but if the motion is not spatially uniform, it is the predicted motion vector obtained without using the target block information Therefore, the motion is different from that of the target block, and the prediction is not sufficient. In addition, in order to capture a large motion, a motion vector detection process over a wide range between reference images is required for both the encoding device and the decoding device, which causes a problem that the amount of calculation increases.

  The present invention has been made in view of such a situation, and an object of the present invention is to compensate for motion compensation by improving the quality of a predicted image while suppressing an increase in the amount of calculation in an encoding device and a decoding device in a mode in which no motion vector is transmitted. It is to provide a technique for improving the efficiency of prediction.

  In order to solve the above-described problem, a moving picture decoding apparatus according to an aspect of the present invention is a mode information indicating that a reference image designation information at the time of encoding and a motion vector designation information at the time of encoding are not transmitted. A first reference image and a first motion vector value used for motion compensation processing of the decoding target block when the mode information is detected by the mode detection unit (203) A reference reference image determination unit (214) that selects based on information used for motion compensation prediction in the peripheral blocks of the decoding target block, and a first reference image extracted from the first reference image using the first motion vector value. A reference image synthesis unit (215 and 216) that generates a synthesized reference block obtained by synthesizing one reference block and a predetermined region of at least one other reference image; and the synthesized reference block The provided as the prediction block, and the prediction block by adding the prediction difference block decoded from the decoding target block, decoding section that generates a decoded image and a (202).

  The reference reference image determination unit (214) decodes a motion vector value for a reference image that is not used for motion compensated prediction in a peripheral block of the encoding target block, and a motion vector value for a reference image used for motion compensated prediction. It is calculated using the image signal, and it is determined whether or not to add a reference image that is not used for the motion compensation prediction to the first reference image selection candidate according to the calculated motion vector value. In this case, by adding the calculated motion vector value to a motion vector candidate for a reference image that is not used for the motion compensation prediction, the frequency of the reference image that is not used for the motion compensation prediction is increased. To select the first reference image and to use the motion vector value added to the candidate in the selected first reference image Generating a first motion vector values (Fig. 12, S1202~S1225).

  Based on a plurality of second motion vector values between a plurality of reference images detected by the reference image synthesizing unit (215 and 216) in block units having a size equal to or smaller than the first reference block. A reference image to be used for the second reference block is selected, and a synthesized reference block obtained by synthesizing the first reference block and the second reference block is generated, and the synthesized reference block is set as a prediction block, or the first It may be determined whether to use the reference block as a prediction block (FIG. 13, S1303 to S1308).

  Another aspect of the present invention is a video decoding method. In this method, a step of detecting mode information indicating a mode that does not transmit the designation information of the reference image at the time of encoding and the designation information of the motion vector at the time of encoding, and when the mode information is detected, Selecting a first reference image and a first motion vector value used for motion compensation processing of a decoding target block based on information used for motion compensation prediction in a peripheral block of the decoding target block; Generating a synthesized reference block by synthesizing a first reference block extracted from a first reference image using one motion vector value and a predetermined region of at least one other reference image; and the synthesized reference block As a prediction block, a decoded image is generated by adding the prediction block and a prediction difference block decoded from the decoding target block That and a step.

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

  According to the present invention, it is possible to improve the quality of motion compensated prediction by reducing the number of motion vectors to be encoded and suppressing the increase in the amount of calculation in the encoding device and the decoding device, while improving the quality of the predicted image.

It is a block diagram which shows the structure of the moving image encoder of Embodiment 1 of this invention. It is a block diagram which shows the structure of the moving image decoding apparatus of Embodiment 1 of this invention. It is a conceptual diagram which shows the operation | movement of the direct motion compensation prediction of Embodiment 1 of this invention. It is a flowchart explaining the operation | movement of the direct mode motion compensation prediction by the side of an encoding of Embodiment 1 of this invention. It is a flowchart explaining the operation | movement of the direct mode motion compensation prediction by the decoding side of Embodiment 1 of this invention. It is a flowchart explaining operation | movement of the reference | standard reference image determination part of Embodiment 1 of this invention. It is a figure which shows the management form of the motion compensation prediction information of the surrounding block in Embodiment 1 of this invention. It is a block diagram which shows the structure of the motion vector detection part between reference images of Embodiment 1 of this invention. It is a flowchart explaining operation | movement of the motion vector detection part between reference images of Embodiment 1 of this invention. FIG. 5 is a diagram showing an example of a motion vector detection range between reference images in Embodiment 1 of the present invention. It is a block diagram which shows the structure of the moving image encoder of Embodiment 2 of this invention. It is a block diagram which shows the structure of the moving image decoding apparatus of Embodiment 2 of this invention. It is a flowchart explaining operation | movement of the reference | standard reference image determination part of Embodiment 2 of this invention. It is a flowchart explaining operation | movement of the motion vector detection part between reference images of Embodiment 2 of this invention. It is a conceptual diagram which shows the interpolation object of the block in which the motion compensation information with respect to the object reference image in Embodiment 2 of this invention does not exist. In Embodiment 2 of this invention, it is a figure which shows an example of the selection result of the bidirectional | one-way in the small block unit of the output direct motion compensation prediction.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(Embodiment 1)
First, the moving picture coding apparatus according to the first embodiment will be described. FIG. 1 is a block diagram showing a configuration of the moving picture encoding apparatus according to the first embodiment.

  As shown in FIG. 1, the moving picture coding apparatus according to Embodiment 1 includes an input terminal 100, an input picture buffer 101, a block division unit 102, an intra-frame prediction unit 103, a motion vector detection unit 104, and a motion compensation prediction unit 105. , Motion vector prediction unit 106, standard reference image determination unit 107, inter-reference image motion vector detection unit 108, direct motion compensation prediction unit 109, prediction mode determination unit 110, subtractor 111, orthogonal transform unit 112, quantization unit 113, An inverse quantization unit 114, an inverse orthogonal transform unit 115, an adder 116, an intra-frame decoded image memory 117, a decoded reference image memory 118, an entropy encoding unit 119, a stream buffer 120, an output terminal 121, and a code amount control unit 122 are included. . Here, the inter-reference image motion vector detection unit 108 and the direct motion compensation prediction unit 109 are examples of the “reference image synthesis unit” of the video encoding device of the present invention.

  The point that the reference reference image determination unit 107 and the inter-reference image motion vector detection unit 108 are provided, and the operation in the processing block and the direct motion compensation prediction unit 109 are the features in Embodiment 1 of the present invention. As for the processing block, the same processing as the processing block constituting the encoding processing in the moving image encoding apparatus such as MPEG4-AVC can be applied.

  A digital image signal input from the input terminal 100 is stored in the input image buffer 101. The digital image signal stored in the input image buffer 101 is supplied to the block dividing unit 102, and is cut out as an encoding target block in units of macroblocks composed of 16 × 16 pixels. The block dividing unit 102 supplies the extracted encoding target block to the intra-frame prediction unit 103, the motion vector detection unit 104, the motion compensation prediction unit 105, the direct motion compensation prediction unit 109, and the subtractor 111.

  In the intra-frame prediction unit 103, the decoding target image input from the block division unit 102 and the decoded image of the area that has been encoded with respect to the periphery of the encoding target block, stored in the intra-frame decoding image memory 117. Is input, and prediction using correlation within the frame is performed. For example, the pixel value is predicted in a plurality of predetermined directions in units of 4 × 4 pixels, 8 × 8 pixels, and 16 × 16 pixels for the encoding target block, and is selected as a unit of prediction processing Prediction using the correlation of adjacent pixels in the screen is performed using a method called intra prediction that generates a predicted image together with information indicating the direction (intra prediction mode). The predicted image and the selected intra prediction mode are output from the intra-frame prediction unit 103 to the prediction mode determination unit 110.

  In the motion vector detection unit 104, the block to be encoded input from the block division unit 102 and the decoded image of the frame that has been encoded on the entire screen and stored in the decoded reference image memory 118 are input as the reference image. Then, motion estimation is performed between the encoding target block and the reference image. As a general motion estimation process, a reference image at a position moved by a predetermined movement amount from the same position on the screen is cut out, and the movement amount that minimizes the prediction error when the image is used as a prediction block is determined as a motion vector. As a value, a block matching process that is obtained while changing the movement amount is used. The detected motion vector value is output to the motion compensation prediction unit 105.

  The motion compensation prediction unit 105 receives the motion vector value obtained by the motion vector detection unit 104, and receives motion compensated prediction images for a plurality of block sizes of 16 × 16 or less and a plurality of reference images from the decoded reference image memory 118. The prediction signal with the least difference information to be encoded is selected and the prediction signal with the least difference information to be encoded is selected for the encoding target block acquired and input from the block dividing unit 102. The motion compensation prediction unit 105 outputs the selected motion compensation prediction mode and the prediction signal to the prediction mode determination unit 110.

  The motion vector prediction unit 106 calculates a predicted motion vector value using the motion vectors of the surrounding encoded blocks, and supplies the motion vector detection unit 104 and the motion compensated prediction unit 105.

  By using the predicted motion vector value, the motion vector detection unit 104 takes into account the amount of code necessary for encoding the difference between the motion vector predicted value and the motion vector value, and determines the optimal motion vector value. To detect. Similarly, the motion compensation prediction unit 105 takes into account the amount of code required when encoding the difference between the motion vector prediction value and the motion vector value, and the reference image used as the block unit for the optimal motion compensation prediction and Select a motion vector value.

  The motion vector prediction unit 106 further stores the stored motion vectors of the surrounding encoded blocks, the reference image that is the target of the motion vector, and the information indicating the block size of the motion compensation, the standard reference image determination unit 107. Output to.

  The reference reference image determination unit 107 is a motion vector that is motion compensated prediction information of a surrounding encoded block input from the motion vector prediction unit 106, and a reference image or motion compensation block that is a target of the motion vector. Based on the size, it has a function of determining a reference image that can predict a motion vector for an encoding target block and calculating a predicted motion vector value for the determined reference image. Information indicating the reference image corresponding to the calculated predicted motion vector value is output to the inter-reference image motion vector detection unit 108. The detailed operation of the reference reference image determination unit 107 will be described later.

  The inter-reference image motion vector detection unit 108 uses the decoded reference image memory 118 for the motion compensated prediction based on the information indicating the reference image corresponding to the predicted motion vector value input from the standard reference image determination unit 107. Extract reference blocks.

  Subsequently, the inter-reference image motion vector detection unit 108 calculates an error value by block matching or the like for the motion vector between the extracted first reference block and another reference image, and refers to a motion vector having a small value. Calculated as an inter-image motion vector. The inter-reference image motion vector detection unit 108 outputs the second reference block extracted from the other reference image based on the calculated inter-reference image motion vector to the direct motion compensation prediction unit 109 together with the first reference block. The inter-reference image motion vector detection unit 108 calculates a motion vector between another reference image and the encoding target block, which is calculated from the predicted motion vector value for the reference reference image, the predicted motion vector value, and the inter-reference image motion vector. The value is also output to the direct motion compensation prediction unit 109. The detailed operation of the inter-reference image motion vector detection unit 108 will be described later.

  The direct motion compensation prediction unit 109 performs motion averaging for the direct mode in which no motion vector is transmitted by averaging the first reference block and the second reference block input from the inter-reference image motion vector detection unit 108 for each pixel. A compensated prediction block is generated and output to the prediction mode determination unit 110. Similarly, a prediction motion vector value for the reference reference image and a motion vector value between another reference image and the encoding target block are also predicted by the prediction mode determination unit. To 110.

  The prediction mode determination unit 110 uses the prediction mode and the prediction image for each prediction method input from the intra-frame prediction unit 103, the motion compensation prediction unit 105, and the direct motion compensation prediction unit 109, and the encoding input from the block division unit 102. For the target block, a prediction signal with the least amount of difference information to be encoded is selected, a prediction image block for the selected prediction method is output to the subtractor 111 and the adder 116, and the entropy encoding unit 119 is Thus, prediction mode information as additional information and information that requires encoding according to the prediction mode are output.

  The prediction mode determination unit 110 outputs, to the motion vector prediction unit 106, the motion vector used in the selected prediction method, the reference image that is the target of the motion vector, and the motion compensation block size and the like. The output information is used to generate a predicted motion vector value in the subsequent encoding target block.

  The subtractor 111 calculates the difference between the encoding target block supplied from the block dividing unit 102 and the predicted image block supplied from the prediction mode determining unit 110, and supplies the result to the orthogonal transform unit 112 as a difference block. .

  The orthogonal transform unit 112 generates DCT coefficients corresponding to the orthogonally transformed frequency component signal by performing DCT transform on the difference block in units of 4 × 4 pixels or 8 × 8 pixels. Further, the orthogonal transform unit 112 collects the generated DCT coefficients in units of macroblocks and outputs them to the quantization unit 113.

  In the quantization unit 113, the quantization process is performed by dividing the DCT coefficient by a different value for each frequency component. The quantization unit 113 supplies the quantized DCT coefficient to the inverse quantization unit 114 and the entropy encoding unit 119.

  The inverse quantization unit 114 performs inverse quantization by multiplying the quantized DCT coefficient input from the quantization unit 113 by a value divided at the time of quantization, and the result of the inverse quantization is obtained. The decoded DCT coefficient is output to the inverse orthogonal transform unit 115.

  The inverse orthogonal transform unit 115 performs inverse DCT processing to generate a decoded difference block. The inverse orthogonal transform unit 115 supplies the decoded difference block to the adder 116.

  The adder 116 adds the prediction image block supplied from the prediction mode determination unit 110 and the decoded difference block supplied from the inverse orthogonal transform unit 115 to generate a local decoding block. The local decoded block generated by the adder 116 is stored in the intra-frame decoded image memory 117 and the decoded reference image memory 118 in a form subjected to inverse block conversion. In the case of MPEG-4 AVC, adaptive filtering is applied to a block boundary where coding distortion of each block is likely to appear as a boundary before the local decoded block is input to the decoded reference image memory 118. In some cases, processing to be performed is performed.

  The entropy encoding unit 119 converts the quantized DCT coefficient supplied from the quantization unit 113, the prediction mode information supplied from the prediction mode determination unit 110, and information that requires encoding according to the prediction mode. On the other hand, variable length encoding of each information is performed. Specifically, intra prediction mode and prediction block size information in the case of intra-frame prediction, and prediction block size, reference image designation information, and motion vector in the case of motion compensation prediction and synthesized image motion compensation prediction, The difference value from the predicted motion vector value is information that requires encoding. Information subjected to variable length coding is output to the stream buffer 120 from the entropy coding unit 119 as a coded bit stream.

  The encoded bit stream stored in the stream buffer 120 is output to a recording medium or a transmission path via the output terminal 121. Regarding the code amount control of the encoded bit stream, the code amount control unit 122 is supplied with the code amount of the encoded bit stream stored in the stream buffer 120 and compared with the target code amount. In order to approach the target code amount, the quantization level (quantization scale) of the quantization unit 113 is controlled.

  Subsequently, a moving picture decoding apparatus that decodes an encoded bitstream generated by the moving picture encoding apparatus according to Embodiment 1 will be described. FIG. 2 is a configuration diagram of the moving picture decoding apparatus according to the first embodiment.

  As shown in FIG. 2, the moving picture decoding apparatus according to Embodiment 1 includes an input terminal 200, a stream buffer 201, an entropy decoding unit 202, a prediction mode decoding unit 203, a prediction image selection unit 204, an inverse quantization unit 205, and an inverse. Orthogonal transformation unit 206, adder 207, intra-frame decoded image memory 208, decoded reference image memory 209, output terminal 210, intra-frame prediction unit 211, motion vector prediction decoding unit 212, motion compensation prediction unit 213, reference reference image determination unit 214, an inter-reference image motion vector detection unit 215, and a direct motion compensation prediction unit 216. Here, the inter-reference image motion vector detection unit 215 and the direct motion compensation prediction unit 216 are examples of the “reference image synthesis unit” of the video decoding device of the present invention.

  The point in which the reference reference image determination unit 214 and the inter-reference image motion vector detection unit 215 are provided, and the operations in these processing blocks and the direct motion compensation prediction unit 216 are the features in Embodiment 1 of the present invention. The operation is the same as that of the same functional block of the moving picture coding apparatus shown in FIG. 1, thereby generating a motion compensated prediction block for the direct mode without transmission of additional information. For the other processing blocks, the same processing as the processing blocks constituting the decoding processing in the moving picture decoding apparatus such as MPEG4-AVC can be applied.

  The encoded bit stream input from the input terminal 200 is supplied to the stream buffer 201, and the stream buffer 201 absorbs the code amount variation of the encoded bit stream and is supplied to the entropy decoding unit 202 in a predetermined unit such as a frame. The The entropy decoding unit 202 performs variable-length decoding on the encoded prediction mode information, the additional information corresponding to the prediction mode, and the quantized DCT coefficient from the encoded bitstream input via the stream buffer 201. Then, the quantized DCT coefficient is output to the inverse quantization unit 205, and the prediction mode information and additional information corresponding to the prediction mode are output to the prediction mode decoding unit 203.

  Regarding the inverse quantization unit 205, the inverse orthogonal transform unit 206, the adder 207, the intra-frame decoded image memory 208, and the decoded reference image memory 209, the local decoding process of the moving image coding apparatus according to the first embodiment of the present invention. Processing similar to that of a certain inverse quantization unit 114, inverse orthogonal transform unit 115, adder 116, intra-frame decoded image memory 117, and decoded reference image memory 118 is performed. The decoded image stored in the decoded reference image memory 209 is displayed as a decoded image signal on the display device via the output terminal 210.

  In the prediction mode decoding part 203, when motion compensation prediction is selected as a prediction mode from the prediction mode information input from the entropy decoding part 202 and the additional information corresponding to the prediction mode, the motion vector prediction decoding part 212 The prediction mode information indicating the predicted block unit or the prediction mode information indicating the direct mode and the decoded difference vector value in the case of the motion compensation prediction mode are output and the prediction image selection unit 204 is also predicted. Output mode information. Also, the prediction mode decoding unit 203 provides information indicating that the prediction mode decoding unit 203 has selected to the intra-frame prediction unit 211, the motion compensation prediction unit 213, and the direct motion compensation prediction unit 216 according to the decoded prediction mode information. Outputs additional information according to the mode.

  The prediction image selection unit 204 is output from any of the intra-frame prediction unit 211, the motion compensation prediction unit 213, and the direct motion compensation prediction unit 216 according to the prediction mode information input from the prediction mode decoding unit 203. A predicted image for the decoding target block is selected and output to the adder 207.

  When the decoded prediction mode indicates intra-frame prediction, the intra-frame prediction unit 211 receives the intra prediction mode as additional information according to the prediction mode from the prediction mode decoding unit 203, and according to the intra prediction mode. The decoded image of the region where decoding is completed is input to the periphery of the decoding target block stored in the intra-frame decoded image memory 208, and prediction using the intra-frame correlation is performed in the same intra prediction mode as the encoding device. Done. The intra-frame prediction unit 211 outputs the intra-frame prediction image generated by the prediction to the prediction image selection unit 204.

  The motion vector predictive decoding unit 212 uses the motion vector of the neighboring decoded block for the decoded difference vector value input from the prediction mode decoding unit 203, and performs the motion prediction using the same method as that performed by the encoding device. A vector value is calculated, and a value obtained by adding the difference vector value and the predicted motion vector value is output to the motion compensated prediction unit 213 as a motion vector value of the decoding target block. The motion vector is decoded by the number encoded according to the block unit of the prediction process indicated in the motion compensation prediction mode or the composite image motion compensation prediction mode.

  The motion vector predictive decoding unit 212 further stores the stored motion vectors of the surrounding encoded blocks, the reference image that is the target of the motion vector, and information indicating the block size of motion compensation, as a reference reference image determination unit. To 214.

  The motion compensation prediction unit 213 generates a motion compensated prediction image from the decoded reference image memory 209 from the motion vector value input from the motion vector prediction decoding unit 212, and sends the generated motion compensation prediction image to the prediction image selection unit 204. Output.

  The reference reference image determination unit 214 is the motion compensation prediction information of the surrounding encoded blocks input from the motion vector prediction decoding unit 212, the motion vector, the reference image that is the target of the motion vector, and motion compensation. Based on the block size, it has a function of determining a reference image that can predict a motion vector for a decoding target block and calculating a predicted motion vector value for the determined reference image. Information indicating the reference image corresponding to the calculated predicted motion vector value is output to the inter-reference image motion vector detection unit 215.

  The inter-reference image motion vector detection unit 215 uses the decoded reference image memory 209 for the first motion compensation prediction based on information indicating the reference image corresponding to the predicted motion vector value input from the standard reference image determination unit 214. Extract reference blocks.

  Subsequently, the inter-reference image motion vector detection unit 215 calculates an error value by block matching or the like for the motion vector between the extracted first reference block and another reference image, and refers to a motion vector having a small value. Calculated as an inter-image motion vector. The inter-reference image motion vector detection unit 215 outputs the second reference block extracted from the other reference images based on the calculated inter-reference image motion vector together with the first reference block to the direct motion compensation prediction unit 216. The inter-reference image motion vector detecting unit 215 calculates a motion vector between another reference image and the encoding target block, which is calculated from the predicted motion vector value for the reference reference image, the predicted motion vector value, and the inter-reference image motion vector. The value is also output to the direct motion compensation prediction unit 216. Detailed operations of the reference reference image determination unit 214 and the inter-reference image motion vector detection unit 215 are paired with the reference reference image determination unit 107 and the inter-reference image motion vector detection unit 108 in the moving image encoding apparatus of the first embodiment. The detailed operation will be described later.

  The direct motion compensation prediction unit 216 averages the first reference block and the second reference block input from the inter-reference image motion vector detection unit 215 for each pixel, thereby performing motion for the direct mode in which no motion vector is transmitted. A compensated prediction block is generated and output to the prediction image selection unit 204, and a motion vector value between a prediction motion vector value for the reference reference image and another reference image and the current block is encoded to the motion vector prediction decoding unit 212. Output. The output information is used to generate a predicted motion vector value in the subsequent decoding target block.

  Hereinafter, a motion compensation prediction prediction image generation method using a synthesized reference image, which operates in the video encoding device and the video decoding device according to Embodiment 1, will be described with reference to FIG.

  FIG. 3C is a conceptual diagram showing direct motion compensation prediction. FIGS. 3A and 3B are conceptual diagrams of motion compensation prediction using a plurality of reference images used in MPEG4-AVC.

  FIG. 3 (a) detects a motion vector between two reference pictures called bi-directional prediction and a coding target block, transmits a motion vector for each reference picture, In this method, the average value of reference blocks indicated by motion vectors is used as a predicted image. A prediction image having a function of removing an encoding degradation component as a motion adaptive filter in the time direction by synthesizing two reference images and a function of following a minute luminance change component of an encoding object by averaging. Can be generated.

  In MPEG4-AVC, a plurality of decoded reference images are stored, and a reference image number and a motion vector used for prediction are transmitted in a predetermined block unit to adaptively select the reference image. Yes. In the case of FIG. 3A, four decoded images are secured as reference images, and two prediction images are acquired using the reference image 1 and the reference image 3, and bidirectional prediction is performed. ing. As the motion vectors, mvL0 and mvL1 take the difference value from the predicted motion vector and are transmitted.

  FIG. 3B is a technique called “temporal direct mode” that performs prediction using two reference images without transmission of motion vectors. In the temporal direct mode, an anchor block, which is a temporally closest reference image (referred to as reference image 3 here) reproduced after the encoding target image, is located at the same position as the encoding target block. When a block is generated by motion compensated prediction from the reference image 2, it is encoded from a motion vector mvCol between the reference image 2 and the reference image 3 on the assumption that the motion is temporally continuous. A motion vector value mvL0 between the target block and the reference image 2 and a motion vector value mvL1 between the encoding target block and the reference image 3 are generated, and bidirectional prediction is performed using the motion vector.

  In the temporal direct mode, a prediction image obtained by synthesizing two reference images can be generated without transmitting the motion vectors mvL0 and mvL1, but when mvCol is large as shown in FIG. 3b), it is expressed by mvCol. Therefore, when there is no spatial motion continuity, appropriate mvL0 and mvL1 cannot be calculated.

  Further, the motion vectors mvL0 and mvL1 are implicitly generated only when the motion is temporally continuous, and an appropriate motion vector can be selected even when the motion vector value is less continuous in time. It cannot be generated, and the temporal direct mode does not function effectively unless it is limited in terms of spatial and temporal movement.

  In the method disclosed in Patent Document 1, for the purpose of improving the quality of the above-described temporal direct mode, both the encoding side and the decoding side detect motion in a block existing at a symmetrical position around the encoding target block between reference images. This is a method of generating a motion vector having temporal continuity passing through the position of the encoding target block by recreating mvCol, but the temporal continuity is limited in that the motion is uniform as in the temporal direct mode. In the case where mvL0 and mvL1 are generated and the temporal continuity is small, it does not function effectively like the temporal direct mode.

  As shown in FIG. 3C, the prediction configuration of the direct motion compensated prediction according to the present invention determines a reference reference image (reference image 3 in FIG. 3C) from a plurality of existing reference images, and sets the reference reference image. The predicted motion vector value between the current block and the current block is calculated as mvL0. Subsequently, with respect to the reference block indicated by the motion vector mvL0 of the standard reference image, motion vector detection between reference images for another reference image (reference image 1 in FIG. 3C) is performed on the encoding side and decoding. The motion vector value mvL1 between the encoding target block and the reference image 1 is calculated from the inter-reference image motion vector mvInterRef, and bi-directional prediction using two reference images using the motion vectors mvL0 and mvL1 Take the configuration to do.

  As an implicit calculation of motion vectors for images that are not spatially and temporally continuous, a reference image that can predict a motion vector most appropriately is selected as a reference image using peripheral decoded motion compensation prediction information. By detecting the motion vector between the reference images for the same part of the other reference images for both the encoding and decoding of the prediction image block indicated by the reference image and the motion vector as a reference point with the function The point of tracking and implicitly calculating the other motion vector value is the point of the direct motion compensated prediction method in the moving picture coding apparatus and the moving picture decoding apparatus according to the embodiment of the present invention. Even in video signals with little temporal continuity, more appropriate bi-directional prediction can be realized without the transmission of motion vectors, and encoding efficiency can be greatly improved.

  4A and 4B are flowcharts showing the direct motion compensated prediction processing procedures in the encoding device and decoding device of Embodiment 1, and the flow of prediction processing will be described. 4A is the encoding side, and FIG. 4B is the decoding side.

On the encoding side, first, a reference reference image for an encoding target block is determined (S400).
Subsequently, a predicted motion vector value for the standard reference image is calculated (S401). Subsequently, the motion vector value mvL0 is used as the predicted motion vector value, and a reference reference image at a position moved by mvL0 from the in-screen position of the encoding target block is acquired to generate a first predicted block (S402).

  Subsequently, a motion vector is detected between the first prediction block and another reference image, and an inter-reference image motion vector value mvInterRef is calculated (S403). Regarding the selection of other reference images, the method of explicitly specifying in a predetermined unit or the reference image closest to the encoding target block that is in the temporal position across the encoding target block is acquired. It is possible to use the method to do. In the first embodiment, it is assumed that information indicating the other reference image with respect to the standard reference image (hereinafter referred to as a reference ID) is transmitted in units similar to the slice header in MPEG4-AVC as information in units of screens.

From the inter-reference image motion vector value mvInterRef, another reference image moved by mvInterRef from the position in the screen of the first prediction block is acquired, and a second prediction block is generated (S404). The motion vector value mvL1 indicating the motion between the encoding target block and the second prediction block is an implicitly generated value and does not need to be calculated for motion compensated prediction. Created for use as information for generating a block prediction vector. Specifically, mvL1 is calculated by adding mvL0 and mvInterRef (S405).
mvL1 = mvL0 + mvInterRef
The motion vector value has two components, horizontal and vertical, and each value is calculated by the above formula.

  Finally, the first prediction block and the second prediction block are averaged for each pixel, and the prediction block subjected to bi-directional prediction is output as a direct motion compensation prediction block together with the motion vector values mvL0 and mvL1 (S406). ).

  On the decoding side, when the decoding target block is in the direct motion compensation prediction mode (S410: YES), first, a reference reference image for the decoding target block is determined (S411).

  Subsequently, processing is performed in the same manner as on the encoding side, a predicted motion vector value for the reference reference image is calculated (S412), and a first predicted block is generated using the predicted motion vector value as the motion vector value mvL0 (S413). ), For the first prediction block, a motion vector is detected with another reference image, a motion vector value mvInterRef between reference images is calculated (S414), and the motion vector value mvInterRef between reference images is 2 prediction blocks are generated (S415), and mvL1 is calculated by adding mvL0 and mvInterRef (S416).

  Finally, the first prediction block and the second prediction block are added and averaged for each pixel, and the prediction block subjected to bidirectional prediction is output as a direct motion compensated prediction block together with the motion vector values mvL0 and mvL1 (S417). ).

  Since the decoded image signal is used as the reference image for these operations, the same operation can be performed on the encoding side and the decoding side, and the decoding is performed without transmitting the motion vector values mvL0, mvInterRef, and mvL1. Can be reproduced on the side.

  Next, the operation of the standard reference image determination process that operates according to the standard reference image determination units 107 and 214 and S400 and S411 of the process flowchart of FIG. 4 will be described with reference to the flowchart of FIG. Since the same operation is performed on the encoding side and the decoding side, the following description is realized as a common function.

  In the reference reference image determination units 107 and 214, first, motion compensation prediction motion vector values, reference IDs, and motion compensation block size information used in the peripheral blocks of the encoding target block (decoding target block in the case of a decoding device). Is acquired (S500).

  Next, the motion compensation prediction information adjacent to the encoding target block is stored in a form in which the motion compensation prediction block size is combined (S501). Specific behavior will be described with reference to FIG. FIG. 6A shows the relationship between the encoding target block and the adjacent decoded block. If the prediction block size of direct motion compensated prediction is 16 × 16, the encoding process is performed from left to right and from top to bottom in units of 16 × 16 macroblocks, so left (A), top ( The motion compensated prediction information for the four macroblocks B), upper right (C), and upper left (D) is adjacent block information.

  Each macroblock is motion-compensated with various block sizes. As shown in FIG. 6B, a plurality of motion compensations are performed in a macroblock that is motion-compensated and predicted in units smaller than the macroblock. Predictive information exists.

  In the standard reference image determination process according to the first embodiment, a configuration is adopted in which the motion compensated prediction information is weighted and evaluated by the block size, and each block position adjacent to the encoding target for the minimum motion compensated prediction block is taken. The motion compensation prediction information used for the motion compensation of that portion is stored. In motion compensation prediction of MPEG4-AVC, a minimum 4 × 4 pixel unit is used when the adaptive motion compensation block size in the sub macroblock is used, and a minimum 8 × 8 pixel unit is used when the adaptive motion compensation block size is not used. The following description will be given in the case where 8 × 8 pixels are the minimum motion compensation prediction block size, but the same processing can be realized in the case of 4 × 4 pixels.

  FIG. 6C shows the motion compensation prediction information to be stored when the motion compensation prediction of the neighboring blocks as shown in FIG. 6B is performed. Motion compensated prediction includes unidirectional prediction performed using only one reference image and bidirectional prediction performed using two reference images. In the case of unidirectional prediction, one motion is used as motion compensated prediction information at the same position. A vector value and one reference ID are stored. On the other hand, in the case of bidirectional prediction, two motion vector values and two reference IDs are stored.

  In the example shown in FIG. 6C, the encoding target block has a configuration of 16 × 8 blocks on the left, 8 × 16 blocks on the top, 8 × 8 blocks on the top right, and 16 × 16 blocks on the top left. The upper macroblock is unidirectionally predicted only by the first reference, and the upper left macroblock is unidirectionally predicted only by the second reference.

  In this case, mvA1, mvA2 and the respective reference IDs for the first reference are stored as the motion compensation prediction information of the left macroblock, and the mvA1, mvA2 and the respective reference IDs for the second reference are stored. MvB1, mvB2 and their respective reference IDs for the first reference, mvC3 and reference ID for the first reference as motion compensation prediction information for the upper right macroblock, mvC3 and reference ID for the second reference, and motion compensation prediction information for the upper left macroblock A total of nine pieces of motion compensation prediction information of mvD and reference ID for the second reference are stored.

  Subsequently, the reference IDs of the stored motion compensation prediction information are totaled for each reference image, and the most reference images are calculated (S502). For example, when motion compensation prediction of 16 × 16 pixels is selected in the left or upper macroblock, the same motion compensation prediction is stored when the motion compensation prediction block is divided into the smallest motion compensation prediction blocks. The information is weighted according to the size. Further, the first reference and the second reference are aggregated without distinction, and the number of times of application for each reference image finally indicated by the motion vector is counted.

  In the first embodiment, when there is one reference image having the largest number (S503: YES), the reference image is selected as a standard reference image (S504). If there are a plurality of reference images (S503: NO), the reference image is preferentially selected as a reference image closer to the encoding target block, and if the distance is equal, the code is preceded in time. It is determined that the converted reference image has a smaller number of times of predictive encoding (S505).

  In addition, when there are a plurality of reference images, the quantization value when the reference image is encoded is acquired on the encoding side / decoding side, and the reference image is subjected to finer quantization. However, it can be preferentially selected based on the judgment that the quality is good, which is one derivation example of the standard reference image determination processing in the first embodiment. The reference image that is most frequently referred to as motion compensated prediction in the peripheral region is judged to have high reliability as the reference image quality and the temporal correlation with the image in which the block to be encoded exists are high. In the first embodiment, it is a point of the standard reference image determination process, a large amount of information that can be used as a motion vector prediction candidate can be acquired, and the motion vector prediction accuracy is increased.

  Subsequently, a process of calculating a predicted motion vector value for the selected standard reference image is performed. First, when there are three or more pieces of motion compensated prediction information whose standard reference image is the reference ID (S506: YES), the horizontal and vertical median values of the motion vector values in the motion compensated prediction are calculated ( S509). The method for calculating the median is the same as the method for generating a predicted vector value in MPEG4-AVC, but is suitable for prediction because the motion vector value is weighted to improve the priority of a motion vector having a large area. As a result of the improvement in accuracy by using a reference image for which a large amount of information can be acquired, it is possible to perform prediction with higher accuracy than the conventional motion vector predictor. If an even-numbered motion vector is acquired (S507: YES), the median value is acquired (S509) after excluding the motion vector value farthest from the average of the two central motion vector values (S508).

  When the motion compensation prediction information whose reference ID is the reference reference image is two or less (S506: NO), the average value of the motion vector values is calculated (S511) in two cases (S510: YES). In one case (S510: NO), the motion vector value is set as a predicted motion vector value (S512).

  In this way, the calculated reference ID and predicted motion vector value are output to the inter-reference image motion vector detection units 108 and 215 (S513), and the standard reference image determination process ends.

  Subsequently, the inter-reference image motion vector detecting unit shown in FIG. 7 will be described with reference to the inter-reference image motion vector detecting process, which operates according to the inter-reference image motion vector detecting units 108 and 215 and S403 and S414 in the processing flowchart of FIG. This is performed using the internal configuration diagram and the processing flowchart shown in FIG.

  The inter-reference image motion vector detection unit 108 includes a standard reference image acquisition unit 700, a motion vector detection range setting unit 701, a standard reference image memory 702, a reference image acquisition unit 703, and a block matching evaluation unit 704.

  First, based on the reference ID and the predicted motion vector value mvL0 of the reference reference image input from the reference reference image determination unit 107, the reference reference image acquisition unit 700 encodes the reference reference image from the decoded reference image memory 118. The image block that has moved mvL0 from the block is cut out, and the first reference block is acquired (S800). The acquired reference block is stored in the standard reference image memory 702.

  Subsequently, the motion vector detection range setting unit 701 determines the reference ID of the second reference image based on the reference ID of the standard reference image and the predicted motion vector value mvL0 input from the standard reference image determination unit 107 (S801). ). As described above, in the first embodiment, the other reference ID for the standard reference image is transmitted in units similar to the slice header in MPEG4-AVC as information in units of screens, and the corresponding information is referred to. Thus, the reference ID of the second reference image can be determined.

  Subsequently, a detection range of the inter-reference image motion vector to be detected between the standard reference image and the second reference image is set (S802). Regarding the detection range, it is also possible to take the entire area of the second reference image as the motion vector detection range for the first reference block, and by performing detection processing with the same definition as the encoding device and the decoding device. Although functioning, in order to reduce the amount of calculation in detecting a motion vector between reference images, a detection range as shown in FIG. 9 is set.

FIG. 9 is an example of a motion vector detection range between reference images in the first embodiment. If the input time of the encoding target image is Poc_Cur, the input time of the standard reference image is Poc_Ref1, and the input time of the second reference image is Poc_Ref2, the second motion vector mvL0 from the standard reference image for the encoding target block is When the search range of the reference image is based on the position of the encoding target block, the search center position is α = mvL0 × (Poc_Cur−Poc_Ref2) / (Poc_Cur−Poc_Ref1).
As shown in the above, the motion vector prediction value between the encoding target block and the second reference image when it is assumed that the motion is temporally continuous is set.

  However, since there are many situations that are not temporally continuous changes, such as camera movements and object movements, an appropriate reference block of the second reference image can be obtained by searching for a motion vector for a specific region with the search position as the center. To be able to get In the example shown in FIG. 9, an area of ± 8 pixels is designated as the specific area.

  The reference image acquisition unit 703 acquires the reference block of the second reference image in the motion vector detection range specified by the motion vector detection range setting unit 701 from the decoded reference image memory 118 (S803), and the block matching evaluation unit 704. Output to.

  The block matching evaluation unit 704 calculates an error sum for each pixel between the first reference block stored in the standard reference image memory 702 and the reference block of the second reference image input from the reference image acquisition unit 703. Then, a reference block having a smaller total sum and a motion vector value when the reference block is acquired are stored (S804).

  After calculating the total error for each pixel for all the motion vector detection ranges (S805: YES), the stored reference block is set as the second reference block, and the motion vector value is set as the inter-reference image motion vector. The value mvInterRef is set (S806).

  Regarding the detection accuracy of motion vectors between reference images, it is possible to apply a method of implicitly detecting motion vectors with the same detection accuracy in the encoding device and the decoding device, but detection of motion vectors for each frame or reference image used. It is also possible to use a technique of transmitting accuracy as encoded information. Here, as an implicit setting, the detection accuracy is 1/4 pixel accuracy. The calculated inter-reference image motion vector value mvInterRef is output to the direct motion compensation prediction unit 109 together with mvL0, the first reference block, and the second reference block (S807).

  In the moving image encoding device and the moving image decoding device in Embodiment 1, a function of selecting a reference image suitable for prediction based on the motion compensated prediction information that has been encoded in the peripheral blocks of the encoding target block, By having both the encoding side and the decoding side, the accuracy of the motion vector that is implicitly generated and the reference image to be used can be made more appropriate, and the prediction block in which the obtained reference image is extracted as the standard reference image And a function to perform bi-directional prediction using the detected motion vector on both the encoding side and the decoding side, so that additional information such as a motion vector can be obtained. Without transmission, it is possible to improve the prediction efficiency for a moving image signal accompanied by a temporal change in motion, and to improve the coding efficiency.

(Embodiment 2)
Next, a video encoding device and a video decoding device according to Embodiment 2 will be described. Also in the second embodiment, the configurations of the moving image encoding device and the moving image decoding device are the same as those of the first embodiment, but as shown in FIGS. 10 and 11, the reference reference image determination units 107 and 214 are used. Is different in that the reference image signal is input from the decoded reference image memories 118 and 209 and the adjacent intra-frame decoded signal from the intra-frame decoded image memories 117 and 208 is used for the reference image determination process.

  In the process, since the reference image determination process and the inter-reference image motion vector detection process perform different operations, FIG. 12 shows a process flowchart of the reference image determination process in the second embodiment, and FIG. 13 shows a reference image in the second embodiment. A process flowchart of the inter-motion vector detection process is shown, and the operation of the second embodiment will be described. Note that, as in the first embodiment, these functions function by performing similar processing on the encoding side and decoding side, and therefore will be described as a common configuration.

  In the second embodiment, in order to further improve the accuracy of determining the reference image in the first embodiment, the motion vector for the reference image that is not used with respect to the motion compensated prediction information used in the peripheral blocks of the encoding target block. As a result, it is possible to evaluate whether or not satisfactory motion compensation prediction was possible when using motion vector information spatially and temporally adjacent to each other by performing motion compensation prediction between decoded images.

  As a result, it is possible to calculate more motion vectors that can be referenced for prediction for each reference image of the peripheral block, and a reference image that more appropriately reflects the importance of motion compensated prediction information, and a predicted motion vector value Can be selected.

  Furthermore, the second embodiment is based on the error evaluation value between reference images when detecting a motion vector between the first reference block acquired from the standard reference image and a plurality of other reference images. It has a function of determining which reference image is used for bidirectional prediction and determining whether bidirectional prediction is performed in units smaller than the size of the reference block.

  This makes it possible to perform bi-directional prediction in small units without additional information, using reference image information that is partially suitable for moving image signals accompanying screen changes such as scene changes or when objects are hidden. It becomes possible.

  First, reference image determination processing will be described. As shown in the flowchart of FIG. 12, in the second embodiment as well, the motion compensated prediction motion first used in the peripheral block of the encoding target block (decoding target block in the case of a decoding device) is the same as in the first embodiment. Vector values, reference IDs, and motion compensation block size information are acquired (S1200), and motion compensation prediction information adjacent to the block to be encoded is stored in a form in which the motion compensation prediction block size is combined (S1201).

  Subsequently, when motion compensation prediction information as a target is stored for each reference image (S1202: YES), the motion compensation prediction information is determined and used as information of a reference image indicated by a reference ID ( S1203).

  If the target motion compensation prediction information is not stored (S1202: NO), a search is made for a block in which the reference image of the corresponding reference ID is used in the motion compensation information in a block further adjacent to the corresponding adjacent block. Process.

  Specifically, as shown in FIG. 14, when a block around the block that does not have motion compensation information for the target reference image uses a reference image with the corresponding reference ID (S1204: YES), The motion vector value of the adjacent block is set as a candidate vector value (S1205). When the target block uses a reference image of another reference ID (S1206: YES), the motion vector value in the reference ID closest to the reference image that is the target in the reference ID is defined as the encoding target block. The value reduced or enlarged according to the distance ratio is set as a candidate vector value (S1207).

Specifically, when the input time of the reference image RefB indicated by the motion compensation information in FIG. 14 is Poc_RefB, the input time of the target reference image RefA is Poc_RefA, and the input time of the encoding target block is Poc_Cur, the motion vector for RefB If the value is mvRefB,
mvRefA_cand = mvRefB × (Poc_Cur−Poc_RefA) / (Poc_Cur−Poc_RefB)
Thus, a candidate vector value is calculated.

  Subsequently, for the block of the target reference image for which the candidate vector can be calculated, the reference image and the intra-frame decoded image are obtained from the reference image memory and the intra-frame decoded image memory, and the motion compensation is performed on the peripheral blocks using the candidate vector. An error value at the time of prediction is calculated (S1208). Although a plurality of candidate vector values can be taken, error values are evaluated for all, and the candidate motion vector value having the smallest error value is set as a temporary motion vector value (S1209).

  Next, in order to determine whether or not the temporary motion vector value is a reliable peripheral motion vector value, the error value is compared with a threshold value α. If the error value is smaller than α (S1210: YES), The motion vector value is determined as motion compensation prediction information and stored as a motion vector of the target reference image (S1211). On the other hand, when the error value is greater than or equal to the threshold value α (S1210: NO), the temporary motion vector value is not adopted as the motion compensation prediction information, and it is determined that there is no information on the target reference image (S1212).

  Regarding the threshold value α, it is possible to transmit a fixed value as an implicit value (eg, 2 × block size) or information in units of slices, but an error when the target block is originally motion compensated predicted. If the value is calculated using the decoded image and compared with the value, for example, 4/3 or less, it can be determined that the value is reliable.

  When the evaluation of the candidate vectors and the interpolation of the motion compensated prediction information are completed for all adjacent blocks and reference IDs (S1213: YES), the stored motion compensated prediction information, which is the same process as in the first embodiment. Are counted by reference image, and the most reference images are calculated (S1214). The first reference and the second reference are also aggregated without distinction, and finally the number of times of application for each reference image indicated by the motion vector is counted.

  Also in the second embodiment, when the number of reference images is the largest as in the first embodiment (S1215: YES), the reference image is selected as a standard reference image (S1216). If there are a plurality of reference images (S1216: NO), the reference image is preferentially selected as a reference image closer to the encoding target block, and if the distance is equal, the code is preceded in time. It is determined that the converted reference image has a smaller number of times of predictive encoding (S1217).

  The process of calculating a predicted motion vector value for the selected reference reference image is the same as that in the first embodiment, but since the adopted candidate vector values are stored, a predicted vector suitable for more motion vector values. Can be calculated. First, when there are three or more pieces of motion compensated prediction information whose standard reference image is a reference ID (S1218: YES), the horizontal and vertical median values of the motion vector values in the motion compensated prediction are calculated ( S1221). If an even-numbered motion vector is acquired (S1219: YES), the motion vector value farthest from the average of the two central motion vector values is excluded (S1220), and then the median value is acquired (S1221).

  When the motion compensation prediction information whose reference ID is the reference reference image is two or less (S1218: NO), the average value of the motion vector values is calculated (S1223) in two cases (S1222: YES). In one case (S1222: NO), the motion vector value is set as a predicted motion vector value (S1224).

  In this way, the calculated reference ID and predicted motion vector value are output to the inter-reference image motion vector detection units 108 and 215 (S1225), and the standard reference image determination process is terminated.

  Next, reference inter-motion vector detection processing will be described. As shown in the flowchart of FIG. 13, in the second embodiment as well, in the same manner as in the first embodiment, the reference ID and predicted motion vector of the reference reference image input from the reference reference image determination unit 107 by the reference reference image acquisition unit 700. Based on the value mvL0, the image block moved mvL0 from the coding target block of the standard reference image is extracted from the decoded reference image memory 118, and the first reference block is acquired (S1300). The acquired reference block is stored in the standard reference image memory 702.

  Subsequently, the motion vector detection range setting unit 701, based on the reference ID and the predicted motion vector value mvL0 of the standard reference image input from the standard reference image determination unit 107, The detection range of the motion vector between the reference images is set (S1301). The detection method can be performed in the same manner as in the first embodiment, but is different in that it is performed on a plurality of reference images.

  The reference image acquisition unit 703 acquires, from the decoded reference image memory 118, the reference blocks in the motion vector detection range in the corresponding reference image specified by the motion vector detection range setting unit 701 for all reference images ( S1302) and output to the block matching evaluation unit 704.

  The block matching evaluation unit 704 blocks the error sum for each pixel between the first reference block stored in the standard reference image memory 702 and the reference block of the corresponding reference image input from the reference image acquisition unit 703. The calculation is performed for each unit smaller than the size, and the reference block having a small sum by the small unit and the motion vector value when the reference block is acquired are stored (S1303). In the second embodiment, the block size of the first reference block is 16 × 16 pixels, the unit smaller than the block size is 4 × 4 pixels, and reference to each reference image is performed for 16 small block sizes. An inter-image motion vector value mvInterRef and an error evaluation value in block matching are stored.

  When the detection of motion vectors between reference images is completed for all reference images (S1304: YES), the error evaluation values of all reference images are compared for each 4 × 4 block, and the reference image having the smallest value is determined. The second reference image is determined (S1305). Regarding the error evaluation value in this case, it is possible to use the error evaluation value at the time of block matching as it is, but it is also possible to make a comparison with the error value from which the DC component of the error is removed. It is possible to generate a bidirectional prediction signal that makes use of a change in DC.

  Subsequently, when the calculated error evaluation value in the second reference image is larger than the threshold β (eg, 100) (S1306: YES), a signal that is partially suitable for bidirectional prediction due to the visibility of the object or the like It is determined that the second reference image is not selected, and one-way prediction is performed (S1307).

  As a result of performing the optimum reference image and bidirectional / one-way determination for each 4 × 4 block in this way, the reference image and motion vector conforming to the block shape as shown in the example of FIG. A value can be generated. It is possible to realize a configuration in which an appropriate predicted image is synthesized for each part with respect to a reference block of a reference image serving as a reference.

  If these processes are performed for all 4 × 4 blocks (S1308: YES), the reference ID and the inter-reference image motion vector value mvInterRef in units of 4 × 4 blocks, and unidirectional prediction or bidirectional prediction are shown. A second reference block is generated based on the information (S1309).

  The image of the second reference block for the block for which unidirectional prediction is performed is not used in the direct motion compensation prediction unit. For example, if the same pixel data of the first reference block is stored, the average value is calculated for all the blocks in the block. Even when applied to the value of, a unidirectional prediction block can be generated.

  Finally, the reference ID and reference image motion vector value mvInterRef in units of 4 × 4 blocks, and information indicating unidirectional prediction or bi-directional prediction, together with mvL0, the first reference block, and the second reference block, It outputs to the direct motion compensation prediction part 109 (S1310).

  According to the moving picture coding apparatus and the moving picture decoding apparatus in the second embodiment, the motion vector result for the reference image that is not used is obtained with respect to the motion compensated prediction information used in the peripheral blocks of the coding target block. A reference image that more appropriately reflects the importance of motion compensated prediction information by having a function to evaluate using the decoded image with spatially and temporally adjacent motion vector information on both the encoding side and the decoding side The accuracy of the standard reference image determination process in Embodiment 1 can be improved, and the direct mode motion compensation prediction efficiency generated using the result can be further improved.

  Furthermore, in Embodiment 2, motion vector detection between reference images is performed between a plurality of reference images for a motion compensated prediction block of a reference image serving as a reference, and bi-directional prediction is performed based on error evaluation values. A video that accompanies a screen change such as a scene change or when an object is hidden by having both the encoding side and the decoding side have a function of determining whether the other reference image is to be determined and performing bidirectional prediction. It is possible to generate a motion compensated prediction signal using only reference image information suitable for the signal without additional information, and the motion compensation prediction efficiency in the direct mode can be further improved.

  According to each embodiment of the present invention described above, a function of selecting a reference image suitable for prediction (reference numerals 107, 215, 5)) on both the encoding side and the decoding side, a reference image suitable for the direct mode in which motion compensation prediction is performed without transmitting motion vector information can be recognized without additional information, so that prediction efficiency is improved. Therefore, it is possible to realize moving picture coding with improved coding efficiency.

  In addition, according to each embodiment, a motion vector between another reference image is detected for a motion compensated prediction block of a reference image serving as a reference (see reference numerals 108, 215, FIG. 7, and FIG. 8). ), By generating a motion compensated prediction block of another reference image based on the detected motion vector and performing bi-directional prediction (see reference numerals 109 and 216) on both the encoding side and the decoding side, motion vector information Therefore, it is possible to improve the prediction efficiency for a moving image signal accompanied by a temporal change in motion without transmitting the image, and to realize moving image coding with improved coding efficiency.

  Furthermore, according to the second embodiment, an error evaluation value between reference images when a motion vector between a plurality of other reference images is detected with respect to a motion compensated prediction block of a reference image serving as a reference is obtained. Based on the fact that both the encoding side and the decoding side have a function (see S1303 to S1308 in FIG. 13) for determining whether to perform bi-prediction and determination of the other reference image for bi-directional prediction, A motion compensated prediction signal that uses only reference image information suitable for moving image signals accompanying screen changes such as scene changes or when objects are hidden can be generated without additional information, improving coding efficiency. It is possible to realize the encoded video encoding.

  Furthermore, according to the second embodiment, the motion vector results for the reference images that are not used are spatially and temporally adjacent to the motion compensated prediction information used in the peripheral blocks of the encoding target block. A reference image that more appropriately reflects the importance of motion compensated prediction information by having a function of evaluating a motion vector information using a decoded image (see S1202 to S1213 in FIG. 12) on both the encoding side and the decoding side. Can be selected, and the prediction efficiency of the direct mode can be further improved.

  The moving picture encoding apparatus and moving picture decoding apparatus presented as the first, second, third, fourth, and fifth embodiments are physically a CPU (central processing unit), a memory, and the like. Can be realized by a computer equipped with a recording device, a display device such as a display, and a communication means to the transmission path, and the means having each presented function is realized as a program on the computer and executed. Is possible. In addition, the program can be provided by being recorded on a computer-readable recording medium, provided from a server through a wired or wireless network, or provided as data broadcasting of terrestrial or satellite digital broadcasting. is there.

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

  101 input image buffer, 102 block division unit, 103 intra-frame prediction unit, 104 motion vector detection unit, 105 motion compensation prediction unit, 106 motion vector prediction unit, 107 reference reference image determination unit, 108 inter-reference image motion vector detection unit, 109 direct motion compensation prediction unit, 110 prediction mode determination unit, 111 subtractor, 112 orthogonal transform unit, 113 quantization unit, 114 inverse quantization unit, 115 inverse orthogonal transform unit, 116 adder, 117 intra-frame decoded image memory, 118 Decoding Reference Image Memory, 119 Entropy Coding Unit, 120 Stream Buffer, 122 Code Quantity Control Unit, 201 Stream Buffer, 202 Entropy Decoding Unit, 203 Prediction Mode Decoding Unit, 204 Predicted Image Selection Unit, 205 Quantization unit, 206 inverse orthogonal transform unit, 207 adder, 208 intra-frame decoded image memory, 209 decoded reference image memory, 211 intra-frame prediction unit, 212 motion vector prediction decoding unit, 213 motion compensation prediction unit, 214 reference reference image Determination unit, 215 inter-reference image motion vector detection unit, 216 direct motion compensation prediction unit, 700 standard reference image acquisition unit, 701 motion vector detection range setting unit, 702 standard reference image memory, 703 reference image acquisition unit, 704 block matching evaluation Department.

Claims (5)

A mode detection unit for detecting mode information indicating a mode in which the designation information of the reference image at the time of encoding and the specification information of the motion vector at the time of encoding are not transmitted;
When the mode information is detected by the mode detection unit, the first reference image and the first motion vector value used for the motion compensation process of the decoding target block are used as the motion compensation prediction in the peripheral blocks of the decoding target block. A reference reference image determination unit that selects based on the information used in
A reference image synthesizing unit that generates a synthesized reference block obtained by synthesizing the first reference block extracted from the first reference image using the first motion vector value and a predetermined region of at least one other reference image; ,
A moving picture decoding apparatus comprising: a decoding unit that generates a decoded image by adding the prediction block and a prediction difference block decoded from the decoding target block using the synthesized reference block as a prediction block .   The reference reference image determination unit obtains a motion vector value for a reference image that is not used for motion compensated prediction in a peripheral block of the encoding target block, a motion vector value for the reference image used for motion compensated prediction, and a decoded image signal. And determining whether or not to add a reference image that is not used for motion compensation prediction to the first reference image selection candidate according to the calculated motion vector value. Adding the calculated motion vector value to a motion vector candidate for a reference image that is not used in the motion compensated prediction, thereby increasing the frequency of the reference image that is not used in the motion compensated prediction. And the first motion image using the motion vector value added to the candidate in the selected first reference image. Video decoding apparatus according to claim 1, characterized in that the generating vector values.   Based on a plurality of second motion vector values between a plurality of reference images detected by the reference image synthesizing unit in units of blocks having a size equal to or smaller than the first reference block, a second reference block is obtained. A reference image to be used is selected, and a combined reference block is generated by combining the first reference block and the second reference block, and the combined reference block is used as a prediction block, or the first reference block is predicted. The moving picture decoding apparatus according to claim 1, wherein it is determined whether to use as a block. Detecting mode information indicating a mode in which the designation information of the reference image at the time of encoding and the designation information of the motion vector at the time of encoding are not transmitted;
When the mode information is detected, the first reference image and the first motion vector value used for the motion compensation process of the decoding target block are used for the motion compensation prediction in the peripheral blocks of the decoding target block. A step to select based on
Generating a synthesized reference block obtained by synthesizing the first reference block extracted from the first reference image using the first motion vector value and a predetermined region of at least one other reference image;
A moving picture decoding method comprising: generating a decoded image by adding the prediction block and a prediction difference block decoded from the decoding target block using the synthesized reference block as a prediction block. A function for detecting mode information indicating a mode that does not transmit designation information of a reference image at the time of encoding and designation information of a motion vector at the time of encoding;
When the mode information is detected, the first reference image and the first motion vector value used for the motion compensation process of the decoding target block are used for the motion compensation prediction in the peripheral blocks of the decoding target block. A function to select based on
A function of generating a synthesized reference block obtained by synthesizing the first reference block extracted from the first reference image using the first motion vector value and a predetermined region of at least one other reference image;
A moving image characterized by causing a computer to realize a function of generating a decoded image by adding the prediction block and a prediction difference block decoded from the decoding target block using the synthesized reference block as a prediction block Decryption program.