JP2010233135A - Moving image encoding device and moving image decoding device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a moving image encoding device and a moving image decoding device for improving predictive vector performance while suppressing the quantity of additional information for various image patterns. <P>SOLUTION: The moving image encoding device AA that allows inter-coding in encoding for the unit of a macro block, includes: a predictive vector control section 22 for inferring reduction in encoding performance in a processing block in inter-coding; and a first predictive vector generation section 23 for generating a predictive vector in the processing block on the basis of inference with the predictive vector control section 22. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention divides a screen into a plurality of macroblocks (hereinafter referred to as “MB”), a moving picture coding apparatus that allows inter coding when coding in MB, and decoding in MB In particular, the present invention relates to a moving picture decoding apparatus that allows inter coding and a moving picture decoding apparatus that generates a prediction vector of a motion vector in a processing block from a motion vector in a coded block. .

  2. Description of the Related Art Conventionally, in a video encoding apparatus that allows inter encoding when encoding is performed in units of MB, inter encoding is performed in units of MBs or blocks obtained by subdividing MBs (hereinafter simply referred to as “blocks”). Is going. In this case, for example, a prediction vector is generated by a method described in Non-Patent Document 1 described later or a method described in Non-Patent Document 2 described later. According to these methods, by encoding only the difference between the prediction vector and the motion vector obtained by motion compensation, the amount of code required for encoding the motion vector can be suppressed.

  Here, a block that generates a prediction vector is a processing block, an encoded block adjacent to the processing block is a reference block, and a motion vector of the reference block is a reference vector. Then, in the prediction vector production | generation method of a nonpatent literature 1, the vector formed by the median of a reference vector is employ | adopted as a prediction vector.

  Specifically, in the prediction vector generation method described in Non-Patent Document 1, the median (center) regarding the x component and the y component of all reference vectors is calculated for two or three reference blocks, and the calculated x component is calculated. A vector formed by the y component and the y component is a prediction vector. The prediction vector generation method described in Non-Patent Document 1 will be described below with reference to FIG.

In FIG. 7, B _pred indicates a processing block that is a block for generating a prediction vector. On the other hand, B _a , B _b , and B _c indicate reference blocks that are adjacent to the processing block B _pred .

V _pred indicates a prediction vector in the processing block B _pred , and each of V _a , V _b , and V _c indicates a reference vector in each of the reference blocks B _a , B _b , and B _c . Here, the reference vectors V _a , V _b , and V _c can be expressed as in Expression (1).

In the formula _(1), each of x _a, x b, _{x c,} reference vector _{V a,} V b, shows the respective x component of _{V c,} each _{y a, y b, y c} , reference vector Each y component of V _a , V _b , and V _c is shown.

_Assuming that the x component is x _pred and the y component is y _pred for the prediction vector V _pred , in the prediction vector generation method described in Non-Patent Document 1, the x component x _pred and the y component y _pred of the prediction vector V _pred are It is expressed as equation (2).

  Note that the function median in Expression (2) is a function for obtaining the median of the argument, and can be expressed as Expression (3).

  However, since the predictive vector generation method described in Non-Patent Document 1 does not consider the pattern feature, the difference between the predictive vector and the motion vector obtained by motion compensation becomes large, which is necessary for encoding the motion vector. There is a possibility that the amount of codes increases.

  On the other hand, in the prediction vector generation method described in Non-Patent Document 2, not only a vector formed by a median of reference vectors but also a reference vector itself can be adopted as a prediction vector. For this reason, compared with the prediction vector generation method described in Non-Patent Document 1, the types that can become prediction vectors are increased. Therefore, the difference between the prediction vector and the motion vector obtained by motion compensation is reduced to reduce the prediction vector. Performance can be improved.

Joint Video Team (JVT) of ISO / IEC MPEG and ITU-VCEG, “Text of ISO / IEC 14496 10 Advanced Video Coding 3rd Edition”, July 2004. G. Laroche, “RD Optimized Coding for Motion Vector Predictor Selection”, IEEE Trans. On Circuits and systems, Dec. 2008

  In the prediction vector generation method described in Non-Patent Document 1, as described above, the vector formed by the median of the reference vector is adopted as the prediction vector. In some cases, the decrease in the predicted vector performance was significant. Deterioration of the prediction vector performance near the edge in the prediction vector generation method described in Non-Patent Document 1 will be described below with reference to FIG. Note that an edge indicates a portion where a change in signal is locally large, such as a boundary between objects having different motions.

In FIG. 8, MV indicates a motion vector in the processing block B _pred . In FIG. 8, the reference block B _a and the processing block B _pred include horizontal edges extending in the horizontal direction, and the reference block B _b and the reference block B _c , the reference block B _a and the processing block are included. The motion vector is different from B _pred .

Here, in order to simplify the description, the reference vectors V _a , V _b , and V _c are expressed as in Expression (4).

Further, a reference vector _{V a,} the motion vector MV in the processing block _{B pred,} assuming equal, it is possible to represent this motion vector MV as in Equation (5).

On the other hand, assuming that a vector formed by medians of the reference vectors V _a , V _b , and V _c is a predicted vector V _pred by the predicted vector generation method described in Non-Patent Document 1, the predicted vector V _pred is _expressed by the following equation (6). It can be expressed as

According to the above equations (5) and (6), the prediction vector V _pred deviates from the motion vector MV in the processing block B _pred . For this reason, since the difference between the prediction vector V _pred and the motion vector MV increases, the amount of code required to encode the motion vector increases, and the prediction vector performance near the edge decreases.

  On the other hand, in the prediction vector generation method described in Non-Patent Document 2, as described above, not only the vector formed by the median of the reference vector but also the reference vector itself can be adopted as the prediction vector, so that the prediction vector The design adaptability is improved. According to this method, even in the vicinity of an edge, an appropriate prediction vector can be selected by adopting a reference vector at a position along the direction in which the edge extends as a prediction vector. It can be prevented from lowering.

  However, in the prediction vector generation method described in Non-Patent Document 2, control information is required to select an appropriate prediction vector. For this reason, in the pattern in which sufficient prediction vector performance can be obtained by the prediction vector generation method described in Non-Patent Document 1, only an increase in additional information called control information is caused.

  Therefore, the present invention has been made in view of the above-described problems, and a moving picture coding apparatus and a moving picture decoding capable of improving the prediction vector performance while suppressing the amount of additional information for various patterns. An object is to provide an apparatus.

  The present invention proposes the following matters in order to solve the above problems.

  (1) The present invention provides a prediction vector for estimating a decrease in coding performance in a processing block in inter coding in a moving image coding apparatus that allows inter coding in coding in units of macroblocks. Based on the estimation by the control means (for example, the prediction vector control unit 22 in FIG. 2) and the prediction vector control means, the prediction vector generation means (for example, the first vector in FIG. 2) generates the prediction vector in the processing block. The prediction vector generation unit 23).

  According to the present invention, a moving picture coding apparatus that allows inter coding when coding in units of macroblocks, and a prediction vector control unit that estimates a decrease in coding performance in a processing block during inter coding. And prediction vector generation means for generating a prediction vector in the processing block based on the estimation by the prediction vector control means. For this reason, the prediction vector performance can be improved and the coding efficiency can be improved while suppressing the amount of additional information for various patterns.

  (2) The present invention relates to the video encoding apparatus according to (1), wherein the prediction vector control means performs encoding in the processing block based on information on a motion vector in an encoded block adjacent to the processing block. A prediction vector generation method is determined by inferring a decrease in performance, and the prediction vector generation means is a coded block adjacent to the processing block based on the prediction vector generation method determined by the prediction vector control means. The video encoding device is characterized in that an optimal vector is selected as the prediction vector from the motion vectors in the above and used as the prediction vector.

  According to the present invention, the prediction vector control means determines a prediction vector generation method by estimating a decrease in the encoding performance in the processing block based on the information on the motion vector in the encoded block adjacent to the processing block. It was decided. Further, the prediction vector generation means selects an optimal vector as a prediction vector from the motion vectors in the encoded block adjacent to the processing block based on the prediction vector generation method determined by the prediction vector control means, It was decided to. According to this, an effect similar to the effect mentioned above can be produced.

  (3) The present invention relates to the moving picture coding apparatus according to (1) or (2), wherein the prediction vector control means is based on a motion vector component or magnitude in a coded block adjacent to the processing block. The video encoding apparatus is characterized by estimating a decrease in encoding performance in the processing block.

  According to the present invention, the prediction vector control means estimates a decrease in coding performance in the processing block based on the motion vector component or magnitude in the coded block adjacent to the processing block. According to this, an effect similar to the effect mentioned above can be produced.

  (4) The present invention relates to the moving picture coding apparatus according to (1) or (2), wherein the prediction vector control means is based on a component or magnitude of a difference vector in a coded block adjacent to the processing block. The video encoding apparatus is characterized by estimating a decrease in encoding performance in the processing block.

  According to the present invention, the prediction vector control means estimates a decrease in encoding performance in the processing block based on the component or magnitude of the difference vector in the encoded block adjacent to the processing block. According to this, an effect similar to the effect mentioned above can be produced.

  (5) In the moving picture encoding apparatus according to any one of (1) to (4), the present invention provides that the prediction vector control means includes a motion vector component or magnitude in an encoded block adjacent to the processing block. Alternatively, a median of a difference vector component or magnitude in an encoded block adjacent to the processing block is calculated, and the calculated median and a motion vector component or magnitude in the encoded block, or Based on the evaluation by the evaluation unit that calculates a difference between the component or the size of the difference vector in the encoded block and evaluates the encoded block having the maximum calculated difference, the processing block A video encoding device comprising: an estimation unit that estimates a decrease in encoding performance in It has proposed a location.

  According to the present invention, the prediction vector control means is provided with the evaluation unit and the estimation unit. Then, the evaluation unit calculates the median of the motion vector component or magnitude in the encoded block adjacent to the processing block, or the difference vector component or magnitude in the encoded block adjacent to the processing block. The difference between the calculated median and the motion vector component or magnitude in the encoded block or the difference vector component or magnitude in the encoded block is calculated, and the encoded block having the maximum difference is calculated. We decided to evaluate. In addition, the estimation unit estimates a decrease in coding performance in the processing block based on the evaluation by the evaluation unit. According to this, an effect similar to the effect mentioned above can be produced.

  (6) In the moving image encoding apparatus according to any one of (1) to (4), the present invention provides that the prediction vector control means includes a motion vector component or magnitude in an encoded block adjacent to the processing block. Alternatively, the average of the difference vector component or magnitude in the encoded block adjacent to the processing block is calculated, and the calculated average and the motion vector component or magnitude in the encoded block, or Based on the evaluation by the evaluation unit that calculates a difference between the component or the size of the difference vector in the encoded block and evaluates the encoded block having the maximum calculated difference, the processing block And a estimator for estimating a decrease in encoding performance of the video encoding device. There.

  According to the present invention, the prediction vector control means is provided with the evaluation unit and the estimation unit. Then, the evaluation unit calculates and calculates the motion vector component or magnitude in the encoded block adjacent to the processing block, or the average of the difference vector component or magnitude in the encoded block adjacent to the processing block. The difference between the average and the motion vector component or magnitude in the encoded block or the difference vector component or magnitude in the encoded block is calculated, and the encoded block having the maximum difference is calculated. We decided to evaluate. In addition, the estimation unit estimates a decrease in coding performance in the processing block based on the evaluation by the evaluation unit. According to this, an effect similar to the effect mentioned above can be produced.

  (7) The present invention relates to the video encoding device according to (5), in which the prediction vector generation unit is configured to perform the processing when the estimation unit estimates that there is a concern about a decrease in encoding performance in the processing block. There has been proposed a moving picture coding apparatus characterized in that a motion vector in a coded block adjacent to a block is used as the prediction vector.

  According to this invention, when the prediction vector generation unit estimates that there is a concern that the encoding unit in the processing block may be deteriorated by the estimation unit, the motion vector in the encoded block adjacent to the processing block is used as the prediction vector. It was decided. According to this, an effect similar to the effect mentioned above can be produced.

  (8) The present invention relates to the video encoding device of (5), wherein the prediction vector generation means predicts the prediction when the estimation unit estimates that there is a concern about a decrease in encoding performance in the processing block. There has been proposed a moving picture coding apparatus characterized in that control information for identifying a motion vector adopted as a vector is added to a bit stream as coded data of the processing block.

  According to the present invention, when the prediction vector generation unit estimates that there is a concern about a decrease in the coding performance in the processing block by the estimation unit, the control information for identifying the motion vector adopted as the prediction vector is processed. The bit stream is encoded data of the block. According to this, an effect similar to the effect mentioned above can be produced.

  (9) According to the present invention, in a video decoding apparatus that allows inter-coding when decoding in units of macroblocks, when decoding an inter-coded block, a decrease in coding performance in the processing block is estimated. Prediction vector control means (for example, equivalent to the prediction vector control unit 132 in FIG. 5) and prediction vector generation means (for example, FIG. 5) for generating a prediction vector in the processing block based on the estimation by the prediction vector control means. The third prediction vector generation unit 133), and a video decoding device is proposed.

  According to the present invention, a moving picture decoding apparatus that allows inter-coding when decoding in units of macroblocks, and that predicts a decrease in coding performance in a processing block when decoding an inter-coded block. Vector control means and prediction vector generation means for generating a prediction vector in the processing block based on estimation by the prediction vector control means are provided. For this reason, the prediction vector performance can be improved and the coding efficiency can be improved while suppressing the amount of additional information for various patterns.

  (10) In the moving picture decoding apparatus according to (9), the prediction vector control unit is configured to perform encoding performance in the processing block based on information on a motion vector in an encoded block adjacent to the processing block. The prediction vector generation method is determined based on the prediction vector generation method determined by the prediction vector control unit, and the prediction vector generation method determines the prediction vector generation method. A moving picture decoding apparatus is proposed in which an optimal vector is selected as the prediction vector from a motion vector and is used as the prediction vector.

  According to the present invention, the prediction vector control means determines a prediction vector generation method by estimating a decrease in the encoding performance in the processing block based on the information on the motion vector in the encoded block adjacent to the processing block. It was decided. Further, the prediction vector generation means selects an optimal vector as a prediction vector from the motion vectors in the encoded block adjacent to the processing block based on the prediction vector generation method determined by the prediction vector control means, It was decided to. According to this, an effect similar to the effect mentioned above can be produced.

  (11) The present invention relates to the moving picture decoding apparatus according to (9) or (10), wherein the prediction vector control means is based on a component or magnitude of a motion vector in an encoded block adjacent to the processing block. A moving picture decoding apparatus has been proposed in which a decrease in encoding performance in the processing block is estimated.

  According to the present invention, the prediction vector control means estimates a decrease in coding performance in the processing block based on the motion vector component or magnitude in the coded block adjacent to the processing block. According to this, an effect similar to the effect mentioned above can be produced.

  (12) In the moving image decoding apparatus according to (9) or (10), the present invention provides the prediction vector control means, based on a component or magnitude of a difference vector in an encoded block adjacent to the processing block, A moving picture decoding apparatus has been proposed in which a decrease in encoding performance in the processing block is estimated.

  According to the present invention, the prediction vector control means estimates a decrease in encoding performance in the processing block based on the component or magnitude of the difference vector in the encoded block adjacent to the processing block. According to this, an effect similar to the effect mentioned above can be produced.

  (13) In the moving image decoding device according to any one of (9) to (12), the prediction vector control unit may include a motion vector component or magnitude in a coded block adjacent to the processing block, Alternatively, a median of a difference vector component or magnitude in an encoded block adjacent to the processing block is calculated, and the calculated median and a motion vector component or magnitude in the encoded block, or the code A difference between the component or the size of the difference vector in the converted block, an evaluation unit that evaluates the encoded block having the maximum calculated difference, and the evaluation in the processing block based on the evaluation by the evaluation unit A video decoding device comprising: an estimation unit that estimates a decrease in encoding performance. It has proposed a location.

  According to the present invention, the prediction vector control means is provided with the evaluation unit and the estimation unit. Then, the evaluation unit calculates the median of the motion vector component or magnitude in the encoded block adjacent to the processing block, or the difference vector component or magnitude in the encoded block adjacent to the processing block. The difference between the calculated median and the motion vector component or magnitude in the encoded block or the difference vector component or magnitude in the encoded block is calculated, and the encoded block having the maximum difference is calculated. We decided to evaluate. In addition, the estimation unit estimates a decrease in coding performance in the processing block based on the evaluation by the evaluation unit. According to this, an effect similar to the effect mentioned above can be produced.

  (14) In the moving image decoding device according to any one of (9) to (12), the prediction vector control unit may include a motion vector component or magnitude in a coded block adjacent to the processing block, Alternatively, the average of the difference vector component or magnitude in the encoded block adjacent to the processing block is calculated, and the calculated average and the motion vector component or magnitude in the encoded block, or the code A difference between the component or the size of the difference vector in the converted block, an evaluation unit that evaluates the encoded block having the maximum calculated difference, and the evaluation in the processing block based on the evaluation by the evaluation unit Proposing a video decoding device comprising: an estimation unit that estimates a decrease in encoding performance There.

  According to the present invention, the prediction vector control means is provided with the evaluation unit and the estimation unit. Then, the evaluation unit calculates and calculates the motion vector component or magnitude in the encoded block adjacent to the processing block, or the average of the difference vector component or magnitude in the encoded block adjacent to the processing block. The difference between the average and the motion vector component or magnitude in the encoded block or the difference vector component or magnitude in the encoded block is calculated, and the encoded block having the maximum difference is calculated. We decided to evaluate. In addition, the estimation unit estimates a decrease in coding performance in the processing block based on the evaluation by the evaluation unit. According to this, an effect similar to the effect mentioned above can be produced.

  (15) In the video decoding device according to (13), the prediction vector generation unit may be configured such that when the estimation unit estimates that there is a concern about a decrease in encoding performance in the processing block, the processing block A moving picture decoding apparatus is proposed in which a motion vector in an encoded block adjacent to is used as the prediction vector.

  According to the present invention, when the prediction vector generation unit estimates that there is a concern about a decrease in encoding performance in the processing block by the estimation unit, the motion vector in the encoded block adjacent to the processing block is used as the prediction vector. It was decided to. According to this, an effect similar to the effect mentioned above can be produced.

  (16) The present invention provides a video decoding device that allows inter-coding when the bit stream encoded by the video encoding device of (8) is decoded in units of macroblocks. When decoding a block, prediction vector control means (for example, equivalent to the prediction vector control unit 132 in FIG. 5) for estimating a decrease in coding performance in the processing block and prediction based on the prediction by the prediction vector control means Prediction vector generation means (for example, equivalent to the third prediction vector generation unit 133 in FIG. 5) for generating a vector, and the prediction vector generation means performs coding performance in the processing block by the prediction vector control means. If it is estimated that there is a concern about a decrease in the data rate, the prediction vector is based on the control information given to the bitstream. It proposes a video decoding apparatus characterized by determining a torque.

  According to the present invention, a moving picture decoding apparatus that allows inter coding when a bit stream coded by a moving picture coding apparatus is decoded in units of macroblocks is processed when decoding an inter coded block. Prediction vector control means for estimating a decrease in coding performance in the block and prediction vector generation means for generating a prediction vector based on the estimation by the prediction vector control means are provided. When the prediction vector generation unit estimates that the prediction vector control unit has a concern about a decrease in coding performance in the processing block, the prediction vector generation unit determines a prediction vector based on the control information given to the bitstream; did. According to this, an effect similar to the effect mentioned above can be produced.

  ADVANTAGE OF THE INVENTION According to this invention, the fall of the prediction vector performance in the edge vicinity can be detected, and control information can be efficiently provided by providing the control information which shows an appropriate prediction vector only when a fall is detected. For this reason, it is possible to solve the problem caused by the edge (particularly the moving object boundary) and to suppress the amount of additional information necessary for solving the problem. As a result, motion vector prediction performance in the vicinity of the edge can be improved, and coding efficiency can be improved.

It is a block diagram which shows the structure of the moving image encoder which concerns on one Embodiment of this invention. It is a block diagram which shows the structure of the inter coding prediction value production | generation part with which the said moving image encoder is provided. It is a block diagram which shows the structure of the 1st prediction vector production | generation part 23 with which the said moving image encoder is provided. It is a block diagram which shows the structure of the moving image decoding apparatus which concerns on one Embodiment of this invention. It is a block diagram which shows the structure of the inter estimated value production | generation part with which the said moving image decoding apparatus is provided. It is a figure for demonstrating the edge detection process by the said moving image encoder and the said moving image decoder. It is a figure for demonstrating the prediction vector production | generation method which concerns on a prior art example. It is a figure for demonstrating the prediction vector production | generation method which concerns on a prior art example.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. Note that the constituent elements in the following embodiments can be appropriately replaced with existing constituent elements and the like, and various variations including combinations with other existing constituent elements are possible. Accordingly, the description of the following embodiments does not limit the contents of the invention described in the claims.

<Configuration of Video Encoding Device AA>
FIG. 1 is a block diagram showing a configuration of a moving picture coding apparatus AA according to an embodiment of the present invention. The moving image encoding device AA allows inter encoding when encoding in units of MB in the screen, determines a motion vector by motion compensation, and converts it into a motion vector of an encoded block adjacent to the processing block. Based on this, a prediction vector is generated, and a difference between the motion vector and the prediction vector is encoded. This moving image encoding apparatus AA includes an intra encoded prediction value generation unit 1, an inter encoded prediction value generation unit 2, a mode determination control unit 3, a DCT / quantization unit 4, an IDCT / inverse quantization unit 5, an entropy code. , A first local memory 7 and a second local memory 8.

  The intra-coded prediction value generation unit 1 includes the input image a, the locally decoded pixel value c in the encoded block, and prediction information (motion vector and invention) in the encoded block adjacent to the processing block in the same frame. D) including information indicating whether the method is applied or not. The intra encoded prediction value generation unit 1 generates a prediction value in each intra prediction direction in the processing block based on the locally decoded pixel value c in the encoded block, and an input signal is generated for the generated prediction value. The encoding distortion in intra coding is calculated by calculating | requiring the difference with these. Then, based on the calculated encoding distortion and the prediction information d in the encoded block, the cost value in each intra prediction direction is calculated, and the calculated cost value is compared, which is most suitable for encoding the processing block. Select the intra prediction direction. And the prediction value e in the optimal intra prediction direction, the information f regarding the optimal intra prediction direction, and the cost value g in the optimal intra prediction direction are output.

  The inter-coded prediction value generation unit 2 inputs the input video a, the locally decoded pixel value c in the encoded block, and the prediction information d in the encoded block adjacent to the processing block in the same frame. And a predicted value h in the optimal motion vector, optimal motion vector prediction information (including control information i1 and difference vector i2 described later) i, and cost value j in the optimal motion vector prediction information are output. To do. Details of the inter-coded prediction value generation unit 2 will be described later with reference to FIG.

  The mode determination control unit 3 uses the cost value g in the optimal intra prediction direction output from the intra encoded prediction value generation unit 1 and the optimal motion vector prediction information output from the inter encoded prediction value generation unit 2. The cost value j and the input cost values g and j are compared, and an encoding mode suitable for the processing block is selected. By selecting the coding mode, the prediction value e in the optimal intra prediction direction output from the intra-coded prediction value generation unit 1 and the prediction value in the optimal motion vector output from the inter-coding prediction value generation unit 2 Of h, the prediction value used for encoding is switched. Also, by selecting this encoding mode, the information f used for the optimal intra prediction direction output from the intra encoded prediction value generation unit 1 and the optimal motion vector prediction information i are used for encoding. Information will be switched.

  The DCT / quantization unit 4 receives the difference between the predicted value e or the predicted value h for the input video a, performs DCT processing and quantization processing on the input signal, and quantized DCT coefficient k Is output.

  The IDCT / inverse quantization unit 5 receives the quantized DCT coefficient k, performs inverse quantization processing and inverse DCT processing on the input signal, and outputs a pixel signal m subjected to inverse DCT.

  The entropy encoding unit 6 receives the quantized DCT coefficient k and the prediction information f or the prediction information i, performs entropy encoding on the input signal, and outputs the encoded data b.

  The first local memory 7 receives a signal obtained by summing the predicted value e or the predicted value h and the inverse DCT pixel signal m, that is, the locally decoded pixel value c in the encoded block. . Then, the locally decoded pixel value c in the encoded block is accumulated and appropriately supplied to the intra-coded prediction value generation unit 1 and the inter-coding prediction value generation unit 2.

  The second local memory 8 receives the prediction information f or the prediction information i, that is, the prediction information d in the encoded block. Then, the prediction information d in the encoded block is accumulated and appropriately supplied to the intra-coded prediction value generation unit 1 and the inter-coding prediction value generation unit 2. Note that the prediction information d in the encoded block includes information indicating whether or not the method of the present invention is applied.

<Configuration of Inter-coded Predicted Value Generation Unit 2>
FIG. 2 is a block diagram illustrating a configuration of the inter-coded prediction value generation unit 2. The inter-coded prediction value generation unit 2 includes an MV (Motion Vector) extraction unit 21, a prediction vector control unit 22, a first prediction vector generation unit 23, a second prediction vector generation unit 24, a reference destination determination unit 25, a vector A difference extraction unit 26, a distortion calculation unit 27, and a cost value calculation unit 28 are provided.

  The MV extraction unit 21 receives the prediction information (difference vector) d in the encoded block. The MV extraction unit 21 extracts information on the motion vector n in the encoded block adjacent to the processing block from the input prediction information d, and outputs the motion vector n in the encoded block adjacent to the processing block. .

  The prediction vector control unit 22 receives the prediction information d in the encoded block. And based on the motion vector contained in the input prediction information d, the deviation between each block is evaluated, the applicability of the invention method in a processing block is judged, and the control signal p according to the judgment is output. Based on this determination, a prediction vector generation method is selected.

  The first prediction vector generation unit 23 selects an optimal vector as a prediction vector from the motion vectors n in N encoded blocks adjacent to the processing block, and outputs the selected vector as a prediction vector q. Control information i1 representing the prediction vector q is output. Details of the first prediction vector generation unit 23 will be described later with reference to FIG.

  The second prediction vector generation unit 24 obtains the median of the x component and the y component for the motion vector n in the encoded block adjacent to the processing block, respectively, and obtains the vector composed of the obtained x component and y component as the prediction vector. Output as r.

  The reference destination determination unit 25 receives the input video a and the locally decoded pixel value c in the encoded block. The reference destination determination unit 25 searches the pixel value c of the input image a in the processing block for the pixel value c that has been locally decoded and searches for the position of the same rectangle that is closest to the pixel signal within the search range. For a close position, a spatial relative position from the processing block is set as a motion vector s. In addition, regarding the locally decoded pixel value at the position where the pixel signal is closest, the pixel value is set as a predicted value h. Furthermore, the difference of the predicted value h with respect to the pixel value of the input video a is set as the prediction error t. Then, the motion vector s, the predicted value h, and the prediction error t are output.

  The vector difference extraction unit 26 receives the motion vector s and the prediction vector q or the prediction vector r selected according to the control signal p. The vector difference extraction unit 26 sets a difference between the prediction vector q or the prediction vector r with respect to the motion vector s as a difference vector i2. Further, entropy coding is performed on the difference vector i2, and the code amount u generated by the difference vector i2 is obtained. Then, a difference vector (information regarding prediction used when decoding) i2 and a code amount u required for encoding the difference vector i2 are output.

  The distortion calculation unit 27 receives the input video a, the prediction error t, and the prediction value h. The distortion calculation unit 27 performs DCT, quantization, inverse quantization, and inverse DCT processing on the input prediction error t, and a signal obtained by adding a prediction value h to the result of these processing (local (Decoded signal) is calculated. Here, a difference between the local decoded signal and the input signal is obtained, a sum of squares is calculated for the difference signal, and a sum of squares (encoding distortion) v of the difference signal is output.

  The cost value calculation unit 28 receives the code amount u and the coding distortion v of the difference vector as inputs. The cost value calculation unit 28 inputs the code amount u and the coding distortion v into a cost function, calculates a cost value j, and outputs it.

<Configuration of First Prediction Vector Generation Unit 23>
FIG. 3 is a block diagram illustrating a configuration of the first prediction vector generation unit 23. The first prediction vector generation unit 23 includes a candidate MV extraction unit 231 and an optimal prediction MV determination unit 232.

  The candidate MV extraction unit 231 includes N unit candidate MV extraction units from the first unit candidate MV extraction unit B1 to the Nth unit candidate MV extraction unit BN. The candidate MV extraction unit 231 selects N candidate motion vectors from the motion vectors n in the encoded block adjacent to the processing block.

  The optimal prediction MV determination unit 232 evaluates the cost value indicating the encoding efficiency for the N motion vectors selected by the candidate MV extraction unit 231 based on the error and the code amount generated by the encoding. Then, the cost values of the N motion vectors are compared, the most efficient motion vector is set as the prediction vector q, and the prediction vector q and control information i1 representing the prediction vector q are output.

<Configuration of Moving Picture Decoding Device BB>
FIG. 4 is a block diagram showing a configuration of the video decoding device BB according to one embodiment of the present invention. The video decoding device BB allows extended intra coding when coding in units of MB in the screen. This moving picture decoding apparatus BB includes an encoded data analysis unit 101, an intra prediction value generation unit 102, an inter prediction value generation unit 103, a prediction technique control unit 104, and a memory 105.

  The encoded data analysis unit 101 receives the encoded data b. The input encoded data b includes information indicating whether the inventive method is applied. The encoded data analysis unit 101 includes an entropy decoding unit (not shown). The entropy decoding unit performs entropy decoding on the encoded data b, and analyzes the contents described in the encoded data b according to the syntax. Further, in each block, the presence / absence of an edge is evaluated based on the same processing as that of the prediction vector control unit 22, and information indicating the presence / absence of application of the inventive method is acquired from the encoded data of the block including the edge. Then, the residual signal A obtained as a result of the syntax analysis and the prediction information B in the encoded block adjacent to the processing block are output.

  The intra prediction value generation unit 102 receives the prediction information B obtained from the above-described entropy decoding unit included in the encoded data analysis unit 101 and the decoded pixel value C obtained from the memory 105. The intra prediction value generation unit 102 generates and outputs an intra prediction value D according to the prediction information B based on the decoded pixel value C.

  The inter prediction value generation unit 103 receives the decoded pixel value C and the prediction information B in adjacent decoded blocks and processing blocks in the same frame, and outputs a prediction value E in inter coding. Details of the inter prediction value generation unit 103 will be described later with reference to FIG.

  The prediction technique control unit 104 receives the prediction information B obtained from the above-described entropy decoding unit included in the encoded data analysis unit 101. The prediction technique control unit 104 identifies whether the prediction information B relates to intra prediction or inter prediction, and outputs a control signal F for switching between intra prediction and inter prediction.

  The memory 105 receives a signal obtained by summing the intra prediction value D or the inter prediction value E and the residual signal A, that is, the decoded pixel value C. The memory 105 accumulates the decoded pixel value C, and supplies the decoded pixel value C to the intra prediction value generation unit 102 and the inter prediction value generation unit 103 as appropriate when performing decoding processing on the undecoded block.

<Configuration of Inter Predicted Value Generation Unit 103>
FIG. 5 is a block diagram illustrating a configuration of the inter prediction value generation unit 103. The inter prediction value generation unit 103 includes an MV (Motion Vector) extraction unit 131, a prediction vector control unit 132, a third prediction vector generation unit 133, a fourth prediction vector generation unit 134, an MV generation unit 135, and a reference destination determination. Part 136 is provided.

  The MV extraction unit 131 receives the prediction information B (difference from the prediction vector with respect to the actual motion vector) in the encoded block supplied from the local memory. The MV extraction unit 131 extracts information on the motion vector G in the encoded block adjacent to the processing block from the input prediction information B, and outputs the motion vector G.

  The prediction vector control unit 132 receives the prediction information B in the encoded block supplied from the local memory. The prediction vector control unit 132 evaluates the divergence between the blocks based on the motion vector included in the input prediction information B, determines whether the invention method is applicable to the processing block, and controls the control signal according to the determination. H is output. Based on this determination, a prediction vector generation method is selected.

  The third prediction vector generation unit 133 receives the prediction information B in the encoded block supplied from the local memory and the motion vector G in the encoded block adjacent to the processing block. The third prediction vector generation unit 133 selects the prediction vector I of the encoded block adjacent to the processing block according to the method control information included in the prediction information B, and outputs this prediction vector I.

  The fourth prediction vector generation unit 134 receives as input the motion vector G in the encoded block adjacent to the processing block. The fourth prediction vector generation unit 134 obtains the median of the x component and the y component of the input motion vector G, and outputs a vector composed of the obtained x component and y component as the prediction vector J.

  The MV generation unit 135 receives the prediction information B in the encoded block supplied from the local memory and the prediction vector I or the prediction vector J as inputs. The MV generation unit 135 calculates the sum of components of the difference vector in the processing block included in the prediction information B and the prediction vector I or the prediction vector J, and processes the vector formed by the calculated components. Output as motion vector K in the block.

  The reference destination determination unit 136 receives the prediction information B in the encoded block supplied from the local memory, the pixel value C obtained by decoding the encoding result in the encoded block, and the motion vector K as inputs. Based on the motion vector K and information for identifying the reference block included in the prediction information B, the reference destination determination unit 136 specifies a pixel value to be a prediction value in the encoded block, and uses this pixel value as a prediction value. Output as the predicted value L in the processing block.

<Edge Detection Processing by Prediction Vector Control Unit 22 and Prediction Vector Control Unit 132>
Below, the edge detection process by the prediction vector control part 22 and the prediction vector control part 132 is demonstrated. In the edge detection process, an edge in a processing block is detected based on information on a motion vector in a coded block adjacent to the processing block.

  Specifically, the edge detection in the processing block is performed based on the motion vector divergence between the encoded blocks adjacent to the processing block. As this deviation determination method, there are a determination method based on a deviation from the median and a determination method based on a deviation from the average.

<Judgment method by deviation from median>
Below, the determination method by the deviation from a median is demonstrated. Here, a function that takes three numeric values or vectors as inputs and outputs the most deviated numeric value or vector among these inputs based on the median of these inputs is defined as a function minor_med, as shown in Equation (7) To express.

  In Expression (7), α, β, and γ are numerical values or vectors. The function median is a function that receives three numerical values or vectors as inputs and outputs these median values or median vectors. Further, | · | is a function that outputs an absolute value when α, β, and γ are numerical values, and a function that outputs a distance when α, β, and γ are vectors.

<Judgment method by deviation from average>
Hereinafter, a determination method based on deviation from the average will be described. Here, a function that takes three numerical values or vectors as inputs and outputs the most deviated numerical value or vector among these inputs based on the average of these inputs is defined as a function minor_avg, as shown in Expression (8). To express.

  In Expression (8), α, β, and γ are numerical values or vectors. The function average is a function that receives three numerical values or vectors as inputs and outputs an average value or average vector thereof. Further, | · | is a function that outputs an absolute value when α, β, and γ are numerical values, and a function that outputs a distance when α, β, and γ are vectors.

<Edge detection method using function minor>
Hereinafter, using the motion vector information in the four encoded blocks adjacent to the processing block will be considered with reference to FIG. In FIG. 6, B _PRED indicates a processing block, and B _A , B _B , B _C , and B _D indicate four encoded blocks adjacent to the processing block B _PRED .

Here, when the information about the motion vectors of the encoded blocks B _{A to} B _D are M _{A to} M _D , the above-described function minor_med or the above-mentioned function minor_avg (hereinafter collectively referred to as “function minor”). Is expressed as in Expression (9), it is determined that a horizontal edge exists. On the other hand, when the function minor is expressed as in Expression (10), it is determined that a vertical edge exists.

When determining from motion vectors in the encoded blocks B _{A to} B _D , there are a method of determining using the motion vector itself and a method of determining using the magnitude of the motion vector.

In the determination method using the motion vector itself, the motion vector in each of the encoded blocks B _{A to} B _D is used as the information M _{A to} M _D regarding each motion vector of the encoded blocks B _{A to} B _D described above. Is used.

The method for determining by using the magnitude of the motion vector, as the information _M A ~M _D for each of the motion vectors of coded blocks _B A .about.B _D above, in each of the encoded block _B A .about.B _D Use the magnitude of the motion vector. The magnitude of the vector is a value obtained by taking the square root of the sum of squares of the x component and the y component of this vector.

On the other hand, when determining from the difference vectors in the encoded blocks B _{A to} B _D , there are a method of determining using the difference vector itself and a method of determining using the magnitude of the difference vector. Here, the difference vector in the encoded block is a vector of the difference between the motion vector and the prediction vector in each block.

In the determination method using the difference vector itself, the difference vector in each of the encoded blocks B _{A to} B _D is used as the information M _{A to} M _D regarding the motion vectors of the encoded blocks B _{A to} B _D described above. Is used.

In the method of determining using the magnitude of the difference vector, information M _{A to} M _D regarding each motion vector of the encoded blocks B _{A to} B _D described above is used as each of the encoded blocks B _{A to} B _D. Use the magnitude of the difference vector. The magnitude of the vector is a value obtained by taking the square root of the sum of squares of the x component and the y component of this vector.

  According to the above moving image encoding device AA and moving image decoding device BB, it is possible to detect a decrease in predicted vector performance in the vicinity of an edge, and only when a decrease is detected, control information indicating an appropriate prediction vector is added. The control information can be efficiently given. For this reason, it is possible to solve problems caused by edges (particularly moving object boundaries) and to suppress the amount of additional information necessary for solving the problems. As a result, motion vector prediction performance in the vicinity of the edge can be improved, and coding efficiency can be improved.

In addition, when the above-described moving image encoding device AA and moving image decoding device BB are applied to an encoder and an encoding experiment is performed, the average code amount can be reduced by about 0.57% compared to the conventional case. Here, in the above-described coding experiments, the edge detection in the processing block, with using a determination method according to deviation from the above median, information M _A for each of the motion vectors of coded blocks B _A .about.B _D above as ~M _D, using motion vectors of coded blocks _B a .about.B _D.

  Note that the processing of the moving image encoding device AA and the processing of the moving image decoding device BB of the present invention are stored in a computer-readable recording medium, and the program recorded on the recording medium is stored in the moving image encoding device AA or the moving image decoding. The present invention can be realized by reading the apparatus BB and executing it.

  The above-described program is transferred from the moving image encoding device AA or the moving image decoding device BB storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. May be transmitted. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line.

  Further, the above-described program may be for realizing a part of the above-described function. Furthermore, what can implement | achieve the above-mentioned function in combination with the program already recorded on the moving image encoder AA and the moving image decoder BB, what is called a difference file (difference program) may be sufficient.

  The embodiments of the present invention have been described in detail with reference to the drawings. However, the specific configuration is not limited to the embodiments, and includes designs and the like that do not depart from the gist of the present invention.

2 ... Inter-coded prediction value generation unit 22, 132 ... Prediction vector control unit 23 ... First prediction vector generation unit 103 ... Inter prediction value generation unit 133 ... Third prediction vector Generation unit 231 ... Candidate MV extraction unit 232 ... Optimal prediction MV determination unit AA ... Video encoding device BB ... Video decoding device

Claims (16)

In a video encoding apparatus that allows inter encoding when encoding in units of macroblocks,
Prediction vector control means for estimating a decrease in coding performance in the processing block during inter coding;
Prediction vector generation means for generating a prediction vector in the processing block based on the prediction by the prediction vector control means;
A moving picture encoding apparatus comprising: The moving image encoding device according to claim 1,
The prediction vector control means estimates a reduction in encoding performance in the processing block based on information on a motion vector in an encoded block adjacent to the processing block, determines a prediction vector generation method,
The prediction vector generation unit selects an optimal vector as the prediction vector from motion vectors in an encoded block adjacent to the processing block based on a prediction vector generation method determined by the prediction vector control unit, A moving picture coding apparatus characterized by using a prediction vector. The moving picture encoding apparatus according to claim 1 or 2,
The prediction vector control means estimates a decrease in encoding performance in the processing block based on a motion vector component or magnitude in an encoded block adjacent to the processing block. apparatus. The moving picture encoding apparatus according to claim 1 or 2,
The prediction vector control means estimates a decrease in encoding performance in the processing block based on a component or magnitude of a difference vector in an encoded block adjacent to the processing block. apparatus. The moving image encoding device according to any one of claims 1 to 4,
The prediction vector control means includes
Calculate the median of the motion vector component or magnitude in the encoded block adjacent to the processing block, or the difference vector component or magnitude in the encoded block adjacent to the processing block, and the calculated median The difference between the motion vector component or magnitude in the coded block or the difference vector component or magnitude in the coded block is calculated, and the coded block having the maximum calculated difference is calculated. An evaluation unit to evaluate;
Based on the evaluation by the evaluation unit, an estimation unit that estimates a decrease in encoding performance in the processing block;
A moving picture encoding apparatus comprising: The moving image encoding device according to any one of claims 1 to 4,
The prediction vector control means includes
The motion vector component or magnitude in the encoded block adjacent to the processing block, or the average of the difference vector component or magnitude in the encoded block adjacent to the processing block is calculated, and the calculated average The difference between the motion vector component or magnitude in the coded block or the difference vector component or magnitude in the coded block is calculated, and the coded block having the maximum calculated difference is calculated. An evaluation unit to evaluate;
Based on the evaluation by the evaluation unit, an estimation unit that estimates a decrease in encoding performance in the processing block;
A moving picture encoding apparatus comprising: The moving image encoding device according to claim 5,
When the estimation unit estimates that there is a concern about a decrease in encoding performance in the processing block, the prediction vector generation unit sets a motion vector in an encoded block adjacent to the processing block as the prediction vector. A video encoding apparatus characterized by the above. The moving image encoding device according to claim 5,
The prediction vector generation means, when the estimation unit estimates that there is a concern about a decrease in encoding performance in the processing block, provides control information for identifying a motion vector to be used as the prediction vector. A moving picture coding apparatus, characterized in that the moving picture coding apparatus assigns the bit stream as encoded data. In a video decoding device that allows inter-coding when decoding in units of macroblocks,
Prediction vector control means for estimating a decrease in coding performance in a processing block when decoding an inter-coded block;
Prediction vector generation means for generating a prediction vector in the processing block based on the prediction by the prediction vector control means;
A moving picture decoding apparatus comprising: The moving picture decoding apparatus according to claim 9, wherein
The prediction vector control means estimates a reduction in encoding performance in the processing block based on information on a motion vector in an encoded block adjacent to the processing block, determines a prediction vector generation method,
The prediction vector generation unit selects an optimal vector as the prediction vector from motion vectors in an encoded block adjacent to the processing block based on a prediction vector generation method determined by the prediction vector control unit, A moving picture decoding apparatus characterized in that a prediction vector is used. The moving picture decoding apparatus according to claim 9 or 10,
The motion vector decoding apparatus, wherein the prediction vector control means estimates a decrease in coding performance in the processing block based on a motion vector component or magnitude in a coded block adjacent to the processing block . The moving picture decoding apparatus according to claim 9 or 10,
The video decoding device, wherein the prediction vector control means estimates a decrease in coding performance in the processing block based on a component or magnitude of a difference vector in a coded block adjacent to the processing block . The video decoding device according to any one of claims 9 to 12,
The prediction vector control means includes
Calculate the median of the motion vector component or magnitude in the encoded block adjacent to the processing block, or the difference vector component or magnitude in the encoded block adjacent to the processing block, and the calculated median The difference between the motion vector component or magnitude in the coded block or the difference vector component or magnitude in the coded block is calculated, and the coded block having the maximum calculated difference is calculated. An evaluation unit to evaluate;
Based on the evaluation by the evaluation unit, an estimation unit that estimates a decrease in encoding performance in the processing block;
A moving picture decoding apparatus comprising: The video decoding device according to any one of claims 9 to 12,
The prediction vector control means includes
The motion vector component or magnitude in the encoded block adjacent to the processing block, or the average of the difference vector component or magnitude in the encoded block adjacent to the processing block is calculated, and the calculated average The difference between the motion vector component or magnitude in the coded block or the difference vector component or magnitude in the coded block is calculated, and the coded block having the maximum calculated difference is calculated. An evaluation unit to evaluate;
Based on the evaluation by the evaluation unit, an estimation unit that estimates a decrease in encoding performance in the processing block;
A moving picture decoding apparatus comprising: The moving picture decoding apparatus according to claim 13, wherein
When the estimation unit estimates that there is a concern about a decrease in encoding performance in the processing block, the prediction vector generation unit sets a motion vector in an encoded block adjacent to the processing block as the prediction vector. A video decoding apparatus characterized by the above. In the moving picture decoding apparatus that allows inter coding when decoding the bitstream encoded by the moving picture encoding apparatus according to claim 8 in units of macroblocks,
Prediction vector control means for estimating a decrease in coding performance in a processing block when decoding an inter-coded block;
Prediction vector generation means for generating a prediction vector based on the prediction by the prediction vector control means,
The prediction vector generation unit determines the prediction vector based on control information given to the bitstream when the prediction vector control unit estimates that there is a concern about a decrease in encoding performance in the processing block. A moving picture decoding apparatus characterized by:

Images