JP2014068361A - Moving image encoder, moving image decoder, moving image encoding method, moving image decoding method, moving image encoding program, and moving image decoding program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a moving image encoder and a moving image decoder, which can improve efficiency of information compression of encoding and decoding.SOLUTION: A moving image encoder 10 includes: a motion vector prediction section for predicting an optimum prediction motion vector after correction by performing correction for scaling motion vectors 751a, 751b and 751c of adjacent blocks with a target reference frame image 702 as reference and by determining the optimum prediction motion vector in response to the motion vector of the adjacent block in response to adjacent reference frame images 703a, 703b and 703c which are referred to for detecting the motion vector of the adjacent block adjacent to the block of an encoding object, a target reference frame image 702 which is referred to for detecting the motion vector of the target block, and a temporal relation with a target frame image 701 being a frame image of the encoding object or time information of the images.

Description

  The present invention relates to a moving image encoding device, a moving image decoding device, a moving image encoding method, a moving image decoding method, a moving image encoding program, and a moving image decoding program.

  As an example of a conventional moving image encoding method, H.264 has been described. Examples thereof include a moving image encoding device and a moving image decoding device based on the H.264 / AVC encoding method (see Non-Patent Document 1). This method reduces the redundancy existing in the temporal direction by motion compensation inter-frame prediction, and further reduces the redundancy remaining in the spatial direction by orthogonal transformation, thereby compressing information of the moving image (input video signal). It is.

  In motion compensation inter-frame prediction (hereinafter referred to as “INTER prediction mode”) in the above method, a plurality of reference frame images for detecting a motion vector can be prepared. The motion vectors may be motion compensated using different reference frame images.

  In addition, when calculating the prediction motion vector of the encoding target region, regardless of which reference frame image the motion vector of the surrounding encoded region has been motion compensated, the values of those motion vectors are compared, These intermediate values are used as motion vector prediction values for the region to be encoded.

Joint Video Team (JVT) of ISO / IEC MPEG and ITU-VCEG, "Editor's Proposed Draft Text Modifications for Joint Video Specification (ITU-TRec.H.264 | ISO / IEC 14496-10 AVC), Geneva modifications draft 37"

  However, when the surrounding encoded region is motion-compensated using a reference frame image different from the reference frame image of the encoding target region as described above, the difference between the reference frames is obtained by using the intermediate value as a predicted motion vector. Therefore, there is a problem that the efficiency of encoding information compression is reduced due to a large deviation from the actual motion vector. Along with this, there is a problem that the efficiency of information compression for decoding also decreases at the same time.

  The present invention has been made to solve the above-described problem, and is a moving image encoding device, a moving image decoding device, and a moving image capable of improving the efficiency of encoding and decoding information compression. It is an object to provide an encoding method, a moving image decoding method, a moving image encoding program, and a moving image decoding program.

  The moving image encoding apparatus according to the present invention divides a frame image to be encoded in a moving image signal composed of a time sequence of frame image signals into a plurality of target regions, and the frame image to be encoded for each target region. In a moving picture coding apparatus that performs motion compensation coding by referring to a plurality of different frame images, the motion vector is referenced to detect a motion vector of an adjacent area adjacent to the target area. Based on the temporal relationship between the adjacent reference frame image, the target reference frame image referred to for detecting the motion vector of the target region, and the target frame image that is the frame image to be encoded, or their time information , Correction for scaling the motion vector of the adjacent area adjacent to the target area based on the target reference frame image, and adjacent to the target area Characterized by having a motion vector prediction means for predicting the optimal predicted motion vector after correction by performing the determination of the optimum predicted motion vector based on the motion vector of the bearing region.

  The moving image encoding method of the present invention divides a frame image to be encoded in a moving image signal composed of a time sequence of frame image signals into a plurality of target regions, and for each target region, In a moving image coding method for performing motion compensation coding by detecting a motion vector with reference to a plurality of frame images different from a frame image, the motion vector predicting means includes a motion vector of an adjacent region adjacent to the target region. Temporal relationship between adjacent reference frame images that are referenced to detect motion, target reference frame images that are referenced to detect motion vectors of the target region, and target frame images that are encoding target frame images Or, based on the time information, the motion vector of the adjacent area adjacent to the target area is scaled based on the target reference frame image A correction that is characterized by having a motion vector prediction step of predicting the optimal predicted motion vector after correction by performing the determination of the optimum predicted motion vector based on the motion vector of the adjacent region adjacent to the target area.

  Further, the moving image encoding program of the present invention divides a frame image to be encoded in a moving image signal composed of a time sequence of frame image signals into a plurality of target regions, and for each target region, In a moving image encoding program that causes a computer to execute an encoding process based on motion compensation by detecting a motion vector with reference to a plurality of frame images different from the frame image, the computer moves the motion of the adjacent region adjacent to the target region. Temporal comparison between an adjacent reference frame image referred to for detecting a vector, a target reference frame image referred to for detecting a motion vector of the target region, and a target frame image that is a frame image to be encoded An adjacent area adjacent to the target area based on the target reference frame image based on the relationship or their time information It is possible to function as a motion vector predictor that predicts the corrected optimal prediction motion vector by performing correction for scaling the motion vector and determining an optimal prediction motion vector based on the motion vector of the adjacent region adjacent to the target region. Features.

  According to the moving image encoding apparatus, the moving image encoding method, and the moving image encoding program of the present invention, the frame referred to by the motion vector predicting unit to detect the motion vector of the target region from the motion vector of the adjacent region. Scaling and correction is performed based on the time difference between the image and the frame image to be encoded, and the optimal prediction motion vector is predicted based on the motion vector of the adjacent region, so prediction is performed in consideration of temporal motion continuity. By determining the motion vector, the difference between the actual motion vector of the target region and the predicted motion vector can be further reduced.

  In the video encoding device of the present invention, the motion vector predicting means includes the target reference frame image based on the temporal relationship between the adjacent reference frame image, the target reference frame image, and the target frame image, or their time information. It is also preferable to scale and correct each of the motion vectors in the adjacent area based on the above, and determine the optimum predicted motion vector based on the corrected motion vector in the adjacent area. In this way, the motion vector predicting means scales each of the motion vectors of a plurality of adjacent regions on the basis of the time difference between the frame image referred to in order to detect the motion vector of the target region and the frame image to be encoded. After the correction, the optimum predicted motion vector is determined based on the corrected motion vector of the adjacent region, so that the difference between the actual motion vector of the target region and the predicted motion vector can be further reduced.

  In the video encoding device of the present invention, the motion vector prediction means determines the optimal prediction motion vector based on the motion vector of the adjacent region, and the adjacent reference of the motion vector of the adjacent region determined as the optimal prediction motion vector Preferably, based on the temporal relationship between the frame image, the target reference frame image, and the target frame image or their time information, the optimal predicted motion vector is scaled and corrected based on the target reference frame image. In this case, after the motion vector prediction means determines the optimal prediction motion vector based on the motion vector of the adjacent region, the determined optimal prediction motion vector is used as a frame image that is referred to in order to detect the motion vector of the target region. Since the time difference between the image and the frame image to be encoded is scaled and corrected, the difference between the motion vector of the actual target area and the predicted motion vector can be further reduced, and the process for predicting the motion vector Time can be shortened.

  The moving picture decoding apparatus of the present invention divides a frame image to be decoded in a moving picture signal composed of a time sequence of frame image signals into a plurality of target areas, In a video decoding device that performs decoding by motion compensation by using difference information between a motion vector detected with reference to a plurality of different frame images and a predicted motion vector, an adjacent region adjacent to the target region Temporal of an adjacent reference frame image referred to detect a motion vector, a target reference frame image referred to detect a motion vector of the target region, and a target frame image that is a frame image to be encoded The motion vector of the adjacent area adjacent to the target area based on the target reference frame image based on the current relationship or their time information. And having a correction to the ring, the motion vector prediction means for predicting the optimal predicted motion vector after correction by performing the determination of the optimum predicted motion vector based on the motion vector of the adjacent region adjacent to the target area.

  Also, the moving picture decoding method of the present invention divides a decoding target frame image in a moving picture signal composed of a time sequence of frame image signals into a plurality of target areas, and the decoding target frames for each target area. In a moving picture decoding method for performing decoding by motion compensation by using difference information between a motion vector detected with reference to a plurality of frame images different from an image and a predicted motion vector, the motion vector prediction means includes: An adjacent reference frame image referred to detect a motion vector of an adjacent region adjacent to the target region, a target reference frame image referred to detect a motion vector of the target region, and a frame image to be encoded Based on the temporal relationship with a certain target frame image or their time information, the target reference frame image is used as a reference to be adjacent to the target region. A motion vector prediction step for predicting the corrected optimal prediction motion vector by performing correction for scaling the motion vector of the adjacent region and determining an optimal prediction motion vector based on the motion vector of the adjacent region adjacent to the target region; It is characterized by that.

  Further, the moving picture decoding program of the present invention divides a decoding target frame image in a moving picture signal composed of a time sequence of frame image signals into a plurality of target areas, and the decoding target frame for each target area. In a moving image decoding program for causing a computer to execute a decoding process by motion compensation by using difference information between a motion vector detected with reference to a plurality of frame images different from the image and a predicted motion vector, the computer An adjacent reference frame image referred to for detecting a motion vector of an adjacent area adjacent to the target area, a target reference frame image referred to for detecting a motion vector of the target area, and a frame image to be encoded The target reference frame image based on the temporal relationship with the target frame image or the time information thereof The optimal predicted motion vector after correction is determined by performing a correction to scale the motion vector of the adjacent region adjacent to the target region with reference and determining an optimal predicted motion vector based on the motion vector of the adjacent region adjacent to the target region. It is made to function as a motion vector prediction means to predict.

  According to the moving picture decoding apparatus, the moving picture decoding method, and the moving picture decoding program of the present invention, the frame referred to by the motion vector predicting unit to detect the motion vector of the target area from the motion vector of the adjacent area. Scaling and correction is performed based on the time difference between the image and the frame image to be encoded, and the optimal prediction motion vector is predicted based on the motion vector of the adjacent region, so prediction is performed in consideration of temporal motion continuity. By determining the motion vector, the difference between the actual motion vector of the target region and the predicted motion vector can be further reduced.

  In the moving picture decoding apparatus according to the present invention, the motion vector predicting means includes the target reference frame image based on the temporal relationship between the adjacent reference frame image, the target reference frame image, and the target frame image or their time information. It is also preferable to scale and correct each of the motion vectors in the adjacent area based on the above, and determine the optimum predicted motion vector based on the corrected motion vector in the adjacent area. In this way, the motion vector predicting means scales each of the motion vectors of a plurality of adjacent regions on the basis of the time difference between the frame image referred to in order to detect the motion vector of the target region and the frame image to be encoded. After the correction, the optimum predicted motion vector is determined based on the corrected motion vector of the adjacent region, so that the difference between the actual motion vector of the target region and the predicted motion vector can be further reduced.

  In the moving picture decoding apparatus of the present invention, the motion vector prediction means determines an optimal prediction motion vector based on the motion vector of the adjacent region, and adjoins the motion vector of the adjacent region determined as the optimal prediction motion vector. It is also preferable that the optimal predicted motion vector is scaled and corrected based on the target reference frame image based on the temporal relationship between the frame image, the target reference frame image, and the target frame image or their time information. In this case, after the motion vector prediction means determines the optimal prediction motion vector based on the motion vector of the adjacent region, the determined optimal prediction motion vector is used as a frame image that is referred to in order to detect the motion vector of the target region. Since the time difference between the image and the frame image to be encoded is scaled and corrected, the difference between the motion vector of the actual target area and the predicted motion vector can be further reduced, and the process for predicting the motion vector Time can be shortened.

  According to the moving image encoding device and the moving image decoding device of the present invention, the motion vector predicting means includes the frame image referred to for detecting the motion vector of the target region as the motion vector of the adjacent region and the encoding target. Scaling and correcting based on the time difference from the frame image and predicting the optimal predicted motion vector based on the motion vector of the adjacent region, so determine the predicted motion vector in consideration of temporal motion continuity Thus, the difference between the motion vector of the actual target region and the predicted motion vector can be further reduced. Accordingly, it is possible to provide a moving image encoding device and a moving image decoding device that can improve the efficiency of encoding and decoding information compression.

It is the schematic which shows an example of the moving image encoder concerning this embodiment. It is a block diagram of the motion detection part shown in FIG. It is the schematic which shows an example of the moving image decoding apparatus concerning this embodiment. It is a block diagram of the motion vector decompression | restoration part shown in FIG. It is the figure which represented typically the block adjacent to the block of encoding object. It is a flowchart which shows operation | movement of the motion vector estimation part in 1st Embodiment. It is a figure which shows the motion vector of the block adjacent to the encoding object block on a time space. It is a flowchart which shows operation | movement of the motion vector estimation part in 2nd Embodiment. (A) is a figure which shows an example of the block divided | segmented for motion vector prediction, (b) is a figure which shows the other example of the block divided | segmented for motion vector prediction. It is a figure which shows the structure of the moving image encoding program concerning this embodiment. It is a figure which shows the structure of the moving image decoding program concerning this embodiment.

A video encoding device and a video decoding device according to an embodiment of the present invention will be described with reference to the drawings. In addition, in each figure, the same code | symbol is attached | subjected to the same element and the overlapping description is abbreviate | omitted.
[First Embodiment]

FIG. 1 is a schematic diagram illustrating an example of a video encoding device according to the present embodiment, and FIG. 3 is a schematic diagram illustrating an example of a video decoding device according to the present embodiment.
(Configuration of video encoding device)

  First, a moving picture encoding apparatus 10 according to the present invention will be described with reference to FIG. The moving picture encoding apparatus 10 to be described below is an H.264 standard. This is an encoding device compliant with the H.264 / AVC encoding system.

  Here, an input video signal (moving image signal) as a moving image signal input to the moving image encoding device 10 is composed of a time sequence of frame images. Further, the frame image signal represents a signal of the frame image unit of the input video signal. Hereinafter, the frame image signal to be encoded is referred to as a “current frame”. The current frame is divided into macroblocks that are square pixels of 16 pixels × 16 lines, and the following encoding process and decoding process are performed in units of macroblocks.

  H. In the H.264 / AVC encoding method, a motion vector is detected for each macroblock with reference to a plurality of encoded frame image signals (reference frame image signals) that are temporally different from the frame image signal as a prediction mode. A plurality of “INTER prediction modes” for performing motion compensation inter-frame prediction and a plurality of “INTRA prediction modes” for performing spatial prediction using pixel values of neighboring macroblocks that have been encoded in the same space are prepared. Has been. In the “INTER prediction mode”, motion detection, motion prediction, and motion compensation processes (details) are performed for each block (target region) obtained by further dividing the macroblock into arbitrary regions (for example, 8 pixels × 16 lines). Will be described later). The moving picture coding apparatus 10 is configured to perform efficient information compression by switching the prediction mode in units of macroblocks according to the local nature of the input video signal.

  As shown in FIG. 1, the moving image encoding apparatus 10 includes an input unit 101, a motion detection unit 102, a motion compensation unit 103, a frame memory 104, a spatial prediction unit 105, and a functional component as shown in FIG. A switch 106, a subtractor 107, an orthogonal transform unit 108, a quantization unit 109, a variable length coding unit 110, an inverse quantization unit 111, an inverse orthogonal transform unit 112, and an adder 113 are provided. Composed. Hereinafter, each component will be described.

  The input unit 101 receives an input video signal 121 as a moving image signal input from the outside, decomposes it into a frame image signal, and sends it to the subtracter 107 and the motion detection unit 102 as frame image signals 122 and 123. Part.

  The frame memory 104 is a part for storing previously encoded frame image signals.

  The motion detection unit 102 is a part that performs prediction mode selection and motion vector detection. More specifically, when “INTER prediction mode” is selected, the motion detection unit 102 uses a reference frame image signal 124 to select a predetermined frame image from a plurality of encoded frame images stored in advance in the frame memory. In this search range, an image signal pattern similar to the image signal pattern in the current frame is searched for. Then, a motion vector that is a spatial displacement amount between both image signal patterns is detected. The motion vector difference value, which is difference information between the detected motion vector and the optimally predicted motion vector (motion vector prediction value) calculated from the motion vector of the encoded adjacent block, and the reference frame image used to detect the motion vector A signal 125 including a reference frame number indicating the signal and the selected prediction mode is sent to the variable length coding unit 110. At the same time, the motion detection unit 102 sends a signal 126 including the selected prediction mode, motion vector, and reference frame number to the motion compensation unit 103.

  Also, the motion compensation unit 103 refers to the encoded image signal (reference frame image signal) of the frame indicated by the reference frame number in the frame memory 104 using the motion vector sent from the motion detection unit 102. The prediction image signal 127 of each block is generated and sent to the switch 106.

  On the other hand, when the “INTRA prediction mode” is selected, the motion detection unit 102 sends the selected prediction mode 128 to the spatial prediction unit 105. In this case, since the motion detection unit 102 performs spatial prediction using pixel values of neighboring blocks that have been encoded in the same space, the motion vector difference value and the reference frame number, which are information related to temporal motion, can be changed. It is not sent to the long encoding unit 110.

  On the other hand, the spatial prediction unit 105 generates a prediction image signal 130 with reference to the image signal (reference frame image signal 129) of the neighboring block that has been encoded, and sends the prediction image signal 130 to the switch 106.

  The switch 106 selects either the prediction image signal 127 or the prediction image signal 130 according to the prediction mode 131 received from the motion detection unit 102, and sends the selected prediction image signal 132 to the subtractor 107.

  On the other hand, the subtracter 107 generates a difference value (prediction residual signal 133) between the frame image signal 122 and the predicted image signal 132 and sends the difference value to the orthogonal transform unit 108.

  The orthogonal transform unit 108 generates an orthogonal transform coefficient 134 by performing orthogonal transform on the prediction residual signal 133 sent from the subtractor 107, and sends it to the quantization unit 109.

  On the other hand, the quantizing unit 109 quantizes the orthogonal transform coefficient 134 transmitted from the orthogonal transform unit 108 to generate a quantized orthogonal transform coefficient 135, and the variable length coding unit 110 and the inverse quantization Send to part 111.

  Next, the variable length coding unit 110 converts the quantized orthogonal transform coefficient 135 transmitted from the quantization unit 109, the prediction mode transmitted from the motion detection unit 102, the motion vector difference value, and the reference frame number. Based on this, entropy encoding is performed, multiplexed into the compressed stream 136, and transmitted to the outside.

  In addition, the inverse quantization unit 111 generates an orthogonal transform coefficient 137 by performing inverse quantization on the quantized orthogonal transform coefficient 135 transmitted from the quantization unit 109, and sends the orthogonal transform coefficient 137 to the inverse orthogonal transform unit 112.

  Then, the inverse orthogonal transform unit 112 performs an inverse orthogonal transform on the orthogonal transform coefficient 137 transmitted from the inverse quantization unit 111 to generate a prediction residual signal 138 and sends the prediction residual signal 138 to the adder 113.

  The adder 113 adds the prediction residual signal 138 transmitted from the inverse orthogonal transform unit 112 and the prediction image signal 132 transmitted from the switch 106 to generate a frame image signal 139 and sends it to the frame memory 104. This frame image signal 139 is stored in the frame memory 104 and used as a reference frame image signal in the subsequent encoding process. In addition, information on motion vectors and reference frame numbers is also included in the reference frame image signal and stored simultaneously.

  Next, the motion detection unit 102 of the video encoding device 10 will be described in detail with reference to FIG. FIG. 2 is a configuration diagram of the motion detection unit of FIG.

  As shown in FIG. 2, the motion detection unit 102 includes a prediction mode determination unit 201, a reference frame determination unit 202, a motion vector detection unit 203, and a motion vector prediction unit (motion vector prediction unit). ) 204 and a motion vector difference unit 205.

  First, the prediction mode determination unit 201 uses “INTER prediction mode” or “INTRA prediction mode” as the encoding mode of a predetermined block to be encoded based on the input frame image signal 123 and reference frame image signal 124. Determine whether to use and determine the prediction mode. When “INTRA prediction mode” is selected, the prediction mode 131 is output and the process is terminated. When “INTER prediction mode” is selected, the prediction mode determination unit 201 outputs the prediction mode 131 and simultaneously outputs a signal 210 including the frame image signal, the reference frame image signal, and the prediction mode to the reference frame determination unit 202. send.

  The reference frame determination unit 202 determines a reference frame for detecting and predicting a motion vector of a predetermined block to be encoded based on the input frame image signal, the reference frame image signal, and the prediction mode. A signal 211 including an image signal, a reference frame image signal, a prediction mode, and a reference frame number is sent to the motion vector detection unit 203. At the same time, the reference frame determination unit 202 sends a signal 212 including the reference frame image signal, the prediction mode, and the reference frame number to the motion vector prediction unit 204.

  Based on the input frame image signal, reference frame image signal, prediction mode, and reference frame number, the motion vector detection unit 203 determines the image signal in the current frame from the image signal indicated by the reference frame number in the reference frame image signal. Find an image signal pattern similar to the pattern. Then, a motion vector that is a spatial displacement amount between both image signal patterns is detected, and a signal 213 including the motion vector, the prediction mode, and the reference frame number is sent to the motion vector difference unit 205. In addition, a signal 126 including a motion vector to be used for motion compensation, a prediction mode, and a reference frame number is output.

  The motion vector prediction unit 204 also includes motion vectors of encoded blocks adjacent to a predetermined block to be encoded included in the reference frame image signal, their reference frame numbers, and a prediction mode of the predetermined block to be encoded. The motion vector prediction value of the predetermined block to be encoded is calculated using the reference frame number. When calculating a motion vector prediction value, a predetermined block to be encoded is used as a reference based on a frame image (target reference frame image) referred to for detecting a motion vector of the predetermined block to be encoded. Correction for scaling the motion vectors of adjacent blocks is performed (details will be described later). Scaling is a temporal relationship between a frame image (adjacent reference frame image) referenced to detect a motion vector of an adjacent block, a target reference frame image, and a frame image to be encoded (target frame image). Based on. Here, the temporal relationship among the adjacent reference frame image, the target reference frame image, and the target frame image indicates a relative time difference between the frame images or time information of each frame image. The motion vector prediction unit 204 sends a signal 215 including the calculated motion vector prediction value, the prediction mode, and the reference frame number to the motion vector difference unit 205.

The motion vector difference unit 205 calculates a motion vector difference value that is a value obtained by subtracting a motion vector prediction value from the input motion vector, and performs a variable length encoding prediction mode, a reference frame number, and a motion vector difference value. A signal 125 including
(Configuration of video decoding device)

  Next, the moving picture decoding apparatus 30 according to the present invention will be described with reference to FIG. The video decoding device 30 described below is similar to the video encoding device 10 in the H.264 format. This is a decoding device compliant with the H.264 / AVC encoding system.

  The moving picture decoding apparatus 30 has a function of using the compressed stream 136 output from the moving picture encoding apparatus 10 as an input signal and decoding it into an input video signal.

  As shown in FIG. 1, the moving picture decoding apparatus 30 includes a variable length decoding unit 301, a motion vector restoration unit 302, a motion compensation unit 303, a frame memory 304, spatial prediction, as illustrated in FIG. A unit 305, a switch 306, an inverse quantization unit 307, an inverse orthogonal transform unit 308, and an adder 309 are configured. Hereinafter, each component will be described.

  After receiving the compressed stream 136, the variable length decoding unit 301 detects a synchronization word representing the head of each frame, and then restores the prediction mode and the quantized orthogonal transform coefficient in units of blocks. When the prediction mode is “INTER prediction mode”, the motion vector difference value and the reference frame number are also decoded. The variable length decoding unit 301 transmits the signal 321 including the reconstructed prediction mode, the motion vector difference value, and the reference frame number to the motion vector reconstructing unit 302, and the reconstructed quantized orthogonal transform coefficient 322 to the inverse quantizing unit 307. The restored prediction mode 326 is sent to the switch 306 and the spatial prediction unit 305, respectively.

  When the prediction mode is “INTER prediction mode”, the motion vector restoration unit 302 calculates motion vector prediction calculated from the motion vector difference value transmitted from the variable length decoding unit 301 and the motion vector of the decoded adjacent block. The motion vector is restored using the value. Then, a signal 323 including the restored motion vector, prediction mode, and reference frame number is sent to the motion compensation unit 303.

  Next, the motion compensation unit 303 generates a predicted image signal 325 using the reference frame image signal 324 transmitted from the frame memory 304 based on the motion vector, the prediction mode, and the reference frame number, and sends it to the switch 306. send. The frame memory 304 stores previously decoded frame image signals.

  In addition, when the prediction mode 326 is the “INTRA prediction mode”, the spatial prediction unit 305 generates the predicted image signal 328 by referring to the image signal (reference frame image signal 327) of the decoded neighboring block, and the switch To 306.

  Next, the switch 306 selects either the predicted image signal 325 or the predicted image signal 328 according to the prediction mode 326 transmitted from the variable length decoding unit 301, and outputs the predicted image signal 329 to the adder 309. send.

  On the other hand, the inverse quantization unit 307 performs inverse quantization on the quantized orthogonal transform coefficient 322 transmitted by the variable length decoding unit 301 to restore the orthogonal transform coefficient 330, and sends it to the inverse orthogonal transform unit 308.

  The inverse orthogonal transform unit 308 performs inverse orthogonal transform on the orthogonal transform coefficient 330 to restore the prediction residual signal 331.

  Then, the adder 309 adds the predicted image signal 329 transmitted from the switch 306 and the predicted residual signal 331 transmitted from the inverse orthogonal transform unit 308 to restore the frame image signal 332.

  Finally, the frame image signal 332 is output to a display device (not shown) at a predetermined display timing, and the input video signal (moving image signal) 121 is reproduced.

  The frame image signal 332 is stored in the frame memory 304 as a reference frame image signal because it is used for the subsequent decoding process. Here, the frame image signal 332 has the same value as the frame image signal 139 having the same number in the moving image coding apparatus 10. In addition, information on motion vectors and reference frame numbers is also included in the reference frame image signal and stored simultaneously.

  Next, the motion vector restoration unit 302 of the moving picture decoding apparatus 30 will be described in detail with reference to FIG. FIG. 4 is a configuration diagram of the motion vector restoration unit of FIG.

  First, the motion vector prediction unit 401 extracts a motion vector of a decoded block adjacent to a predetermined block to be decoded included in the input reference frame image signal 324 and their reference frame numbers. Then, using the prediction mode and the reference frame number of the predetermined block to be decoded included in the input signal 321, the motion vector prediction value of the predetermined block to be decoded is calculated. When calculating a motion vector prediction value, a predetermined block to be decoded is used as a reference based on a frame image (target reference frame image) referred to for detecting a motion vector of the predetermined block to be decoded. Correction for scaling the motion vectors of adjacent blocks is performed (details will be described later). Scaling is a temporal relationship between a frame image (adjacent reference frame image) referenced to detect a motion vector of an adjacent block, a target reference frame image, and a decoding target frame image (target frame image). Based on. Here, the temporal relationship among the adjacent reference frame image, the target reference frame image, and the target frame image indicates a relative time difference between the frame images or time information of each frame image. Thereafter, a signal 421 including the prediction mode, the reference frame number, and the calculated motion vector prediction value is sent to the motion vector adding unit 402.

The motion vector addition unit 402 restores a motion vector based on the input motion vector prediction value, prediction mode, motion vector difference value, and reference frame number. Then, a signal 323 including a motion vector used for motion compensation, a prediction mode, and a reference frame number is output.
(Calculation of optimal prediction motion vector)

  Here, the calculation of the optimal prediction motion vector performed in the motion detection unit 102 of the video encoding device 10 and the motion vector restoration unit 302 of the video decoding device 30 will be described in more detail.

  The motion vector prediction unit 204 in the motion detection unit 102 of the video encoding device 10 is an optimal prediction motion vector (motion vector prediction value) used for taking a difference from the motion vector detected in the block to be encoded. Is calculated. Finally, the information transmitted as the compressed stream 136 by the moving image encoding apparatus 10 is obtained by encoding a motion vector difference value obtained by subtracting the motion vector prediction value from the motion vector. Therefore, the closer the optimum predicted motion vector is to the actual motion vector, the more efficient the encoding can be performed.

  In addition, the motion vector prediction unit 401 in the motion vector restoration unit 302 of the video decoding device 30 calculates a motion vector prediction value based on the reference frame image signal. The calculated motion vector prediction value is used to restore the motion vector by adding it to the transmitted motion vector difference value. Therefore, as in the case of the moving picture coding apparatus 10, the more efficient decoding can be performed as the optimum predicted motion vector is closer to the actual motion vector.

  Since the calculation of the motion vector prediction value by the motion vector prediction unit 204 and the calculation of the motion vector prediction value by the motion vector prediction unit 401 are the same processing, only the operation of the motion vector prediction unit 204 will be described below. .

  FIG. 5 is a diagram schematically showing a block adjacent to a block to be encoded. In FIG. 5, the block to be encoded is E, block A is a block including the pixel immediately left of the upper leftmost pixel of block E, block B is a block including the pixel immediately above the upper leftmost pixel of block E, The block C is a block including the pixel immediately above the upper right pixel of the block E. Further, the block D is a block including the pixel immediately above the leftmost pixel of the block E.

  First, when the block C is outside the screen, the motion vector prediction unit 204 assumes that the motion vector and reference frame number of the block C are the same as the motion vector and reference frame number of the block D.

  When both block B and block C are outside the screen, the motion vector and reference frame number of block B and block C are the same as the motion vector and reference frame number of block A.

  On the premise as described above, the motion vector prediction unit 204 always makes the motion vector and the reference frame number of the block A, block B, and block C adjacent to the block E to be encoded exist, and then the motion vector. Calculate the predicted value.

  FIG. 6 is a flowchart showing the operation of the motion vector prediction unit 204 when calculating the motion vector prediction value.

  First, the motion vector prediction unit 204 refers to the motion vector and reference frame image number of a block (A, B, C) adjacent to the predetermined block (E) to be encoded included in the reference frame image signal (step). S01).

  Next, it is determined whether only one of the reference frame image numbers of the adjacent blocks (A, B, C) is equal to the reference frame image number of the block (E) (step S02). If any one of the reference frame image numbers of the adjacent blocks (A, B, C) is equal to the reference frame image number of the block (E) (step S02: YES), the reference frame of the block (E) The motion vector value of the block having the reference frame image number equal to the image number is determined as the motion vector prediction value of the block (E) (step S03). If it is not the case that any one of the reference frame image numbers of the adjacent blocks (A, B, C) is equal to the reference frame image number of the block (E) (step S02: NO), the process is performed in step S04. Migrate to

  Subsequently, the motion vector prediction unit 204 performs motion compensation on each adjacent block (A, B, C) using the same reference frame image as the reference frame image number in which the motion vector of the block (E) is detected. Is determined (step S04). When the detected reference frame image of the adjacent block (A, B, C) is not the same reference frame image as the reference frame image from which the motion vector of the block (E) is detected (step S04: NO), the block to be encoded The motion vector values of the adjacent blocks (A, B, C) are scaled so as to satisfy the same reference frame conditions as in (E) (step S05).

This scaling method will be described with reference to FIG. FIG. 7 is a diagram illustrating a motion vector of a block adjacent to the encoding target block in time space. As shown in FIG. 7, the temporal distance (t0-te) from the current frame (target frame image) 701 (time t0) to the reference frame (target reference frame image) 702 of the encoding target block (E) is used as a reference. Thus, the magnitudes of the motion vectors 751a, 751b, 751c of the blocks (A, B, C) adjacent to the detected predetermined block to be encoded are changed. Specifically, the reference frame 702 of the encoding target block (E) is the time te, the motion vector 751a of the adjacent block A is (MVxA, MVyA), and the reference frame (adjacent reference frame image) 703a is the time ta. , The motion vectors (MVxA ′, MVyA ′) scaled with reference to the reference frame 702 of the encoding target block (E) are obtained by the following equations (1) and (2).

  In this case, the adjacent blocks (A, B, C) are not necessarily motion-compensated using the past reference frame image, and the future ((t0−t) <0) reference frame image is used. Motion compensation may be performed.

  On the other hand, referring back to FIG. 6, when the detected reference frame image of the adjacent block (A, B, C) is the same reference frame image as the reference frame image from which the motion vector of the block (E) is detected (step S04). : NO), the motion vector values of the blocks (A, B, C) are used as they are after being scaled (step S06).

  Next, it is determined whether all the motion vectors of the adjacent blocks (A, B, C) have been scaled (step S07). If all adjacent blocks (A, B, C) have not been scaled (step S07: NO), the processing from step S04 is repeated.

  When all the adjacent blocks (A, B, C) have been scaled (step S07: YES), they have been scaled based on the reference frame 702 in which the motion vector of the encoding target block (E) is detected. By calculating the intermediate value of the motion vectors of the adjacent blocks (A, B, C), the optimum predicted motion vector is determined (step S08). The intermediate value is calculated by scaling the motion vectors (MVxA ′, MVyA ′), (MVxB ′, MVyB ′), (MVxC ′, MVyC ′) of three adjacent blocks (A, B, C) to x, y. Comparison is made for each component, and an intermediate (second) value in each component is set as a motion vector prediction value. For example, when (MVxA ′, MVyA ′) = (3, −5), (MVxB ′, MVyB ′) = (− 1, 4), (MVxC ′, MVyC ′) = (2, 6) The predicted values (PMVxE, PMVyE) are (2, 4) which are intermediate (second) values for each of the x and y components.

  When the motion vector of the encoding target block is predicted using the intermediate value of the motion vector of the adjacent block as described above, the optimal prediction motion vector is determined from the spatial correlation between the encoding target block and the surrounding blocks. It is considered that the value is relatively close to the actual motion vector of the block. This method is called median prediction.

  Next, a video encoding program 910 for causing a computer to function as the above-described video encoding device 10 and a video decoding program 930 for causing the computer to function as the above-described video decoding device 30 will be described. . 10 and 11 are diagrams showing configurations of the moving image encoding program 910 and the moving image decoding program 930, respectively.

  As illustrated in FIG. 10, the moving image encoding program 910 includes a main module 911 that controls processing, an input module 912, a motion detection module 913, a motion compensation module 914, a spatial prediction module 915, and a switch module 916. A subtraction module 917, an orthogonal transform module 918, a quantization module 919, a variable length coding module 920, an inverse quantization module 921, an inverse orthogonal transform module 922, and an addition module 923. Input module 912, motion detection module 913, motion compensation module 914, spatial prediction module 915, switch module 916, subtraction module 917, orthogonal transform module 918, quantization module 919, variable length encoding module 920, inverse quantization module 921, The functions that the computer performs by the inverse orthogonal transform module 922 and the addition module 923 are the above-described input unit 101, motion detection unit 102, motion compensation unit 103, spatial prediction unit 105, switch 106, subtractor 107, and orthogonal transform unit 108, respectively. , The quantization unit 109, the variable length coding unit 110, the inverse quantization unit 111, the inverse orthogonal transform unit 112, and the adder 113.

  As shown in FIG. 11, the moving picture decoding program 930 includes a main module 931 that supervises processing, a variable length decoding module 932, a motion vector restoration module 933, a motion compensation module 934, and a spatial prediction module. 935, a switch module 936, an inverse quantization module 937, an inverse orthogonal transform module 938, and an addition module 939. The functions that the variable length decoding module 932, motion vector restoration module 933, motion compensation module 934, spatial prediction module 935, switch module 936, inverse quantization module 937, inverse orthogonal transform module 938, and addition module 939 implement in the computer are respectively These are the same as the variable length decoding unit 301, motion vector restoration unit 302, motion compensation unit 303, spatial prediction unit 305, switch 306, inverse quantization unit 307, inverse orthogonal transform unit 308, and adder 309 described above.

According to the video encoding device 10 and the video decoding device 30 according to the first embodiment described above, the motion vector prediction units 204 and 401 use the motion vectors of the adjacent blocks as the motion vectors of the target block. Since the optimal prediction motion vector is predicted based on the corrected motion vector of the adjacent block after scaling and correction based on the time difference between the frame image referred to for detection and the frame image to be encoded as a reference, time By determining the optimal prediction motion vector in consideration of the continuity of the typical motion, the difference between the actual motion vector of the target block and the optimal prediction motion vector can be further reduced.
[Second Embodiment]

  Next, a second embodiment of the present invention will be described. The basic configurations of the moving image encoding device and the moving image decoding device in the present embodiment are the same as the configurations of the moving image encoding device 10 and the moving image decoding device 30 in the first embodiment. Are denoted by the same reference numerals, description thereof is omitted, and differences from the first embodiment will be described in detail below.

  The difference between the video encoding device and video decoding device in the second embodiment and the video encoding device 10 and video decoding device 30 in the first embodiment relates to the calculation of motion vector prediction values. Part. Hereinafter, calculation of a predicted motion vector performed by the motion detection unit 102 of the video encoding device 10 and the motion vector restoration unit 302 of the video decoding device 30 different from the first embodiment will be described.

  Since the calculation of the motion vector prediction value by the motion vector prediction unit 204 and the calculation of the motion vector prediction value by the motion vector prediction unit 401 are the same processing, only the operation of the motion vector prediction unit 204 will be described below. .

  In the calculation of the motion vector prediction value by the motion vector prediction unit 204 according to the present embodiment, the motion vector of each adjacent block is scaled in advance based on the target reference frame, and the optimal prediction motion vector is based on the scaled motion vector. Is not determined. That is, the motion vector predicting unit 204 first determines an optimum one to be used for motion vector prediction based on the motion vector of each adjacent block before scaling, and then uses the determined motion vector prediction value based on the target reference frame. To obtain a motion vector prediction value.

  FIG. 8 is a flowchart showing an operation at the time of calculating a motion vector prediction value of the motion vector prediction unit 204 according to the present embodiment.

  First, the motion vector prediction unit 204 refers to the motion vector and reference frame image number of a block (A, B, C) adjacent to the predetermined block (E) to be encoded included in the reference frame image signal (step). S201).

  Next, it is determined whether only one of the reference frame numbers of the adjacent blocks (A, B, C) is equal to the reference frame number of the block (E) (step S202). When only one of the reference frame numbers of the adjacent blocks (A, B, C) is equal to the reference frame number of the block (E) (step S202: YES), the adjacent block having the same reference frame number is selected. The motion vector value is determined as the motion vector prediction value of the block (E) (step S203). If only one of the reference frame numbers of the adjacent blocks (A, B, C) is equal to the reference frame number of the block (E) (step S202: NO), the process goes to step S204. Transition.

  Subsequently, the reference frame (target reference frame image) in which the motion vector of the encoding target block (E) is detected by the motion vector prediction unit 204 based on the motion vector of each adjacent block (A, B, C). An optimal motion vector is selected as a motion vector prediction value without performing scaling on the basis (step S204). As a method for selecting an optimal motion vector as a motion vector prediction value, as in the first embodiment, x is selected from motion vectors of blocks (A, B, C) adjacent to the block (E) to be encoded. , One having an intermediate value for each y component is selected.

  After the motion vector is selected, the reference frame in which each motion vector selected as the x and y component as the optimum motion vector prediction value of the encoding target block (E) detects the motion vector of the block (E). It is determined whether motion compensation has been performed using the same reference frame image as the image number (step S205). When each selected motion vector is not motion-compensated using the same reference frame image as the reference frame image in which the motion vector of the encoding target block (E) is detected (step S205: NO), The size of the x and y components of the motion vector is scaled so as to satisfy the same reference frame conditions as the block (E) to be converted (step S206). As a scaling method, similar to the first embodiment, the adjacent blocks selected based on the temporal distance from the current frame (target frame image) to the reference frame (target reference frame image) of the block to be encoded , And the x component or y component is used as a motion vector prediction value.

  On the other hand, when each selected motion vector is motion-compensated using the same reference frame image as the reference frame image in which the motion vector of the encoding target block (E) is detected (step S205: YES) ), The x component or y component of the selected motion vector is determined as the motion vector prediction value as it is (step S207).

  According to the video encoding device 10 and the video decoding device 30 according to the second embodiment described above, the motion vector prediction units 204 and 401 have determined the optimal prediction motion vector based on the motion vector of the adjacent block. Thereafter, the determined optimal prediction motion vector is corrected by scaling based on the time difference between the frame image referred to in order to detect the motion vector of the target block and the frame image to be encoded. Thus, the difference between the motion vector and the optimum predicted motion vector can be further reduced, and at the same time, the processing time for predicting the motion vector can be shortened.

  In addition, in the moving picture coding apparatus 10 and the moving picture decoding apparatus 30 according to the first embodiment and the second embodiment described above, the intermediate motion vector is selected from the motion vectors of adjacent blocks before scaling or after scaling. Although the value is selected and determined as the motion vector prediction value, the motion vector prediction value (PMVxE, PMVyE) may be determined as follows.

  That is, in the first embodiment, when the x component of each scaled motion vector satisfies the condition 1: | MVxA′−MVxB ′ | <| MVxB′−MVxC ′ |, PMVxE = MVxA ′ is determined. . When the above condition 1 is not satisfied, PMVxE = MVxB ′ is determined. Similarly, when the y component of each scaled motion vector satisfies the condition 2: | MVyA′−MVyB ′ | <| MVyB′−MVyC ′ |, PMVyE = MVyA ′ is determined. When the above condition 2 is not satisfied, PMVyE = MVyB ′ is determined. In the second embodiment, when the x component of each motion vector before scaling satisfies the condition 1: | MVxA−MVxB | <| MVxB−MVxC |, PMVxE = MVxA is determined. When the above condition 1 is not satisfied, PMVxE = MVxB is determined. Similarly, when the y component of each motion vector before scaling satisfies the condition 2: | MVyA−MVyB | <| MVyB−MVyC |, PMVyE = MVyA is determined. When the above condition 2 is not satisfied, PMVyE = MVyB is determined. Thereafter, the determined motion vector prediction value is scaled.

  Further, the motion vector prediction value may be determined by the following method. That is, the position of a block used as a motion vector prediction value may be uniquely determined according to the method for dividing a predetermined block to be encoded, and the motion vector of that block may be always selected. FIG. 9 is a diagram illustrating an example of blocks divided for motion vector prediction. According to the example in FIG. 9A, in the prediction mode in which the block E to be encoded is divided into two areas of 16 pixels × 8 pixels, the upper area represents the motion vector of the block B and the lower area. Determines the motion vector of block A as a motion vector prediction value. Further, according to the example of FIG. 9B, in the prediction mode in which the block E is divided into two regions of 8 pixels × 16 pixels, the left region is the motion vector of the block A, and the right region is the block C. Are determined as predicted motion vectors. Note that the position of the motion vector selected for the region divided by each prediction mode is an example, and the motion vector at an arbitrary position can be determined as the motion vector prediction value of the block to be encoded. it can.

  Further, in the moving picture coding apparatus 10 and the moving picture decoding apparatus 30 according to the first and second embodiments described above, as shown in FIG. 5, the motion vector prediction of the block (E) to be coded is performed. In order to determine the value, the motion vector of the adjacent block (A, B, C) is used. However, the number of the adjacent blocks and the relative position with the block (E) may be appropriately changed.

  Next, it extracts from each embodiment of the moving image encoding device, moving image decoding device, moving image encoding method, moving image decoding method, moving image encoding program, and moving image decoding program of the present invention described above. The technical ideas to be listed are enumerated according to the description format of the claims. The technical idea according to the present invention can be grasped by various levels and variations from a superordinate concept to a subordinate concept, and the present invention is not limited to the following description.

(Claim 1) A frame image to be encoded in a moving image signal composed of a time sequence of frame image signals is divided into a plurality of target regions, and each target region has a plurality of different frame images from the encoding target frame image. In a moving image coding apparatus that performs coding by motion compensation by detecting a motion vector with reference to a frame image,
An adjacent reference frame image referred to for detecting a motion vector of an adjacent region adjacent to the target region, a target reference frame image referred to for detecting a motion vector of the target region, and the encoding target Correction for scaling a motion vector of an adjacent region adjacent to the target region based on the target reference frame image based on a temporal relationship with the target frame image that is a frame image or their time information;
Determining an optimal prediction motion vector based on a motion vector of an adjacent region adjacent to the target region;
And a motion vector predicting means for predicting the corrected optimally predicted motion vector by performing the correction.

  (Claim 2) The motion vector prediction means uses the target reference frame image as a reference based on a temporal relationship between the adjacent reference frame image, the target reference frame image, and the target frame image or time information thereof. The moving picture encoding apparatus according to claim 1, wherein each of the motion vectors in the adjacent area is scaled and corrected, and an optimal prediction motion vector is determined based on the corrected motion vector in the adjacent area. .

  (Claim 3) The motion vector prediction means determines an optimal prediction motion vector based on the motion vector of the adjacent region, and the adjacent reference frame image of the motion vector of the adjacent region determined as the optimal prediction motion vector and the target The optimal prediction motion vector is scaled and corrected based on the target reference frame image based on a temporal relationship between the reference frame image and the target frame image or their time information. The moving image encoding apparatus described in 1.

(Claim 4) A frame image to be decoded in a moving image signal composed of a time sequence of frame image signals is divided into a plurality of target regions, and a plurality of frames different from the frame images to be decoded for each target region. In a video decoding device that performs decoding by motion compensation by using difference information between a motion vector detected with reference to an image and a predicted motion vector,
An adjacent reference frame image referred to for detecting a motion vector of an adjacent region adjacent to the target region, a target reference frame image referred to for detecting a motion vector of the target region, and the encoding target Correction for scaling a motion vector of an adjacent region adjacent to the target region based on the target reference frame image based on a temporal relationship with the target frame image that is a frame image or their time information;
Determining an optimal prediction motion vector based on a motion vector of an adjacent region adjacent to the target region;
A motion picture decoding apparatus comprising: motion vector prediction means for predicting the corrected optimal predicted motion vector by performing

  (Claim 5) The motion vector prediction means uses the target reference frame image as a reference based on a temporal relationship between the adjacent reference frame image, the target reference frame image, and the target frame image or time information thereof. 5. The moving picture decoding apparatus according to claim 4, wherein each of the motion vectors of the adjacent area is scaled and corrected, and an optimal prediction motion vector is determined based on the corrected motion vector of the adjacent area. .

  (Claim 6) The motion vector prediction means determines an optimal prediction motion vector based on the motion vector of the adjacent region, and the adjacent reference frame image of the motion vector of the adjacent region determined as the optimal prediction motion vector and the target 5. The optimal prediction motion vector is scaled and corrected based on the target reference frame image based on a temporal relationship between the reference frame image and the target frame image or time information thereof. The moving picture decoding apparatus described in 1.

(Claim 7) A frame image to be encoded in a moving image signal composed of a time sequence of frame image signals is divided into a plurality of target regions, and each of the target regions has a plurality of different from the frame image to be encoded. In a moving image coding method for performing motion compensation coding by detecting a motion vector with reference to a frame image,
The motion vector prediction means
An adjacent reference frame image referred to for detecting a motion vector of an adjacent region adjacent to the target region, a target reference frame image referred to for detecting a motion vector of the target region, and the encoding target Correction for scaling a motion vector of an adjacent region adjacent to the target region based on the target reference frame image based on a temporal relationship with the target frame image that is a frame image or their time information;
Determining an optimal prediction motion vector based on a motion vector of an adjacent region adjacent to the target region;
And a motion vector predicting step for predicting the corrected optimal predicted motion vector by performing the correction.

(Claim 8) A frame image to be decoded in a moving image signal composed of a time sequence of frame image signals is divided into a plurality of target regions, and a plurality of frames different from the frame image to be decoded for each target region. In a video decoding method for decoding by motion compensation by using difference information between a motion vector detected with reference to an image and a predicted motion vector,
The motion vector prediction means
An adjacent reference frame image referred to for detecting a motion vector of an adjacent region adjacent to the target region, a target reference frame image referred to for detecting a motion vector of the target region, and the encoding target Correction for scaling a motion vector of an adjacent region adjacent to the target region based on the target reference frame image based on a temporal relationship with the target frame image that is a frame image or their time information;
Determining an optimal prediction motion vector based on a motion vector of an adjacent region adjacent to the target region;
A motion picture decoding method comprising: a motion vector prediction step of predicting the corrected optimal prediction motion vector by performing

(Section 9) A frame image to be encoded in a moving image signal composed of a time sequence of frame image signals is divided into a plurality of target regions, and each target region has a plurality of different frame images that are different from the frame image to be encoded. In a moving image encoding program for causing a computer to execute an encoding process by motion compensation by detecting a motion vector with reference to a frame image,
Computer
An adjacent reference frame image referred to for detecting a motion vector of an adjacent region adjacent to the target region, a target reference frame image referred to for detecting a motion vector of the target region, and the encoding target Correction for scaling a motion vector of an adjacent region adjacent to the target region based on the target reference frame image based on a temporal relationship with the target frame image that is a frame image or their time information;
Determining an optimal prediction motion vector based on a motion vector of an adjacent region adjacent to the target region;
Motion vector predicting means for predicting the corrected optimal predicted motion vector by performing
A video encoding program characterized by being made to function as:

(Claim 10) A frame image to be decoded in a moving image signal composed of a time sequence of frame image signals is divided into a plurality of target regions, and a plurality of frames different from the frame image to be decoded for each target region In a video decoding program for causing a computer to execute a decoding process by motion compensation by using difference information between a motion vector detected with reference to an image and a predicted motion vector,
Computer
An adjacent reference frame image referred to for detecting a motion vector of an adjacent region adjacent to the target region, a target reference frame image referred to for detecting a motion vector of the target region, and the encoding target Correction for scaling a motion vector of an adjacent region adjacent to the target region based on the target reference frame image based on a temporal relationship with the target frame image that is a frame image or their time information;
Determining an optimal prediction motion vector based on a motion vector of an adjacent region adjacent to the target region;
Motion vector predicting means for predicting the corrected optimal predicted motion vector by performing
A moving picture decoding program characterized by being made to function as:

DESCRIPTION OF SYMBOLS 10 ... Moving image encoder, 30 ... Moving image decoder, 101 ... Input part, 102 ... Motion detection part, 103 ... Motion compensation part, 104 ... Frame memory, 105 ... Spatial prediction part, 106 ... Switch, 107 ... Subtractor 108... Orthogonal transform unit 109 109 Quantization unit 110 Variable length encoding unit 111 Inverse quantization unit 112 Inverse orthogonal transform unit 113 Adder 201 Prediction mode determination unit 202 Reference frame determination unit, 203 ... Motion vector detection unit, 204 ... Motion vector prediction unit (motion vector prediction means), 205 ... Motion vector difference unit, 301 ... Variable length decoding unit, 302 ... Motion vector restoration unit, 303 ... Motion compensation unit, 304 ... frame memory, 305 ... spatial prediction unit, 306 ... switch, 307 ... inverse quantization unit, 308 ... inverse orthogonal transform unit, 309 ... adder, 401 ... motion Vector prediction unit (motion vector prediction means), 402 ... motion vector addition unit, 701 ... target frame image, 702 ... target reference frame image 703a, 703b, 703c ... adjacent reference frame images 751a, 751c, 751c ... motion vector.

Claims (10)

A frame image to be encoded in a moving image signal composed of a time sequence of frame image signals is divided into a plurality of target regions, and a plurality of frame images different from the frame image to be encoded are referenced for each target region. In the moving picture coding apparatus that performs coding by motion compensation by detecting a motion vector,
An adjacent reference frame image referred to for detecting a motion vector of an adjacent region adjacent to the target region, a target reference frame image referred to for detecting a motion vector of the target region, and the encoding target Correction for scaling a motion vector of an adjacent region adjacent to the target region based on the target reference frame image based on a temporal relationship with the target frame image that is a frame image or their time information;
A moving picture code comprising: motion vector prediction means for predicting the corrected optimal predicted motion vector by determining an optimal predicted motion vector based on a motion vector of an adjacent area adjacent to the target area Device.   The motion vector predicting unit is configured to determine the adjacent region based on the target reference frame image based on a temporal relationship between the adjacent reference frame image, the target reference frame image, and the target frame image or time information thereof. 2. The moving picture encoding apparatus according to claim 1, wherein each of the motion vectors is corrected by scaling, and an optimal prediction motion vector is determined based on the corrected motion vector of the adjacent region.   The motion vector prediction means determines an optimal prediction motion vector based on the motion vector of the adjacent region, and an adjacent reference frame image of the motion vector of the adjacent region determined as the optimal prediction motion vector, the target reference frame image, 2. The optimal prediction motion vector is scaled and corrected based on the target reference frame image based on a temporal relationship with the target frame image or time information thereof. Video encoding device. A frame image to be decoded in a moving image signal composed of a time sequence of frame image signals is divided into a plurality of target regions, and a plurality of frame images different from the frame images to be decoded are referred to for each target region. In the moving picture decoding apparatus that performs decoding by motion compensation by using the difference information between the detected motion vector and the predicted motion vector,
An adjacent reference frame image referred to for detecting a motion vector of an adjacent region adjacent to the target region, a target reference frame image referred to for detecting a motion vector of the target region, and the encoding target Correction for scaling a motion vector of an adjacent region adjacent to the target region based on the target reference frame image based on a temporal relationship with the target frame image that is a frame image or their time information;
A moving picture decoding unit comprising: a motion vector predicting unit configured to predict an optimal predicted motion vector after the correction by determining an optimal predicted motion vector based on a motion vector of an adjacent region adjacent to the target region; Device.   The motion vector predicting unit is configured to determine the adjacent region based on the target reference frame image based on a temporal relationship between the adjacent reference frame image, the target reference frame image, and the target frame image or time information thereof. 5. The moving picture decoding apparatus according to claim 4, wherein each of the motion vectors is corrected by scaling, and an optimal prediction motion vector is determined based on the corrected motion vector of the adjacent region.   The motion vector prediction means determines an optimal prediction motion vector based on the motion vector of the adjacent region, and an adjacent reference frame image of the motion vector of the adjacent region determined as the optimal prediction motion vector, the target reference frame image, 5. The optimal prediction motion vector is scaled and corrected based on the target reference frame image based on a temporal relationship with the target frame image or time information thereof. Video decoding device. A frame image to be encoded in a moving image signal composed of a time sequence of frame image signals is divided into a plurality of target regions, and a plurality of frame images different from the frame image to be encoded are referenced for each target region. In the moving picture coding method for performing the motion compensation coding by detecting the motion vector,
The motion vector prediction means
An adjacent reference frame image referred to for detecting a motion vector of an adjacent region adjacent to the target region, a target reference frame image referred to for detecting a motion vector of the target region, and the encoding target Correction for scaling a motion vector of an adjacent region adjacent to the target region based on the target reference frame image based on a temporal relationship with the target frame image that is a frame image or their time information;
A motion vector predicting step for predicting the corrected optimal prediction motion vector by determining an optimal prediction motion vector based on a motion vector of an adjacent region adjacent to the target region. Method. A frame image to be decoded in a moving image signal composed of a time sequence of frame image signals is divided into a plurality of target regions, and a plurality of frame images different from the frame images to be decoded are referred to for each target region. In the moving picture decoding method for performing decoding by motion compensation by using difference information between the motion vector detected in this way and the predicted motion vector,
The motion vector prediction means
An adjacent reference frame image referred to for detecting a motion vector of an adjacent region adjacent to the target region, a target reference frame image referred to for detecting a motion vector of the target region, and the encoding target Correction for scaling a motion vector of an adjacent region adjacent to the target region based on the target reference frame image based on a temporal relationship with the target frame image that is a frame image or their time information;
A motion vector predicting step for predicting the corrected optimal prediction motion vector by determining an optimal prediction motion vector based on a motion vector of an adjacent region adjacent to the target region. Method. A frame image to be encoded in a moving image signal composed of a time sequence of frame image signals is divided into a plurality of target regions, and a plurality of frame images different from the frame image to be encoded are referenced for each target region. In the moving picture encoding program for causing the computer to execute the encoding process by motion compensation by detecting the motion vector
Computer
An adjacent reference frame image referred to for detecting a motion vector of an adjacent region adjacent to the target region, a target reference frame image referred to for detecting a motion vector of the target region, and the encoding target Correction for scaling a motion vector of an adjacent region adjacent to the target region based on the target reference frame image based on a temporal relationship with the target frame image that is a frame image or their time information;
A motion vector predicting unit that predicts the corrected optimal predicted motion vector by performing determination of an optimal predicted motion vector based on a motion vector of an adjacent region adjacent to the target region. Image encoding program. A frame image to be decoded in a moving image signal composed of a time sequence of frame image signals is divided into a plurality of target regions, and a plurality of frame images different from the frame images to be decoded are referred to for each target region. In a moving picture decoding program for causing a computer to execute a decoding process by motion compensation by using difference information between a detected motion vector and a predicted motion vector,
Computer
An adjacent reference frame image referred to for detecting a motion vector of an adjacent region adjacent to the target region, a target reference frame image referred to for detecting a motion vector of the target region, and the encoding target Correction for scaling a motion vector of an adjacent region adjacent to the target region based on the target reference frame image based on a temporal relationship with the target frame image that is a frame image or their time information;
A motion vector predicting unit that predicts the corrected optimal predicted motion vector by performing determination of an optimal predicted motion vector based on a motion vector of an adjacent region adjacent to the target region. Image decoding program.

Images