JP5736083B2 - Moving picture predictive coding apparatus, moving picture predictive coding method, moving picture predictive coding program, moving picture predictive decoding apparatus, moving picture predictive decoding method, and moving picture predictive decoding program - Google Patents

Description

  The present invention relates to a video predictive encoding device, a video predictive encoding method, a video predictive encoding program, a video predictive decoding device, a video predictive decoding method, and a video predictive decoding program. Moving picture predictive coding apparatus, moving picture predictive coding method, moving picture predictive coding program, moving picture predictive decoding apparatus, moving picture that generates final block predicted signal (bi-predicted signal) by averaging predicted signals The present invention relates to an image predictive decoding method and a moving image predictive decoding program.

  In order to efficiently transmit and store still images and moving image data, compression coding technology is used. In the case of moving images, MPEG-1 to 4 and ITU (International Telecommunication Union) H.264. 261-H. H.264 is widely used.

  In these encoding methods, encoding / decoding processing is performed after an image to be encoded is divided into a plurality of blocks. In predictive coding within a screen, a predicted signal is generated using an adjacent previously reproduced image signal (reconstructed compressed image data) in the same screen as the target block, and then the predicted signal is The differential signal subtracted from the signal of the target block is encoded. In predictive coding between screens, the adjacent reproduced image signal in the screen different from the target block is referred to, the motion is corrected, the predicted signal is generated, and the predicted signal is subtracted from the signal of the target block. Encode the difference signal.

  For example, H.M. H.264 intra-screen predictive encoding employs a method of generating a prediction signal by extrapolating already reproduced pixel values adjacent to a block to be encoded in a predetermined direction. FIG. 2 is a schematic diagram for explaining an intra-screen prediction method used for H.264. In FIG. 20A, a target block 802 is a block to be encoded, and a pixel group 801 composed of pixels A to M adjacent to the boundary of the target block 802 is an adjacent region, and has already been obtained in past processing. It is the reproduced image signal.

  In this case, a prediction signal is generated by extending the pixel group 801 that is an adjacent pixel directly above the target block 802 downward. In FIG. 20B, the already reproduced pixels (I to L) on the left of the target block 804 are stretched to the right to generate a prediction signal. A specific method for generating a prediction signal is described in Patent Document 1, for example. Thus, the difference between the nine prediction signals generated by the method shown in FIGS. 20A to 20I and the pixel signal of the target block is calculated, and the one with the smallest difference value is set as the optimum prediction signal. . As described above, a prediction signal can be generated by extrapolating pixels. The above contents are described in Patent Document 1 below.

  In normal inter-screen predictive coding, a prediction signal is generated by searching for a signal similar to the pixel signal from a screen that has already been reproduced for a block to be coded. Then, a motion vector, which is a spatial displacement amount between the target block and a region formed by the searched signal, and a residual signal between the pixel signal and the prediction signal of the target block are encoded. Such a method for searching for a motion vector for each block is called block matching.

  FIG. 19 is a schematic diagram for explaining the block matching process. Here, a procedure for generating a prediction signal will be described using the target block 702 on the encoding target screen 701 as an example. The reference screen 703 has already been reproduced, and the area 704 is an area in the same position as the target block 702. In block matching, a search range 705 surrounding the region 704 is set, and a region 706 in which the sum of absolute value errors with the pixel signal of the target block 702 is detected from the pixel signals in the search range 705 is detected. The signal in the area 706 becomes a prediction signal, and the amount of displacement from the area 704 to the area 706 is detected as a motion vector 707. Also, a method of preparing a plurality of reference screens 703, selecting a reference screen for performing block matching for each target block, and detecting reference screen selection information is often used. H. H.264 provides a plurality of prediction types having different block sizes for encoding motion vectors in order to cope with changes in local features of an image. H. H.264 prediction types are described in Patent Document 2, for example.

  In the compression encoding of moving image data, the encoding order of each screen (frame, field) may be arbitrary. For this reason, there are two types of coding order methods for inter-screen prediction in which a prediction signal is generated with reference to a reproduced screen. The first method is uni-prediction in which a prediction signal is generated with reference to one reproduced screen. The second method uses two prediction signals obtained by referring to one or two reproduced screens. Bi-predicting to average. The uni-prediction includes forward prediction that refers to past reproduced screens in display order and backward prediction that refers to future reproduced screens in display order. About the kind of prediction between screens, it describes in patent document 3, for example.

  H. In H.264, as a candidate for the reference screen 703, two reference screen lists including a plurality of reproduced screens are created and the second inter-screen method (bi-prediction) is performed. Block matching is performed for a plurality of reference screens registered in each reference screen list, two regions corresponding to the region 706 are detected, and the two detected prediction signals are averaged.

  An example of the reference screen list will be described with reference to FIGS. In FIG. 5A, a screen 505 indicates an encoding target image, and screens 501 to 504 indicate reproduced screens. In FIG. 5B, the screen 510 indicates the encoding target screen, and the screens 507, 508, 509, and 511 indicate the reproduced screens. Each image (screen) is identified by a frame number (frame_num). List0 and List1 in FIG. 6 show two reference screen lists, FIG. 6 (A) is a reference screen list in FIG. 5 (A), and FIGS. 6 (B) and 6 (C) are references in FIG. 5 (B). An example of a screen list is shown. 6 (A) and 6 (C), four reference screens are registered in each reference screen list, and in FIG. 6 (B), two reference screens are registered in each reference screen list. Yes. Each reference screen is identified by a reference screen index (ref_idx). The reproduced images that can be registered in the reference screen list are basically arbitrary. In this case, in order to facilitate understanding of the contents, names such as 0th motion information and 1st motion information are used in accordance with the reference screen lists List0 and List1.

US Pat. No. 6,765,964 US Patent Publication No. 7003035 US Pat. No. 6,259,739

  In bi-prediction, noise contained in a prediction signal can be removed by a smoothing effect obtained by averaging two similar prediction signals. However, such a block having a large smoothing effect is often a texture region or a flat region including noise, and a plurality of similar signals exist in the reference screen.

  Since signals in these areas are highly random, when two prediction signals similar to the target block are searched from a plurality of reference screens in these areas, motion vectors between adjacent blocks may vary. Since the motion vector is encoded based on the difference from the motion vector of the adjacent block, if the motion vector varies between adjacent blocks, the amount of code increases.

  As a technique for reducing the code amount of bi-prediction, as shown in the conventional example, there is a method of deriving two motion vectors on the decoding side using the situation of adjacent blocks, but there is a strong restriction on the prediction signals that can be used. Therefore, it is difficult to increase the similarity between the two prediction signals, and a sufficient smoothing effect cannot be obtained.

  In order to solve the above-described problem, the present invention encodes one motion vector that generates a prediction signal similar to the target block, and selectively obtains another motion vector from the encoded motion information. Moving picture predictive coding apparatus, moving picture predictive coding method, moving picture predictive coding program, moving picture predictive signal that efficiently suppresses noise of prediction signal with a small amount of code for bi-prediction encoding two motion vectors An object is to provide an image predictive decoding device, a moving image predictive decoding method, and a moving image predictive decoding program.

  A video predictive coding apparatus according to an aspect of the present invention includes: an area dividing unit that divides an input image into a plurality of areas; and a target area that is an encoding target among the areas divided by the area dividing unit. A prediction signal generating means for determining a motion vector for acquiring a signal having a high correlation from the already reproduced image; a motion information recording means for storing the motion vector; a prediction signal for the target area; and a pixel signal for the target area; A residual signal generating means for generating the residual signal, a residual signal compressing means for compressing the residual signal generated by the residual signal generating means, and a reproduction residual obtained by restoring the compressed data of the residual signal A residual signal restoring means for generating a signal, a predicted motion vector similar to the motion vector of the target area selected from the previously reproduced motion vectors stored in the motion information recording means, and the target area A pixel signal of the target region is obtained by adding an encoding unit that encodes additional information determined from a motion vector, compressed data of the residual signal, and the prediction signal and the reproduction residual signal. Image recording means for restoring and storing the restored pixel signal as the already-reproduced image, wherein the motion vector includes a zeroth motion vector required to generate a zeroth prediction signal, and a first prediction A first motion vector required for signal generation, and the predicted signal generation means estimates the 0th motion vector for obtaining the 0th predicted signal, and the estimated 0th motion vector A 0th prediction motion vector similar to a motion vector is selected from a plurality of motion vectors stored in the motion information recording means, and a 0th prediction motion vector for identifying the selected 0th prediction motion vector is selected. 0th motion information estimating means for generating 0th additional information including a motion information index and a differential motion vector determined from the 0th motion vector and the 0th predicted motion vector, and the motion information recording A motion vector for generating a first prediction signal having a high correlation with the target region is selected from a plurality of motion vectors stored in the means, and a first motion vector for identifying the selected motion vector as a first prediction motion vector is selected. 1st motion information estimation means which generates the 1st additional information containing 1 prediction motion information index, sets the 1st prediction motion vector to the 1st motion vector, the 0th prediction signal, Prediction signal synthesis means for synthesizing the first prediction signal and generating a prediction signal of the target region, wherein the first motion information estimation means further acquires the first prediction signal A second motion vector is estimated to detect the second motion vector, and a second predicted motion vector similar to the estimated second motion vector is stored in the motion information recording means A second motion vector selected from a plurality of motion vectors, a second motion vector predictor index for identifying the selected second motion vector predictor, and a difference determined from the second motion vector and the second motion vector predictor A second additional information including a motion vector, and the prediction signal generation unit includes a plurality of already-reproduced images stored in the image recording unit that are all past images in the display order from the encoding target image. Includes the first predicted signal using the first motion vector, and a plurality of already-reproduced images stored in the image recording means include future images in the display order from the encoding target image. The first motion prediction signal is generated using the second motion vector, and the encoding means includes a plurality of already reproduced images stored in the image recording means in the display order. In the case of an image older than the image, a plurality of already-reproduced images stored in the image recording means are encoded for each screen or slice, the instruction information indicating that the first additional information is encoded If a future image is included in the display order in the display order, the instruction information indicating that the second additional information is encoded is encoded for each screen or slice, and the 0th The additional information and either the first additional information or the second additional information are encoded as additional information of each target area based on the instruction information.

  A moving picture predictive decoding apparatus according to an aspect of the present invention decodes additional information of a decoding target area to be decoded and compressed data of a residual signal from among compressed data divided and encoded into a plurality of areas. Decoding means for performing, motion information restoring means for restoring a motion vector required for generating a prediction signal of the target region from the additional information, motion information recording means for saving the motion vector, and the target based on the motion vector A motion compensation unit that generates a prediction signal of the region, a residual signal recovery unit that recovers the reproduction residual signal of the target region from the compressed data of the residual signal, and adds the prediction signal and the reproduction residual signal An image recording means for restoring the pixel signal of the decoding target area and storing the restored pixel signal as an already reproduced image, wherein the decoding means includes the 0th additional information and the first information. The 0th additional information is obtained by selecting one of the 0th differential motion vector and a plurality of motion vectors stored in the motion information recording means as the 0th predicted motion. A first predicted motion information index for identifying as a vector, wherein the first additional information is a first selected from a plurality of motion vectors stored in the motion information recording means. A first predicted motion information index that is identified as a predicted motion vector, wherein the motion information restoration unit generates a 0th predicted motion vector based on the 0th predicted motion information index, and the generated 0th prediction Based on the 0th motion information restoring means for restoring the 0th motion vector by adding a motion vector and the 0th differential motion vector, and the first predicted motion information index First motion information restoring means for generating a first predicted motion vector and restoring the generated first predicted motion vector as a first motion vector, wherein the motion compensation means includes the zeroth motion Based on the vector and the first motion vector, two signals acquired from the already-reproduced image are combined to generate a prediction signal of the target area, and the decoding unit further includes, for each screen or slice, the decoding unit When the instruction information indicating whether or not the first additional information includes a difference motion vector is decoded, and the instruction information indicates that the first additional information includes the first difference motion vector, When the differential motion vector is decoded as the first additional information, and the first motion information restoring means indicates that the instruction information indicates that the first additional information does not include the first differential motion vector, The first A first predicted motion vector is generated based on the predicted motion information index, the generated first predicted motion vector is restored as a first motion vector, and the indication information includes a difference vector in the first additional information. The first motion vector is generated based on the first motion vector predictor index, and the first motion vector generated and the decoded motion vector difference are added to generate the first motion vector. A vector is generated and restored.

  The present invention can also be regarded as an invention related to a video predictive encoding method, an invention related to a video predictive decoding method, an invention related to a video predictive encoding program, an invention related to a video predictive decoding program, and is as follows. Can be described in

  A video predictive encoding method according to an aspect of the present invention is a video predictive encoding method executed by a video predictive encoding device, and includes an area dividing step of dividing an input image into a plurality of areas, A prediction signal generating step for determining a motion vector for acquiring a signal having a high correlation with a target region to be encoded from regions already divided among the regions divided by the region dividing step; and the motion vector as motion information. A motion information recording step stored in a recording means; a residual signal generating step for generating a residual signal between the prediction signal of the target region and a pixel signal of the target region; and a residual signal generated by the residual signal generating step. A residual signal compressing step for compressing the difference signal, a residual signal restoring step for generating a reproduction residual signal obtained by restoring the compressed data of the residual signal, and the motion information recording unit. A predicted motion vector similar to the motion vector of the target region selected from the previously played motion vectors stored in the above, additional information determined from the motion vector of the target region, compressed data of the residual signal, A pixel signal of the target area is restored by adding the prediction signal and the reproduction residual signal, and the restored pixel signal is stored in the image recording unit as the already reproduced image. And the motion vector includes a 0th motion vector required for generating the 0th prediction signal and a first motion vector required for generating the 1st prediction signal. The prediction signal generation step estimates a zeroth motion vector for obtaining the zeroth prediction signal and is similar to the estimated zeroth motion vector. A measured motion vector is selected from a plurality of motion vectors stored in the motion information recording means, and a 0th predicted motion information index for identifying the selected 0th predicted motion vector, and the 0th motion A zeroth motion information estimating step for generating zeroth additional information including a vector and a differential motion vector determined from the zeroth predicted motion vector, and a plurality of motion vectors stored in the motion information recording means First, a motion vector that generates a first predicted signal having a high correlation with the target region is selected, and a first motion information index that includes a first predicted motion information index for identifying the selected motion vector as a first predicted motion vector is selected. A first motion information estimation step for generating the additional information and setting the first motion vector as the first motion vector, the 0th prediction signal, A prediction signal combining step of combining the first prediction signal and generating a prediction signal of the target region, and in the first motion information estimation step, the video predictive encoding device further includes: A second motion vector for obtaining a first prediction signal is estimated, a second motion vector is detected, and a second predicted motion vector similar to the estimated second motion vector is detected as the motion information. A second prediction motion information index selected from a plurality of motion vectors stored in the recording means and identifying the selected second prediction motion vector; the second motion vector and the second prediction; Second difference information including a difference motion vector determined from a motion vector is generated, and in the prediction signal generation step, the video predictive encoding device When the plurality of already-reproduced images stored in the stage are all images past the encoding target image in the display order, the first prediction signal is generated by the first motion vector, and the image recording means When a plurality of already-reproduced images stored in FIG. 4 include future images from the encoding target image in the display order, the first prediction signal is generated using the second motion vector, and the encoding step is performed. In the moving picture predictive encoding device, the plurality of already-reproduced images stored in the image recording means are all images older than the encoding target image in the display order, and each of the images or the slices In the case where instruction information for instructing that one additional information is encoded is encoded, and a plurality of previously reproduced images stored in the image recording means include future images in the display order from the encoding target image. , Screen or slice Each time, the instruction information for instructing that the second additional information is encoded is encoded, and the first additional information or the second additional information is determined based on the zeroth additional information and the instruction information. Either one is encoded as additional information of each target area.

  A moving picture predictive decoding method according to an aspect of the present invention is a moving picture predictive decoding method executed by a moving picture predictive decoding apparatus, and is a decoding target from among compressed data encoded by being divided into a plurality of regions. A decoding step for decoding the additional information of the decoding target region and the compressed data of the residual signal, a motion information restoring step for restoring a motion vector required for generating a prediction signal of the target region from the additional information, and the motion A motion information recording step for storing a vector in a motion information recording means; a motion compensation step for generating a prediction signal for the target region based on the motion vector; and a reproduction residual for the target region from the compressed data of the residual signal A residual signal restoring step for restoring a signal, and restoring the pixel signal in the decoding target area by adding the prediction signal and the reproduction residual signal An image recording step for storing the pixel signal as a reproduced image, and in the decoding step, the moving image predictive decoding device decodes the 0th additional information and the first additional information, The 0th additional information includes a 0th differential motion vector and a 0th motion vector for identifying one selected from a plurality of motion vectors stored in the motion information recording means as a 0th predicted motion vector. A first motion information recording unit that identifies one selected from a plurality of motion vectors stored in the motion information recording means as a first motion vector predictor. A predicted motion information index, wherein the motion information restoration step generates a 0th predicted motion vector based on the 0th predicted motion information index, and the generated 0th predicted motion A 0th motion information restoring step for restoring the 0th motion vector by adding a vector and the 0th differential motion vector, and generating a first predicted motion vector based on the first predicted motion information index And a first motion information restoring step that restores the generated first predicted motion vector as a first motion vector. In the motion compensation step, the moving picture predictive decoding device includes the 0th motion Based on the vector and the first motion vector, two signals acquired from the already-reproduced image are combined to generate a prediction signal of the target region, and in the decoding step, the video predictive decoding device further includes For each screen or slice, the instruction information indicating whether or not the first additional information includes a differential motion vector is decoded, and the instruction information is the first additional information. When the report indicates that the first motion vector includes a first difference motion vector, the video predictive decoding device decodes the motion vector difference as the first additional information, and in the first motion information restoration step, When the instruction information indicates that the first additional information does not include the first differential motion vector, the moving picture predictive decoding device uses the first predicted motion vector based on the first predicted motion information index. And the generated first predicted motion vector is restored as a first motion vector, and when the instruction information indicates that the first additional information includes a difference vector, the first predicted motion information A first predicted motion vector is generated based on the index, and the generated first predicted motion vector and the decoded difference motion vector are added to generate and restore the first motion vector. The features.

  A moving image predictive encoding program according to an aspect of the present invention includes a region dividing unit that divides an input image into a plurality of regions, and a target that is an encoding target among regions divided by the region dividing unit. A prediction signal generating means for determining a motion vector for acquiring a signal having a high correlation with the area from the already reproduced image; a motion information recording means for storing the motion vector; a prediction signal for the target area; Residual signal generating means for generating a residual signal with the pixel signal, residual signal compressing means for compressing the residual signal generated by the residual signal generating means, and decompressed compressed data of the residual signal A residual signal restoring means for generating a reproduction residual signal; and a prediction motion vector similar to the motion vector of the target area selected from the already reproduced motion vectors stored in the motion information recording means. And the additional information determined from the motion vector of the target region and the compressed data of the residual signal, and the prediction signal and the reproduction residual signal are added together. A moving picture predictive encoding program for restoring a pixel signal of the target region and causing the restored pixel signal to function as an image recording unit that stores the restored pixel signal as the already-reproduced image, wherein the motion vector is a 0th prediction A 0th motion vector required for signal generation and a 1st motion vector required for generation of the first prediction signal, so that the prediction signal generation means acquires the 0th prediction signal. And a 0th predicted motion vector similar to the estimated 0th motion vector from a plurality of motion vectors stored in the motion information recording means. And a 0th motion vector information index for identifying the selected 0th motion vector predictor, and a differential motion vector determined from the 0th motion vector and the 0th motion vector predictor. A motion vector for generating a first prediction signal having a high correlation with a target region from a plurality of motion vectors stored in the motion information recording unit and a 0th motion information estimation unit that generates 0th additional information And generating first additional information including a first predicted motion information index for identifying the selected motion vector as a first predicted motion vector, and using the first predicted motion vector as a first motion First motion information estimation means for setting to a vector; and prediction signal synthesis means for synthesizing the 0th prediction signal and the first prediction signal to generate a prediction signal of the target region. The first motion information estimation means further estimates a second motion vector for obtaining the first prediction signal, detects a second motion vector, and estimates the second motion vector. A second predicted motion information index for selecting a second predicted motion vector similar to a vector from a plurality of motion vectors stored in the motion information recording means and identifying the selected second predicted motion vector And a second motion vector and a difference motion vector determined from the second motion vector predictor, and the predicted signal generating means is stored in the image recording means. When the plurality of already reproduced images are all images past the encoding target image in the display order, the first prediction signal is generated with the first motion vector and stored in the image recording means. When the plurality of already reproduced images include images that are future in the display order from the image to be encoded, the first prediction signal is generated using the second motion vector, and the encoding means includes: When the plurality of already reproduced images stored in the image recording means are all images past the encoding target image in the display order, the first additional information is encoded for each screen or slice. When the instruction information to be instructed is encoded, and a plurality of already reproduced images stored in the image recording means include future images in the display order from the image to be encoded, the second information is displayed for each screen or slice. The instruction information for instructing that the additional information is encoded is encoded, and each of the 0th additional information and either the first additional information or the second additional information based on the instruction information Code as additional information of area To, characterized in that.

  A moving picture predictive decoding program according to an aspect of the present invention provides a computer that divides a computer into a plurality of areas and encodes additional information of a decoding target area and compressed data of a residual signal among compressed data encoded. Based on the motion vector, a motion information restoration unit that restores a motion vector required to generate a prediction signal of the target region from the additional information, a motion information recording unit that saves the motion vector, and Motion compensation means for generating a prediction signal of the target area, residual signal restoration means for recovering a reproduction residual signal of the target area from the compressed data of the residual signal, the prediction signal and the reproduction residual signal To restore the pixel signal of the decoding target area and to store the restored pixel signal as an already reproduced image. An image predictive decoding program, wherein the decoding means decodes the 0th additional information and the first additional information, and the 0th additional information is stored in the 0th differential motion vector and the motion information recording means. A first predicted motion information index for identifying one selected from a plurality of stored motion vectors as a zeroth predicted motion vector, wherein the first additional information is the motion information record Including a first predicted motion information index for identifying one selected from a plurality of motion vectors stored in the means as a first predicted motion vector, wherein the motion information restoring means includes the zeroth predicted motion information. A 0th motion vector is generated based on the index, and the 0th motion vector is restored by adding the generated 0th motion vector and the 0th difference motion vector. A first motion information restoration unit that generates a first motion vector predictor based on the motion information restoration unit and the first motion vector predictor index, and that restores the generated first motion vector predictor as a first motion vector. And the motion compensation means synthesizes two signals acquired from the already reproduced image based on the 0th motion vector and the first motion vector to generate a prediction signal of the target region. And the decoding means further decodes instruction information indicating whether or not the first additional information includes a differential motion vector for each screen or slice, and the instruction information is converted into the first additional information. When the information indicates that the first differential motion vector is included, the differential motion vector is decoded as the first additional information, and the first motion information restoring means is configured to display the first additional information. Affection When the report indicates that the first difference motion vector is not included, a first prediction motion vector is generated based on the first prediction motion information index, and the generated first prediction motion vector is When the instruction information is restored as a motion vector and the instruction information indicates that the first additional information includes a difference vector, a first predicted motion vector is generated based on the first predicted motion information index The first motion vector is added to the decoded motion vector difference and the first motion vector is generated and restored.

  As another aspect, the video predictive coding apparatus has a correlation between a region dividing unit that divides an input image into a plurality of regions and a target region that is an encoding target among regions divided by the region dividing unit. A prediction signal generating means for determining a motion vector for acquiring a high signal from the already reproduced image; a motion information recording means for storing the motion vector; a residual signal of the prediction signal of the target area and the pixel signal of the target area; A residual signal generating means for generating a difference signal; a residual signal compressing means for compressing the residual signal generated by the residual signal generating means; and a reproduced residual signal obtained by restoring compressed data of the residual signal. A generated residual signal, a predicted motion vector similar to the motion vector of the target region selected from the previously reproduced motion vectors stored in the motion information recording unit, and a motion of the target region Reconstructing the pixel signal of the target area by adding the additional information determined from the vector, the encoding means for encoding the compressed data of the residual signal, and adding the prediction signal and the reproduction residual signal And an image recording means for storing the restored pixel signal as the already-reproduced image, wherein the motion vector includes a zeroth motion vector required for generating a zeroth prediction signal, and a first prediction signal. A first motion vector required for generating the first motion vector, and the prediction signal generation means estimates a zeroth motion vector for obtaining the zeroth prediction signal, and the motion information recording means Using a motion vector selected from a plurality of stored motion vectors, a 0th predicted motion vector similar to the estimated 0th motion vector is generated, and the generated 0th predicted motion vector is generated. 0th motion information for generating 0th additional information including a 0th predicted motion information index for identifying the first motion vector and a differential motion vector determined from the 0th motion vector and the 0th predicted motion vector A first prediction motion that generates a first prediction signal having a high correlation with a target region, using a motion information estimation unit and a motion vector selected from a plurality of motion vectors stored in the motion information recording unit; Generating a vector, generating first additional information including a first predicted motion information index for identifying the generated first predicted motion vector, and making the first predicted motion vector a first motion vector First motion information estimation means for setting, and prediction signal synthesis means for synthesizing the 0th prediction signal and the first prediction signal to generate a prediction signal of the target region, and Movement The information estimation means further estimates a second motion vector for obtaining the first prediction signal, detects a second motion vector, and secondly resembles the estimated second motion vector. Selected from a plurality of motion vectors stored in the motion information recording means, a second predicted motion information index for identifying the selected predicted motion vector, the second motion vector, Generating second additional information including a differential motion vector determined from the second predicted motion vector, and the predicted signal generating means includes a plurality of already reproduced images stored in the image recording means in display order. In the case where all the images are past the encoding target image, the first prediction signal is generated using the first motion vector, and a plurality of already reproduced images stored in the image recording means are stored. When a future image is included in the display order from the encoding target image, the first prediction signal is generated by the second motion vector, and the encoding unit is stored in the image recording unit In the case where all of the plurality of already reproduced images are images past the encoding target image in the display order, the instruction information indicating that the first additional information is encoded is encoded for each screen or slice, When a plurality of previously reproduced images stored in the image recording means include images of the future from the encoding target image in the display order, the second additional information is encoded for each screen or slice. And encoding the 0th additional information and either the first additional information or the second additional information based on the instruction information as additional information of each target area. Also good.

  The encoding means may encode the instruction information by including it in a screen or slice header.

  As another aspect, the moving picture predictive decoding apparatus decodes the additional information of the decoding target area to be decoded and the compressed data of the residual signal from the compressed data divided and encoded into a plurality of areas. Means, motion information restoration means for restoring a motion vector required for generating a prediction signal of the target area from the additional information, motion information recording means for saving the motion vector, and the target area based on the motion vector Adding a motion compensation means for generating a prediction signal, a residual signal restoration means for restoring the reproduction residual signal of the target area from the compressed data of the residual signal, and adding the prediction signal and the reproduction residual signal. And an image recording unit that restores the pixel signal of the decoding target area and stores the restored pixel signal as an already reproduced image, wherein the decoding unit includes the zeroth additional information and the first information Additional information is decoded, and the 0th additional information is a 0th predicted motion vector obtained by selecting one of a 0th differential motion vector and a plurality of motion vectors stored in the motion information recording means. A first predicted motion information index for identifying the first additional information as one selected from a plurality of motion vectors stored in the motion information recording means. Including a first predicted motion information index to be identified as a motion vector, wherein the motion information restoring means is selected from a plurality of motion vectors stored in the motion information recording means based on the zeroth predicted motion information index The motion vector is used to generate a 0th motion vector predictor and the 0th motion vector predictor and the 0th motion vector difference are added to generate a motion vector 0th. Using a motion vector selected from a plurality of motion vectors stored in the motion information recording unit based on the first predicted motion information index, First motion information restoring means for generating a predicted motion vector and restoring the generated first predicted motion vector as a first motion vector, wherein the motion compensation means includes the zeroth motion vector and the first motion vector. A prediction signal for the target region is generated by combining two signals respectively obtained from the already-reproduced image based on one motion vector, and the decoding means further includes the first signal for each screen or slice. The instruction information indicating whether or not the additional information includes a difference motion vector is decoded, and the instruction information indicates that the first additional information does not include the first difference motion vector. If the first predicted motion information index is decoded as the first additional information, and the indication information indicates that the first additional information includes a first differential motion vector, As the first additional information, a differential motion vector and a first predicted motion information index are decoded, and the first motion information restoring means includes the instruction information and the first additional information as the first differential motion vector. Is included, the first predicted motion vector is determined using a motion vector selected from a plurality of motion vectors stored in the motion information recording unit based on the first predicted motion information index. When the generated first predicted motion vector is restored as the first motion vector and the instruction information indicates that the first additional information includes the first difference vector, the first motion vector is generated. Predicted motion information index first generates a predicted motion vector based on the first and the prediction motion vector and differential motion vector decoding may restore the first motion vector and adds generated.

  The decoding means may decode the instruction information included in the screen or slice header.

  As another aspect, the video predictive encoding method is a video predictive encoding method executed by a video predictive encoding device, and includes an area dividing step of dividing an input image into a plurality of areas, and the area dividing A prediction signal generating step for determining a motion vector for acquiring a signal having a high correlation with the target region to be encoded among the regions divided by the step from the already reproduced image; and the motion vector as motion information recording means. A motion information recording step, a residual signal generation step for generating a residual signal between the prediction signal of the target region and a pixel signal of the target region, and a residual signal generated by the residual signal generation step A residual signal compression step for compressing the residual signal, a residual signal restoration step for producing a reproduction residual signal obtained by restoring the compressed data of the residual signal, and the motion information recording means A predicted motion vector similar to the motion vector of the target region selected from the existing reproduced motion vectors, additional information determined from the motion vector of the target region, and compressed data of the residual signal. An encoding step for encoding, the pixel signal of the target area is restored by adding the prediction signal and the reproduction residual signal, and the restored pixel signal is stored in the image recording unit as the already reproduced image An image recording step, wherein the motion vector includes a zeroth motion vector required for generating a zeroth prediction signal and a first motion vector required for generating a first prediction signal. The prediction signal generation step estimates a zeroth motion vector for obtaining the zeroth prediction signal, and a plurality of motions stored in the motion information recording means A 0th predicted motion for generating a 0th predicted motion vector similar to the estimated 0th motion vector using a motion vector selected from the vector and identifying the generated 0th predicted motion vector A zeroth motion information estimating step for generating zeroth additional information including an information index and a differential motion vector determined from the zeroth motion vector and the zeroth predicted motion vector; and the motion information recording means A first prediction motion vector that generates a first prediction signal having a high correlation with the target region is generated using a motion vector selected from a plurality of motion vectors stored in the first prediction, and the generated first prediction First additional information including a first predicted motion information index for identifying a motion vector is generated, and the first predicted motion vector is set as the first motion vector. A first motion information estimation step; and a prediction signal synthesis step of synthesizing the 0th prediction signal and the first prediction signal to generate a prediction signal of a target region, and the first motion information In the estimation step, the video predictive coding apparatus further estimates a second motion vector for acquiring the first prediction signal, detects a second motion vector, and estimates the second motion vector. A second prediction motion vector similar to the motion vector of the second motion vector is selected from a plurality of motion vectors stored in the motion information recording means, and a second prediction motion information index for identifying the selected prediction motion vector; Generating second additional information including the second motion vector and a differential motion vector determined from the second predicted motion vector, and performing the predicted signal generation step The moving image predictive encoding device uses the first motion vector when the plurality of already-reproduced images stored in the image recording means are all images past the encoding target image in the display order. When a first prediction signal is generated, and a plurality of previously reproduced images stored in the image recording means include images that are future in the display order from the image to be encoded, the second motion vector A first prediction signal is generated, and in the encoding step, the moving image predictive encoding device includes a plurality of already-reproduced images stored in the image recording unit that are all past images to be encoded in display order. In this case, for each screen or slice, the instruction information for instructing that the first additional information is encoded is encoded, and the plurality of already reproduced images stored in the image recording means are encoded in the display order. Future than images to be converted When an image is included, the instruction information for instructing that the second additional information is encoded is encoded for each screen or slice, and the second additional information is encoded based on the zeroth additional information and the instruction information. Either one additional information or second additional information may be encoded as additional information of each target region.

  In the encoding step, the video predictive encoding device may encode the instruction information by including the instruction information in a header of a screen or a slice.

  As another aspect, the video predictive decoding method is a video predictive decoding method executed by a video predictive decoding device, and is a decoding target from compressed data divided and encoded into a plurality of regions. A decoding step of decoding additional information of a decoding target region and compressed data of a residual signal, a motion information restoring step of restoring a motion vector required for generating a prediction signal of the target region from the additional information, and the motion vector A motion information recording step for storing in a motion information recording means; a motion compensation step for generating a prediction signal for the target region based on the motion vector; and a reproduction residual signal for the target region from the compressed data of the residual signal. A residual signal restoration step to restore, and a pixel signal in the decoding target area is restored by adding the prediction signal and the reproduction residual signal, and restored. An image recording step for storing the pixel signal as a reproduced image, and in the decoding step, the video predictive decoding device decodes the 0th additional information and the first additional information, and the 0th The additional information is the 0th motion vector for identifying the 0th differential motion vector and one selected from a plurality of motion vectors stored in the motion information recording means as the 0th motion vector predictor. A first predicted motion that identifies one selected from a plurality of motion vectors stored in the motion information recording means as a first predicted motion vector. The motion information restoring step includes a plurality of motion vectors stored in the motion information recording means based on the 0th predicted motion information index. The 0th motion vector is generated using the selected motion vector, and the 0th motion vector is restored by adding the generated 0th motion vector predictor and the 0th difference motion vector. Generating a first predicted motion vector using a motion vector selected from a plurality of motion vectors stored in the motion information recording unit based on the first predicted motion information index, and a motion information restoring step; A first motion information restoring step for restoring the generated first predicted motion vector as a first motion vector, and in the motion compensation step, the moving picture predictive decoding device includes the zeroth motion vector and Based on the first motion vector, two signals respectively obtained from the already reproduced image are combined to generate a prediction signal of the target region, and the decoding step The video predictive decoding apparatus further decodes instruction information indicating whether or not the first additional information includes a difference motion vector for each screen or slice, and the instruction information includes the first information When the one additional information indicates that the first difference motion vector is not included, the video predictive decoding apparatus decodes the first predicted motion information index as the first additional information, and the instruction information If the first additional information indicates that the first differential motion vector is included, the moving picture predictive decoding apparatus uses the differential motion vector and the first predicted motion information index as the first additional information. And in the first motion information restoration step, the video predictive decoding device, when the instruction information indicates that the first additional information does not include the first differential motion vector A first predicted motion vector is generated using a motion vector selected from a plurality of motion vectors stored in the motion information recording unit based on the first predicted motion information index, and the generated first prediction When the motion vector is restored as the first motion vector and the indication information indicates that the first additional information includes the first difference vector, the first motion vector is based on the first predicted motion information index. A predicted motion vector may be generated, and the generated first predicted motion vector and the decoded difference motion vector may be added to restore the first motion vector.

  In the decoding step, the video predictive decoding apparatus may decode the instruction information included in a screen or slice header.

  As another aspect, the moving image predictive encoding program includes: a region dividing unit that divides an input image into a plurality of regions; and a target region that is an encoding target among regions divided by the region dividing unit. A prediction signal generating means for determining a motion vector for acquiring a signal having a high correlation from the already reproduced image, a motion information recording means for storing the motion vector, a prediction signal for the target area, and a pixel signal for the target area A residual signal generating means for generating a residual signal, a residual signal compressing means for compressing the residual signal generated by the residual signal generating means, and a reproduction residual obtained by decompressing the compressed data of the residual signal. A residual signal restoring means for generating a difference signal, and a predicted motion vector similar to the motion vector of the target area selected from the previously reproduced motion vectors stored in the motion information recording means Adding the additional information determined from the motion vector of the target region and the compressed data of the residual signal, and adding the prediction signal and the reproduction residual signal A moving picture predictive encoding program for causing a pixel signal of a target region to be restored and functioning as an image recording unit that stores the restored pixel signal as the already-reproduced image, wherein the motion vector is a 0th prediction signal A first motion vector required for generating the first motion vector and a first motion vector required for generating the first prediction signal, the prediction signal generating means for acquiring the zeroth prediction signal Estimating the 0th motion vector and using the motion vector selected from a plurality of motion vectors stored in the motion information recording means; A similar 0th predicted motion vector is generated, and is determined from the 0th predicted motion information index for identifying the generated 0th predicted motion vector, and the 0th motion vector and the 0th predicted motion vector. Using a motion vector selected from a plurality of motion vectors stored in the motion information recording means, a 0th motion information estimating means for generating 0th additional information including a difference motion vector A first addition including a first predicted motion information index for generating a first predicted motion vector that generates a first predicted signal having a high correlation with a region and identifying the generated first predicted motion vector A first motion information estimating means for generating information and setting the first motion vector as the first motion vector; and combining the 0th prediction signal and the first prediction signal Prediction signal synthesis means for generating a prediction signal of the region, wherein the first motion information estimation means further estimates a second motion vector for obtaining the first prediction signal, and 2 motion vectors are detected, a second predicted motion vector similar to the estimated second motion vector is selected from a plurality of motion vectors stored in the motion information recording means, and the selected predicted motion vector is selected Generating second additional information including a second predicted motion information index for identifying the second motion vector and a differential motion vector determined from the second motion vector and the second predicted motion vector, and When the plurality of already-reproduced images stored in the image recording unit are all images past the encoding target image in the display order, the signal generating unit uses the first motion vector as the first motion vector. When a plurality of already-reproduced images stored in the image recording means include a future image in the display order from the encoding target image, the first motion vector is used as the first motion vector. In the case where the plurality of already-reproduced images stored in the image recording unit are all images past the encoding target image in the display order, for each screen or slice, When instruction information for instructing that the first additional information is encoded is encoded, and a plurality of already reproduced images stored in the image recording means include images of the future from the encoding target image in the display order. For each screen or slice, the instruction information for instructing that the second additional information is encoded is encoded, and the first additional information or the first additional information based on the zeroth additional information and the instruction information is encoded. Second additional information Re or one and may be encoded as additional information for each target region.

  The encoding means may encode the instruction information by including it in a screen or slice header.

  As another aspect, the moving picture predictive decoding program obtains additional information of a decoding target area to be decoded and compressed data of a residual signal from among compressed data encoded by dividing a computer into a plurality of areas. Decoding means for decoding, motion information restoring means for restoring a motion vector required for generating a prediction signal of the target region from the additional information, motion information recording means for saving the motion vector, and based on the motion vector, the motion vector Motion compensation means for generating a prediction signal of the target area, residual signal restoration means for restoring the reproduction residual signal of the target area from the compressed data of the residual signal, and the prediction signal and the reproduction residual signal. A moving image for functioning as an image recording means for restoring the pixel signal of the decoding target area by adding and saving the restored pixel signal as an already reproduced image A decoding / decoding program, wherein the decoding means decodes the 0th additional information and the first additional information, and the 0th additional information is stored in the 0th differential motion vector and the motion information recording means. A first predicted motion information index for identifying one selected from a plurality of motion vectors as a zeroth predicted motion vector, wherein the first additional information is the motion information recording means Including a first predicted motion information index that identifies one selected from a plurality of motion vectors stored in the first motion vector as a first predicted motion vector, wherein the motion information restoring means includes the 0th predicted motion information index. Is used to generate a 0th predicted motion vector using a motion vector selected from a plurality of motion vectors stored in the motion information recording means, and the generated 0th prediction And 0th motion information restoring means for restoring the 0th motion vector by adding the vector and the 0th differential motion vector, and storing in the motion information recording means based on the first predicted motion information index First motion information restoring means for generating a first predicted motion vector using a motion vector selected from a plurality of motion vectors that have been generated, and restoring the generated first predicted motion vector as a first motion vector And the motion compensation means synthesizes two signals respectively obtained from the already-reproduced images based on the 0th motion vector and the first motion vector to obtain a prediction signal of the target region. The decoding means further decodes instruction information indicating whether the first additional information includes a differential motion vector for each screen or slice, and the instruction information If the first additional information indicates that the first differential motion vector is not included, the first predicted motion information index is decoded as the first additional information, and the instruction information is the first additional information. When the additional information of 1 indicates that the first additional motion vector is included, the differential motion vector and the first predicted motion information index are decoded as the first additional information, and the first motion information restoration is performed. The means is stored in the motion information recording means based on the first predicted motion information index when the instruction information indicates that the first additional information does not include the first differential motion vector. A first predicted motion vector is generated using a motion vector selected from a plurality of motion vectors, the generated first predicted motion vector is restored as a first motion vector, and the instruction information includes When the additional information includes a first difference vector, a first predicted motion vector is generated based on the first predicted motion information index, and the generated first predicted motion vector and the decoded difference The first motion vector may be restored by adding the motion vector.

  The decoding means may decode the instruction information included in the screen or slice header.

  As another aspect, the video predictive coding apparatus has a correlation between a region dividing unit that divides an input image into a plurality of regions and a target region that is an encoding target among regions divided by the region dividing unit. A prediction signal generating means for determining a motion vector for acquiring a high signal from the already reproduced image; a motion information recording means for storing the motion vector; a residual signal of the prediction signal of the target area and the pixel signal of the target area; A residual signal generating means for generating a difference signal; a residual signal compressing means for compressing the residual signal generated by the residual signal generating means; and a reproduced residual signal obtained by restoring compressed data of the residual signal. A generated residual signal, a predicted motion vector similar to the motion vector of the target region selected from the previously reproduced motion vectors stored in the motion information recording unit, and a motion of the target region Reconstructing the pixel signal of the target area by adding the additional information determined from the vector, the encoding means for encoding the compressed data of the residual signal, and adding the prediction signal and the reproduction residual signal And an image recording means for storing the restored pixel signal as the already-reproduced image, wherein the motion vector includes a zeroth motion vector required for generating a zeroth prediction signal, and a first prediction signal. A first motion vector required for generating the first motion vector, and the prediction signal generation means estimates a zeroth motion vector for obtaining the zeroth prediction signal, and the motion information recording means Using a motion vector selected from a plurality of stored motion vectors, a 0th predicted motion vector similar to the estimated 0th motion vector is generated, and the generated 0th predicted motion vector is generated. 0th motion information for generating 0th additional information including a 0th predicted motion information index for identifying the first motion vector and a differential motion vector determined from the 0th motion vector and the 0th predicted motion vector A first prediction motion that generates a first prediction signal having a high correlation with a target region, using a motion information estimation unit and a motion vector selected from a plurality of motion vectors stored in the motion information recording unit; Generating a vector, generating first additional information including a first predicted motion information index for identifying the generated first predicted motion vector, and making the first predicted motion vector a first motion vector First motion information estimation means to be set, and prediction signal synthesis means for synthesizing the 0th prediction signal and the first prediction signal to generate a prediction signal of the target region may be included.

  Here, the 0th additional information further includes a reference screen index for identifying a reference screen related to generation of the 0th prediction signal, and the first additional information is the first prediction signal. A reference screen index for identifying a reference screen related to the generation of the first prediction signal, the motion information recording means, a reference screen index for identifying a reference screen related to the generation of the 0th prediction signal, and the first A reference screen index for identifying a reference screen related to the generation of one prediction signal, and the zeroth predicted motion information index further includes a plurality of reference screen indexes stored in the motion information recording means. To the 0th predicted motion vector and the 0th predicted reference screen index, and the first predicted motion information index further includes the motion information recording The first prediction reference screen index is identified together with the first predicted motion vector from the plurality of reference screen indexes stored in the reference screen, and the zeroth motion information restoring means is a reference screen included in the zeroth additional information. Based on the index and the 0th prediction reference screen index, the 0th prediction motion vector is generated using a motion vector selected from the motion information recording unit based on the 0th prediction motion information index The first motion information restoring means is further based on the first predicted motion information index based on the reference screen index and the first predicted reference screen index included in the first additional information. The first predicted motion vector may be generated using a motion vector selected from the motion information recording means.

  As another aspect, the moving picture predictive decoding apparatus decodes the additional information of the decoding target area to be decoded and the compressed data of the residual signal from the compressed data divided and encoded into a plurality of areas. Means, motion information restoration means for restoring a motion vector required for generating a prediction signal of the target area from the additional information, motion information recording means for saving the motion vector, and the target area based on the motion vector Adding a motion compensation means for generating a prediction signal, a residual signal restoration means for restoring the reproduction residual signal of the target area from the compressed data of the residual signal, and adding the prediction signal and the reproduction residual signal. And an image recording unit that restores the pixel signal of the decoding target area and stores the restored pixel signal as an already reproduced image, wherein the decoding unit includes the zeroth additional information and the first information Additional information is decoded, and the 0th additional information is a 0th predicted motion vector obtained by selecting one of a 0th differential motion vector and a plurality of motion vectors stored in the motion information recording means. A first predicted motion information index for identifying the first additional information as one selected from a plurality of motion vectors stored in the motion information recording means. Including a first predicted motion information index to be identified as a motion vector, wherein the motion information restoring means is selected from a plurality of motion vectors stored in the motion information recording means based on the zeroth predicted motion information index The motion vector is used to generate a 0th motion vector predictor and the 0th motion vector predictor and the 0th motion vector difference are added to generate a motion vector 0th. Using a motion vector selected from a plurality of motion vectors stored in the motion information recording unit based on the first predicted motion information index, First motion information restoring means for generating a predicted motion vector and restoring the generated first predicted motion vector as a first motion vector, wherein the motion compensation means includes the zeroth motion vector and the first motion vector. A prediction signal for the target region may be generated by combining two signals respectively acquired from the already reproduced image based on one motion vector.

  Here, the 0th additional information further includes a reference screen index for identifying a reference screen related to the generation of the 0th prediction signal, and the first additional information is the generation of the first prediction signal. A reference screen index for identifying a reference screen related to the reference information, and the motion information recording means includes a reference screen index for identifying a reference screen related to generation of the 0th prediction signal, and the first screen A reference screen index for identifying a reference screen related to generation of a prediction signal is further stored, and the zeroth predicted motion information index is further stored in a plurality of reference screen indexes stored in the motion information recording means. A 0th prediction reference screen index is identified together with a 0 predicted motion vector, and the first predicted motion information index is further stored in the motion information recording means A first predicted reference screen index together with a first predicted motion vector is identified from a plurality of reference screen indexes, and the zeroth motion information restoring means includes a reference screen index included in the zeroth additional information, Further, based on the 0th prediction reference screen index, generating the 0th prediction motion vector using a motion vector selected from the motion information recording unit based on the 0th prediction motion information index, The first motion information restoration means further includes the reference motion index based on the first predicted motion information index based on the reference screen index and the first predicted reference screen index included in the first additional information. The first predicted motion vector may be generated using a motion vector selected from the information recording means.

  As another aspect, the video predictive encoding method is a video predictive encoding method executed by a video predictive encoding device, and includes an area dividing step of dividing an input image into a plurality of areas, and the area dividing A prediction signal generating step for determining a motion vector for acquiring a signal having a high correlation with the target region to be encoded among the regions divided by the step from the already reproduced image; and the motion vector as motion information recording means. A motion information recording step, a residual signal generation step for generating a residual signal between the prediction signal of the target region and a pixel signal of the target region, and a residual signal generated by the residual signal generation step A residual signal compression step for compressing the residual signal, a residual signal restoration step for producing a reproduction residual signal obtained by restoring the compressed data of the residual signal, and the motion information recording means A predicted motion vector similar to the motion vector of the target region selected from the existing reproduced motion vectors, additional information determined from the motion vector of the target region, and compressed data of the residual signal. An encoding step for encoding, the pixel signal of the target area is restored by adding the prediction signal and the reproduction residual signal, and the restored pixel signal is stored in the image recording unit as the already reproduced image An image recording step, wherein the motion vector includes a zeroth motion vector required for generating a zeroth prediction signal and a first motion vector required for generating a first prediction signal. The prediction signal generation step estimates a zeroth motion vector for obtaining the zeroth prediction signal, and a plurality of motions stored in the motion information recording means A 0th predicted motion for generating a 0th predicted motion vector similar to the estimated 0th motion vector using a motion vector selected from the vector and identifying the generated 0th predicted motion vector A zeroth motion information estimating step for generating zeroth additional information including an information index and a differential motion vector determined from the zeroth motion vector and the zeroth predicted motion vector; and the motion information recording means A first prediction motion vector that generates a first prediction signal having a high correlation with the target region is generated using a motion vector selected from a plurality of motion vectors stored in the first prediction, and the generated first prediction First additional information including a first predicted motion information index for identifying a motion vector is generated, and the first predicted motion vector is set as the first motion vector. A first motion information estimation step and a prediction signal synthesis step of generating a prediction signal of the target region by synthesizing the 0th prediction signal and the first prediction signal may be included.

  Here, in the 0th motion information estimation step, the moving picture predictive coding apparatus further includes the reference screen index for identifying a reference screen related to generation of the 0th prediction signal. The additional information is generated, and in the first motion information estimation step, the video predictive encoding device further includes a reference screen index for identifying a reference screen related to generation of the first prediction signal. 1st additional information is produced | generated, In the said motion information recording step, the said moving image predictive coding apparatus WHEREIN: The reference screen index for identifying the reference screen which concerns on the production | generation of the said 0th prediction signal, and said 1st And a reference screen index for identifying a reference screen related to the generation of the predicted signal is stored in the motion information recording means, and the zeroth predicted motion information index is Furthermore, the 0th prediction reference screen index is identified together with the 0th prediction motion vector from a plurality of reference screen indexes stored in the motion information recording means, and the first prediction motion information index is further recorded in the motion information recording The first predicted reference screen index may be identified together with the first predicted motion vector from a plurality of reference screen indexes stored in the means.

  As another aspect, the video predictive decoding method is a video predictive decoding method executed by a video predictive decoding device, and is a decoding target from compressed data divided and encoded into a plurality of regions. A decoding step of decoding additional information of a decoding target region and compressed data of a residual signal, a motion information restoring step of restoring a motion vector required for generating a prediction signal of the target region from the additional information, and the motion vector A motion information recording step for storing in a motion information recording means; a motion compensation step for generating a prediction signal for the target region based on the motion vector; and a reproduction residual signal for the target region from the compressed data of the residual signal. A residual signal restoration step to restore, and a pixel signal in the decoding target area is restored by adding the prediction signal and the reproduction residual signal, and restored. An image recording step for storing the pixel signal as a reproduced image, and in the decoding step, the video predictive decoding device decodes the 0th additional information and the first additional information, and the 0th The additional information is the 0th motion vector for identifying the 0th differential motion vector and one selected from a plurality of motion vectors stored in the motion information recording means as the 0th motion vector predictor. A first predicted motion that identifies one selected from a plurality of motion vectors stored in the motion information recording means as a first predicted motion vector. The motion information restoring step includes a plurality of motion vectors stored in the motion information recording means based on the 0th predicted motion information index. The 0th motion vector is generated using the selected motion vector, and the 0th motion vector is restored by adding the generated 0th motion vector predictor and the 0th difference motion vector. Generating a first predicted motion vector using a motion vector selected from a plurality of motion vectors stored in the motion information recording unit based on the first predicted motion information index, and a motion information restoring step; A first motion information restoring step for restoring the generated first predicted motion vector as a first motion vector, and in the motion compensation step, the moving picture predictive decoding device includes the zeroth motion vector and Two signals acquired from the already reproduced image based on the first motion vector may be combined to generate a prediction signal for the target region.

  Here, the 0th additional information further includes a reference screen index for identifying a reference screen related to the generation of the 0th prediction signal, and the first additional information is the generation of the first prediction signal. A reference screen index for identifying the reference screen according to the reference information, and in the motion information recording step, the moving picture predictive decoding device refers to the reference screen for identifying the reference screen related to the generation of the zeroth prediction signal. A screen index and a reference screen index for identifying a reference screen related to generation of the first prediction signal are further stored in the motion information recording unit, and the 0th predicted motion information index is further stored in the motion information. The 0th prediction reference screen index is identified together with the 0th prediction motion vector from the plurality of reference screen indexes stored in the recording means, and the first prediction screen index is identified. The motion information index further identifies the first predicted reference screen index together with the first predicted motion vector from a plurality of reference screen indexes stored in the motion information recording means, and in the 0th motion information restoration step, The moving picture predictive decoding device further records the motion information based on the 0th predicted motion information index based on the reference screen index and the 0th predicted reference screen index included in the 0th additional information. Generating the 0th motion vector predictor using the motion vector selected from the means, and in the first motion information restoration step, the video predictive decoding device includes a reference screen included in the first additional information. Based on the index and the first prediction reference screen index, the first prediction motion information Index may generate the first prediction motion vector using the motion vectors selected from the motion information recording means on the basis of.

  As another aspect, the moving image predictive encoding program includes: a region dividing unit that divides an input image into a plurality of regions; and a target region that is an encoding target among regions divided by the region dividing unit. A prediction signal generating means for determining a motion vector for acquiring a signal having a high correlation from the already reproduced image, a motion information recording means for storing the motion vector, a prediction signal for the target area, and a pixel signal for the target area A residual signal generating means for generating a residual signal, a residual signal compressing means for compressing the residual signal generated by the residual signal generating means, and a reproduction residual obtained by decompressing the compressed data of the residual signal. A residual signal restoring means for generating a difference signal, and a predicted motion vector similar to the motion vector of the target area selected from the previously reproduced motion vectors stored in the motion information recording means Adding the additional information determined from the motion vector of the target region and the compressed data of the residual signal, and adding the prediction signal and the reproduction residual signal A moving picture predictive encoding program for causing a pixel signal of a target region to be restored and functioning as an image recording unit that stores the restored pixel signal as the already-reproduced image, wherein the motion vector is a 0th prediction signal A first motion vector required for generating the first motion vector and a first motion vector required for generating the first prediction signal, the prediction signal generating means for acquiring the zeroth prediction signal Estimating the 0th motion vector and using the motion vector selected from a plurality of motion vectors stored in the motion information recording means; A similar 0th predicted motion vector is generated, and is determined from the 0th predicted motion information index for identifying the generated 0th predicted motion vector, and the 0th motion vector and the 0th predicted motion vector. Using a motion vector selected from a plurality of motion vectors stored in the motion information recording means, a 0th motion information estimating means for generating 0th additional information including a difference motion vector A first addition including a first predicted motion information index for generating a first predicted motion vector that generates a first predicted signal having a high correlation with a region and identifying the generated first predicted motion vector A first motion information estimating means for generating information and setting the first motion vector as the first motion vector; and combining the 0th prediction signal and the first prediction signal Prediction signal synthesis means for generating a prediction signal for the region.

  Here, the 0th additional information further includes a reference screen index for identifying a reference screen related to generation of the 0th prediction signal, and the first additional information is the first prediction signal. A reference screen index for identifying a reference screen related to the generation of the first prediction signal, the motion information recording means, a reference screen index for identifying a reference screen related to the generation of the 0th prediction signal, and the first A reference screen index for identifying a reference screen related to the generation of one prediction signal, and the zeroth predicted motion information index further includes a plurality of reference screen indexes stored in the motion information recording means. To the 0th predicted motion vector and the 0th predicted reference screen index, and the first predicted motion information index further includes the motion information recording The first prediction reference screen index is identified together with the first predicted motion vector from the plurality of reference screen indexes stored in the reference screen, and the zeroth motion information restoring means is a reference screen included in the zeroth additional information. Based on the index and the 0th prediction reference screen index, the 0th prediction motion vector is generated using a motion vector selected from the motion information recording unit based on the 0th prediction motion information index The first motion information restoring means is further based on the first predicted motion information index based on the reference screen index and the first predicted reference screen index included in the first additional information. The first predicted motion vector may be generated using a motion vector selected from the motion information recording means.

  As another aspect, the moving picture predictive decoding program obtains additional information of a decoding target area to be decoded and compressed data of a residual signal from among compressed data encoded by dividing a computer into a plurality of areas. Decoding means for decoding, motion information restoring means for restoring a motion vector required for generating a prediction signal of the target region from the additional information, motion information recording means for saving the motion vector, and based on the motion vector, the motion vector Motion compensation means for generating a prediction signal of the target area, residual signal restoration means for restoring the reproduction residual signal of the target area from the compressed data of the residual signal, and the prediction signal and the reproduction residual signal. A moving image for functioning as an image recording means for restoring the pixel signal of the decoding target area by adding and saving the restored pixel signal as an already reproduced image A decoding / decoding program, wherein the decoding means decodes the 0th additional information and the first additional information, and the 0th additional information is stored in the 0th differential motion vector and the motion information recording means. A first predicted motion information index for identifying one selected from a plurality of motion vectors as a zeroth predicted motion vector, wherein the first additional information is the motion information recording means Including a first predicted motion information index that identifies one selected from a plurality of motion vectors stored in the first motion vector as a first predicted motion vector, wherein the motion information restoring means includes the 0th predicted motion information index. Is used to generate a 0th predicted motion vector using a motion vector selected from a plurality of motion vectors stored in the motion information recording means, and the generated 0th prediction And 0th motion information restoring means for restoring the 0th motion vector by adding the vector and the 0th differential motion vector, and storing in the motion information recording means based on the first predicted motion information index First motion information restoring means for generating a first predicted motion vector from the plurality of motion vectors that have been generated and restoring the generated first predicted motion vector as a first motion vector, the motion compensation The means may generate a prediction signal of the target region by combining two signals respectively acquired from the already-reproduced image based on the 0th motion vector and the first motion vector.

  Here, the 0th additional information further includes a reference screen index for identifying a reference screen related to the generation of the 0th prediction signal, and the first additional information is the generation of the first prediction signal. A reference screen index for identifying a reference screen related to the reference information, and the motion information recording means includes a reference screen index for identifying a reference screen related to generation of the 0th prediction signal, and the first screen A reference screen index for identifying a reference screen related to generation of a prediction signal is further stored, and the zeroth predicted motion information index is further stored in a plurality of reference screen indexes stored in the motion information recording means. A 0th prediction reference screen index is identified together with a 0 predicted motion vector, and the first predicted motion information index is further stored in the motion information recording means A first predicted reference screen index together with a first predicted motion vector is identified from a plurality of reference screen indexes, and the zeroth motion information restoring means includes a reference screen index included in the zeroth additional information, Further, based on the 0th prediction reference screen index, generating the 0th prediction motion vector using a motion vector selected from the motion information recording unit based on the 0th prediction motion information index, The first motion information restoration means further includes the reference motion index based on the first predicted motion information index based on the reference screen index and the first predicted reference screen index included in the first additional information. The first predicted motion vector may be generated using a motion vector selected from the information recording means.

  According to the video predictive encoding device, video predictive encoding method, video predictive encoding program, video predictive decoding device, video predictive decoding method, and video predictive decoding program of the present invention, encoded motion Since one motion vector effective for bi-prediction can be indicated based on information, there is an effect of improving bi-prediction performance with a small amount of code.

It is a block diagram which shows the moving image predictive coding apparatus which concerns on this embodiment. It is a block diagram explaining the prediction signal generator shown in FIG. It is a flowchart explaining the 1st motion estimation process shown in FIG. It is a flowchart explaining the 0th motion estimation process shown in FIG. It is a schematic diagram explaining the example of the encoding order of a screen. It is a figure explaining the example of a reference screen list. It is a figure explaining the example of an adjacent block. It is a figure explaining another example of an adjacent block. It is a flowchart which shows the procedure of the moving image predictive encoding method of the moving image predictive encoding apparatus shown in FIG. It is a block diagram which shows the moving image predictive decoding apparatus which concerns on this embodiment. It is a block diagram explaining the motion information decompression device shown in FIG. It is a flowchart explaining the 1st motion information restoration process shown in FIG. It is a flowchart explaining the 0th motion information restoration process shown in FIG. It is a flowchart which shows the procedure of the moving image predictive decoding method of the moving image predictive decoding apparatus shown in FIG. It is a block diagram which shows the program which can perform the moving image predictive coding method which concerns on this embodiment. It is a block diagram which shows the program which can perform the moving image predictive decoding method which concerns on this embodiment. It is a figure which shows the hardware constitutions of the computer for performing the program recorded on the recording medium. It is a perspective view of a computer for executing a program stored in a recording medium. It is a schematic diagram for demonstrating the motion estimation process in the prediction between screens. It is a schematic diagram for demonstrating the conventional intra prediction method.

  Hereinafter, this embodiment will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same or equivalent elements are denoted by the same reference numerals, and redundant description is omitted. Further, “frame”, “screen”, and “image” (501 to 511 in FIG. 5) constituting a moving image have the same meaning in the description in this specification.

  FIG. 1 is a block diagram showing a video predictive encoding device 100 according to the present embodiment. This video predictive coding apparatus 100 includes an input terminal 101, a block divider 102, a prediction signal generator 103, a frame memory 104, a subtractor 105, a converter 106, a quantizer 107, an inverse quantizer 108, and an inverse transform. And an adder 110, an encoder 111, an output terminal 112, and a motion information memory 113. The converter 106 and the quantizer 107 function as residual signal compression means, the inverse quantizer 108 and the inverse transformer 109 function as residual signal restoration means, and the motion information memory functions as motion information recording means. . Further, the motion information memory 113 may be included in the prediction signal generator 103.

  The input terminal 101 is a terminal for inputting a moving image signal composed of a plurality of images.

  The block divider 102 divides an image to be encoded, which is represented by a signal input from the input terminal 101, into a plurality of regions (target block, target region). In this embodiment, the block is divided into 8 × 8 pixels, but may be divided into other block sizes or shapes. Further, blocks of different sizes may be mixed in the screen.

  The prediction signal generator 103 detects motion information necessary for generating a prediction signal of each prediction block in the target block and generates a prediction signal. Also, additional information necessary for restoring the motion information is generated by the decoding device. The method for generating the prediction signal is not limited in the present invention, but inter-screen prediction (single prediction, bi-prediction) and intra-screen prediction (not shown for intra-screen prediction) as described in the background art are applicable.

  In the present embodiment, a prediction signal is generated by bi-prediction. One motion information of bi-prediction is acquired via L104 so that the absolute value error sum between the original signal of the target block input via L102 and the bi-prediction signal is minimized by the block matching shown in FIG. Detection is performed using the image signal. The other motion information is generated based on the encoded motion information.

  In this embodiment, in order to explain bi-prediction, motion information is composed of 0th motion information and first motion information, and a reference screen index (ref_idx [0] or ref_idx [1]) and a motion vector, respectively. (Mv [0] [0/1] or mv [1] [0/1]). Candidates for the reference screen for the 0th motion information are indicated by List0 in FIG. 6, and candidates for the reference screen for the first motion information are indicated by List1 in FIG. Here, [0/1] is information for identifying the horizontal component and the vertical component of the vector. Hereinafter, the description of [0/1] is omitted (the same is omitted in the drawings).

  Note that the playback image registered in the reference screen list shown in FIG. 6 may be automatically determined according to a predetermined rule, or may be explicitly encoded in frame units or sequence units. At this time, a frame number can be used to identify each reference screen as shown in FIGS.

  The motion information generated by the prediction signal generator 103 is output to the motion information memory 113 via the L103b.

  The motion information memory 113 stores the input motion information. The stored motion information is input to the prediction signal generator via L113, and is used for encoding motion information of the subsequent block.

  The additional information generated by the prediction signal generator 103 is output to the encoder 111 via the L103c.

  The prediction signal generated by the prediction signal generator 103 is output to the subtracter 105 and the adder 110 via L103a.

  The subtractor 105 subtracts the prediction signal for the target block input via the line L103a from the pixel signal of the target block divided and input by the block divider 102 input via the line L102, Generate a residual signal. The subtractor 105 outputs the residual signal obtained by the subtraction to the converter 106 via the line L105.

  The converter 106 is a part that performs discrete cosine transform on the input residual signal. The quantizer 107 is a part that quantizes the transform coefficient that has been discrete cosine transformed by the transformer 106.

  The encoder 111 entropy-encodes the additional information input from the prediction signal generator and the quantized transform coefficient input from the quantizer 107, and the encoded data is output to the output terminal 112 via L111. The entropy encoding method is not limited, but arithmetic encoding, variable length encoding, and the like can be applied.

  The output terminal 112 collectively outputs information input from the encoder 111 and outputs the information to the outside.

  The inverse quantizer 108 inversely quantizes the quantized transform coefficient. The inverse transformer 109 restores the residual signal by inverse discrete cosine transform. The adder 110 adds the restored residual signal and the prediction signal input via the L 103 a, reproduces the signal of the encoding target block, and stores it in the frame memory 104. In the present embodiment, the converter 106 and the inverse converter 109 are used, but other conversion processes in place of these converters may be used. Further, the converter 106 and the inverse converter 109 are not essential. Thus, in order to use it for the prediction signal production | generation of a subsequent encoding object block, the reproduced signal of the encoding object block encoded is decompress | restored by reverse processing, and is memorize | stored in the frame memory 104. FIG.

  Next, the prediction signal generator 103 will be described in detail. First, motion information, predicted motion information, and additional information will be described.

  As shown above, motion information in bi-prediction is composed of 0th motion information and 1st motion information, and a reference screen index (ref_idx [0] or ref_idx [1]) and a motion vector (mv [ 0] or mv [1]). Candidates for the reference screen for the 0th motion information are indicated by List0 in FIG. 6, and candidates for the reference screen for the first motion information are indicated by List1 in FIG.

  In the bi-prediction of the present embodiment, the prediction signal generator 103 uses already encoded motion information as predicted motion information. The encoded motion information includes motion information associated with the encoded adjacent block and encoded motion information of the target region. Note that the motion information associated with the adjacent block refers to the motion information used for generating the prediction signal when the adjacent block is the encoding target, and is stored in the motion information memory 113.

  The predicted motion information is also composed of 0th predicted motion information and first predicted motion information, and each includes a reference screen index (ref_idx [0] or ref_idx [1]) and a motion vector (mv [0] or mv [1). ])including. Candidates for the reference screen for the 0th predicted motion information are indicated by List0 in FIG. 6, and candidates for the reference screen for the first predicted motion information are indicated by List1 in FIG.

  A specific method of using the predicted motion information is based on the 0th predicted motion information when the motion vector of the 0th motion information detected by block matching with reference to the reference screen of List0 is differentially encoded. Thus, a predicted motion vector is generated. Also, first motion information using the reference screen of List1 is generated based on the first predicted motion information.

  An example of the predicted motion information will be described with reference to FIG. A block 400 shown in FIG. 7 is a target block, and motion information accompanying blocks 401 to 404 adjacent thereto is a candidate for predicted motion information. The motion information of the adjacent block includes 0th motion information and 1st motion information, respectively. Both may be candidates for predicted motion information, or may be limited to one (when the Nth motion information is predicted, only the Nth motion vector of an adjacent block is a candidate).

  A block 410 indicates a block (co-located block) at the same spatial position as the block 400 in the reference screen. Motion information associated with the block 410 and blocks 411 to 415 adjacent to the block 410 is a candidate for predicted motion information. n indicates a number for identifying a candidate for prediction motion information, and the selected candidate is indicated by a prediction motion information index (mvp_idx [0] or mvp_idx [1]). In this embodiment, since the 0th motion information is encoded first, the 0th motion information associated with the block 400 can be used as the first predicted motion information (in the example, n = 4).

  It should be noted that the position and number of the predicted motion information candidate may be determined in advance on the encoding side and the decoding side, and is not limited in the present invention. The number of prediction motion information candidates may be determined in advance on the encoding side and the decoding side, or may be encoded.

  In addition, when the reference screen identified by the ref_idx of the predicted motion information is different from the reference screen identified by the ref_idx of the target block, the predicted motion is based on the frame numbers of the encoding target screen and the two reference screens. You may perform the scaling process of the motion vector contained in information. That is, the motion vector included in the predicted motion information is scaled and converted into a motion vector that indicates the reference screen identified by the target block, and the converted motion vector is converted into the predicted motion vector (pmv [0] [0/1 ] Or pmv [1] [0/1]). At this time, the reference screen index (ref_idx [0] or ref_idx [1]) included in the predicted motion information is updated to the reference screen index (ref_idx [0] or ref_idx [1]) of the target block. Here, [0/1] is information for identifying the horizontal component and the vertical component of the vector. Hereinafter, the description of [0/1] is omitted (the same is omitted in the drawings).

  The additional information includes 0th additional information and 1st additional information. The 0th additional information includes ref_idx [0], a differential motion vector (mvd [0] [0/1] = mv [0] [0/1] −pmv [0] [0/1]), and mvp_idx [0 ]including. The first additional information includes ref_idx [1] and mvp_idx [1]. In this embodiment, since mv [1] [0/1] = pmv [1] [0/1], the vector value of mvd [1] [0/1] is always 0. Since mvd [1] [0/1] can be restored on the decoding side without being encoded, it is not necessary to include it in the first additional information. Here, [0/1] is information for identifying the horizontal component and the vertical component of the vector. Hereinafter, the description of [0/1] is omitted (the same is omitted in the drawings).

  FIG. 2 is a block diagram illustrating a configuration of the prediction signal generator 103 according to the present embodiment. The prediction signal generator 103 includes a first motion information estimator 121, a zeroth motion information estimator 122, and a prediction signal synthesizer 123.

  The first motion information estimator 121 uses a reference screen of List1 input via L104, and uses a plurality of first prediction motion information candidates input via L113 (according to the reference screen index). The first predicted motion information candidate that generates the first predicted signal most similar to the original signal of the target block input via L102). And a reference screen index pair. The first prediction signal is output to the prediction signal synthesizer 123 via L121a, and the first motion information generated based on the set of the selected first prediction motion information and the reference screen index is represented by L121b and L103b1. And output to the 0th motion information estimator 122 and the motion information memory 113 respectively. Further, the first additional information is generated and output to the encoder 111 via L103c1.

  The 0th motion information estimator 122 generates a first prediction signal using the first motion information input from L121b and the reference screen of List1 input via L104. Then, the reference screen of List0 input via L104 is searched, the bi-prediction signal generated from the first prediction signal and the searched candidate for the 0th prediction signal, and the object input via L102 The 0th motion information that minimizes the sum of absolute value differences from the original signal of the block is detected. The 0th prediction signal generated from the detected 0th motion information is output to the prediction signal synthesizer 123 via L122. Also, the 0th motion information is output to the motion information memory 113 via L103b2. Furthermore, 0th additional information is generated and output to the encoder 111 via the L103c2.

  Note that the processing by the 0th motion information estimator 122 may be executed first to obtain the 0th motion information and the 0th additional information before the 0th prediction signal. In this case, the 0th motion information that minimizes the sum of absolute difference between the predicted signal generated by the searched candidate for the 0th predicted signal and the original signal of the target block input via L102 is obtained. To detect. Then, the first motion information estimator 121 is executed using the 0th prediction signal. That is, by using the reference screen of List1 input via L104, a plurality of first prediction motion information candidates input via L113 (predicted by scaling the motion vector according to the reference screen index) The first prediction signal candidate is generated from the motion vector, and the bi-prediction signal generated by the 0th prediction signal and the first prediction signal candidate is input via L102. A set of the first predicted motion information candidate and the reference screen index is selected so as to be most similar to the first signal. This modification can be performed by inputting the 0th motion information to the first motion information estimator 121.

  The prediction signal synthesizer 123 averages the first prediction signal and the zeroth prediction signal input via L121a and L122 to generate a prediction signal of the target block, and subtracter 105 and adder via L103a. 110 is output.

  FIG. 3 shows a flowchart of the first motion information estimator 121. First, in step S301, the number of List1 reference screens used for prediction of the target block is set to M (M = 4 in FIGS. 6A and 6C, M = 2 in FIG. 6B), and A reference screen index ref_idx [1] of List1 included in the first motion information is set to 0, and a count m of the number of reference screens of List1 is initialized to 0. Further, the motion vector evaluation value D is set to the Max value. Next, in step S302, the number of motion vector predictor candidates is N (N = 11 in FIG. 7, provided that the first motion information estimator 121 is implemented before the zeroth motion information estimator 122). , N = 4 is skipped because the 0th motion information of the block 400 has not been determined), and the predicted motion information index mvp_idx [1] included in the first additional information is set to 0, The count n of the number of predicted motion information is initialized to 0.

  Next, in step S303, the motion vector of the predicted motion vector candidate n is acquired from the motion vectors of adjacent blocks, and in step S304, the nth predicted motion vector pmv [1] [m] [n] [0 / 1] is generated (here, [0/1] is information for identifying the horizontal and vertical components of the vector. Hereinafter, the description of [0/1] is omitted, and is similarly omitted in the drawings). At this time, the motion vector of the adjacent block is scaled according to the distance between the target screen and the reference screen (frame number identified by the reference screen index) to obtain a predicted motion vector. Thereafter, in step S305, a prediction signal of the target block is generated based on the mth reference screen and the nth scaled predicted motion vector (pmv [1] [m] [n]). Thus, it is determined whether or not the sum of absolute values of the difference signal between the generated prediction signal and the original signal of the target block is smaller than the motion vector evaluation value D. If the absolute value sum is greater than or equal to the motion vector evaluation value D, the process proceeds to step S308. When the absolute value sum is smaller than the motion vector evaluation value D, the process proceeds to step S307, where the predicted motion information index mvp_idx [1] included in the first additional information is updated to n, and the reference screen index is ref_idx [1]. Is updated to m, and the motion vector evaluation value D is updated to the sum of absolute values of the difference signal between the prediction signal calculated in step S306 and the original signal of the target block. Furthermore, the motion vector mv [1] included in the first motion information is set to the predicted motion vector pmv [1] [ref_idx [1]] [mvp_idx [1]], and the reference screen index is set to ref_idx [1]. . Thereafter, the process proceeds to step S308.

  In step S308, it is determined whether or not the value of n is smaller than N. If n is smaller than N, the process proceeds to step S309. If n reaches N, the process proceeds to step S310. In step S309, 1 is added to the value of n, and the process returns to step S303. Thereafter, steps S303 to S307 are repeated until n reaches N. In step S310, it is determined whether the value of m is smaller than M. If m is smaller than M, the process proceeds to step S311, 1 is added to the value of m, and the process returns to step S302. Thereafter, steps S302 to S309 are repeated until m reaches M. When m reaches M, the first additional information (ref_idx [1], mvp_idx [1]) is output to the encoder 111 in step S312, and the first motion information (ref_idx [1]) is output. And mv [1]) are stored in the motion information memory 113 and output to the 0th motion information estimator 122, and the process ends.

  FIG. 4 shows a flowchart of the 0th motion information estimator 122. First, in step S351, a first prediction signal in bi-prediction is generated based on the first motion information. Next, in step S352, the number of List0 reference screens used for prediction of the target block is set to M (M = 4 in FIGS. 6A and 6C, M = 2 in FIG. 6B)). At the same time, the reference screen index ref_idx [0] of List0 included in the 0th motion information is set to 0, and the count m of the number of reference screens of List0 is initialized to 0. Further, the motion vector evaluation value D is set to the Max value. Next, in step S353, a predicted motion information index mvp_idx [0] for identifying a predicted motion vector used when differentially encoding the 0th motion vector from a plurality of candidates is determined. As the selection method, for example, the method shown in steps S303 to S309 in FIG. 3 can be used. Then, a predicted motion vector candidate pmv [0] [m] [n] is generated. At this time, as described in step S304 in FIG. 3, the motion vector of the adjacent block is scaled according to the distance between the target screen and the reference screen (the frame number identified by the reference screen index) to obtain a predicted motion vector. .

  Next, in step S354, the bi-prediction signal obtained by acquiring the reference screen indicated by ref_idx [0] stored in the frame memory 104 and averaging it together with the first prediction signal; The 0th motion vector mv [0] that minimizes the sum of absolute values of the difference signal from the original signal is searched. Subsequently, in step S355, a zeroth differential motion vector mvd [0] (= mv [0] −pmv [0] [m] [n]) is generated. Subsequently, in step S356, the absolute value sum of the difference signal between the generated bi-predicted signal and the original signal of the target block and the code amount evaluation value of the 0th additional information (mvd [0], m, and n) ( The sum of λ (QP) × (mvd, m, n code amount), λ is a weight vector determined by a parameter QP that determines quantization accuracy when the transform coefficient obtained by transforming the prediction error signal is quantized is a motion vector. It is determined whether or not the evaluation value D is smaller. If the absolute value sum + code amount evaluation value is greater than or equal to the motion vector evaluation value D, the process proceeds to step S358. If the absolute value sum + the code amount evaluation value is smaller than the motion vector evaluation value D, the process proceeds to step S357, the predicted motion information index mvp_idx [0] included in the 0th additional information is updated to n, and the reference screen index Ref_idx [0] is updated to m, the differential motion vector mvd [0] is updated to (mv [0] −pmv [0] [ref_idx [1]] [mvp_idx [1]]), and D is updated in step S356. The absolute value sum of the difference signal between the bi-predicted signal calculated in step 1 and the original signal of the target block + the code amount evaluation value is updated. Further, the motion vector mv [0] included in the 0th motion information is updated. Thereafter, the process proceeds to step S358.

  In step S358, it is determined whether or not the value of m is smaller than M. If m is smaller than M, the process proceeds to step S359, 1 is added to the value of m, and the process returns to step S353. Thereafter, steps S353 to S359 are repeated until m reaches M. When m reaches M, in step S360, the 0th additional information (ref_idx [0], mvd [0], mvp_idx [0]) is output to the encoder 111 and the 0th motion Information (ref_idx [0] and mv [0]) is stored in the motion information memory 113, and the process ends.

  Note that the 0th motion information estimator 122 may be executed first to obtain the 0th motion information and the 0th additional information before the 0th prediction signal. In this case, step S351 in FIG. 4 is omitted, and in step S356, the absolute value sum of the difference signal between the 0th prediction signal and the original signal is obtained instead of the bi-prediction signal. In FIG. 3, the 0th motion information indicated by n = 4 in FIG. 7 can be used as a candidate for predicted motion information. A step of generating a 0th prediction signal in bi-prediction based on the 0th motion information is added, and the generated first prediction signal and the 0th prediction signal are averaged in step S306. This modification can be realized by changing to calculate the sum of absolute values of the difference signal between the bi-prediction signal and the original prediction signal.

  In this way, a first motion vector of bi-prediction is generated based on already encoded motion information, and a reference screen index and prediction are made from a plurality of candidates as shown in FIGS. 6 and 7 instead of the motion vector. By encoding the identification information for identifying the motion information, it is possible to generate one similar signal that is similar to the signal of the target block in the texture region with high randomness and the flat region including noise with a small code amount. Furthermore, search is performed on the reference screen, and the 0th motion vector that generates the second similar signal is detected and encoded, so that both motion vectors are obtained from the encoded motion information. Thus, an effect of enhancing the smoothing effect of the bi-prediction signal can be expected.

  FIG. 9 is a flowchart showing the procedure of the video predictive coding method in the video predictive coding device 100 according to the present embodiment. First, the input image is divided into 8 × 8 encoded blocks (other block sizes or shapes by the block divider 102. Also, blocks of different sizes may be mixed in the screen. .).

  First, the first motion information estimator 121 that constitutes the prediction signal generator 103 uses the reference screen of List1 obtained from the frame memory 104 and the first prediction motion information candidate obtained from the motion information memory, as a target. A first prediction signal similar to the block is generated, and first motion information and first additional information necessary for generating the first prediction signal are generated (step S100). Details of this step have already been described with reference to FIG. Next, the 0th motion information estimator 122 obtains from the reference screen of List 0 obtained from the frame memory 104 and the 0th motion prediction information candidate obtained from the motion information memory and the first motion information estimator 121. Is used to generate a 0th prediction signal similar to the target block, and to generate 0th motion information and 0th additional information necessary for generating the 0th prediction signal. (Step S150). Details of this step have already been described with reference to FIG.

  Next, the 0th additional information and the first additional information are entropy-encoded by the encoder 111, and the 0th motion information and the first motion information are stored in the motion information memory 113 (step S101). . Subsequently, in step S102, the prediction signal synthesizer 123 constituting the prediction signal generator 103 averages the 0th prediction signal and the first prediction signal to generate a bi-prediction signal of the target block. The residual signal indicating the difference between the pixel signal of the encoding target block and the prediction signal is transform-coded by the transformer 106, the quantizer 107, and the encoder 111 (step S103). The additional information and the encoded data of the quantized transform coefficient are output via the output terminal 112 (step S104). In order to predictively encode the subsequent encoding target block, the residual signal encoded after these processes or in parallel with these processes is decoded by the inverse quantizer 108 and the inverse transformer 109. Then, the adder 110 adds the decoded residual signal and the prediction signal, and reproduces the signal of the encoding target block. The reproduction signal is stored as a reference screen in the frame memory 104 (step S105). If all the encoding target blocks have not been processed, the process returns to step S100, and processing for the next encoding target block is performed. If all the encoding target blocks have been processed, the processing ends (step S106).

  Next, the video predictive decoding method according to this embodiment will be described. FIG. 10 is a block diagram showing a video predictive decoding device 200 according to this embodiment. The moving picture predictive decoding apparatus 200 includes an input terminal 201, a decoder 202, an inverse quantizer 203, an inverse transformer 204, an adder 205, an output terminal 206, a motion compensator 207, a motion information restoration unit 208, and a frame memory 104. A motion information memory 113 is provided. The inverse quantizer 203 and the inverse transformer 204 function as residual signal restoring means, and the motion information memory 113 functions as motion information recording means. The decoding means by the inverse quantizer 203 and the inverse transformer 204 may be performed using other than these. Further, the inverse converter 204 may not be provided.

  The input terminal 201 inputs compressed data that has been compression-encoded by the above-described video predictive encoding method. This compressed data includes encoded data of quantized transform coefficient information obtained by transform-quantizing an error signal and entropy-coding the encoded block divided into a plurality of blocks, and a second prediction signal for generating a block bi-predictive signal. The encoded data of 0 additional information and first additional information is included.

  The decoder 202 analyzes the compressed data input to the input terminal 201, separates the decoding target block into encoded data of quantized transform coefficients and encoded data of additional information, performs entry-pei decoding, L202a, The data is output to the inverse quantizer 203 and the motion information decompressor 208 via L202b, respectively.

  The motion information restoring unit 208 sends the 0th additional information (ref_idx [0], mvd [0], mvp_idx [0]) and the first additional information (ref_idx [1], mvp_idx [1]) via L202b. And using the decoded motion information acquired via L113, the 0th motion information (ref_idx [0], mv [0]) and the first motion information (ref_idx [1], mv Restore [1]). The restored zeroth motion information and first motion information are output to the motion compensator 207 and the motion information memory 113 via L208a and L208b, respectively. The motion information memory stores motion information.

  The motion compensator 207 acquires the already reproduced signal from the frame memory 104 based on the two pieces of motion information, averages the two prediction signals, and generates a bi-prediction signal for the decoding target block. The generated prediction signal is output to the adder 205 via L126.

  The quantized transform coefficient of the residual signal in the decoding target block decoded by the decoder 202 is output to the inverse quantizer 203 via L203. The inverse quantizer 203 inversely quantizes the quantization coefficient of the residual signal in the decoding target block. The inverse transformer 204 performs an inverse discrete cosine transform on the inversely quantized data to generate a residual signal.

  The adder 205 adds the bi-prediction signal generated by the motion compensator 207 to the residual signal restored by the inverse quantizer 203 and the inverse transformer 204, and the reproduced pixel signal of the decoding target block is added to the line L205. Via the output terminal 206 and the frame memory 104. The output terminal 206 outputs to the outside (for example, a display).

  The frame memory 104 stores the playback image output from the adder 205 as a reference screen as a reference playback image for the next decoding process.

  FIG. 11 is a block diagram showing the configuration of the motion information restoring unit 208 according to this embodiment. The motion information restoring unit 208 includes a first motion information restoring unit 211 and a 0th motion information restoring unit 212.

  The first motion information restoration unit 211 and the 0th motion information restoration unit 212 can operate simultaneously.

  The 0th motion information reconstructor 212 receives the 0th additional information (ref_idx [0], mvp_idx [0], mvd [0]) as input, and performs 0th prediction from the motion information of adjacent blocks obtained via L113. A motion vector (pmv [0] [ref_idx [0]] [mvp_idx [0]]) is generated, and the predicted motion vector and the difference motion vector (mvd [0]) included in the additional information are added to obtain the 0th The zeroth motion information is restored by generating a motion vector included in the motion information. Similarly, if the reference screen identified by ref_idx and the reference screen identified by ref_idx of the target block are different, they are included in the predicted motion information based on the frame numbers of the encoding target screen and the two reference screens. A motion vector scaling process may be performed.

  The first motion information decompressor 211 receives the first additional information (ref_idx [1], mvp_idx [1]) as an input, and uses the decoded motion information obtained via L113 as the first predicted motion vector (pmv [ 1] [ref_idx [1]] [mvp_idx [1]]). By using the predicted motion vector as a motion vector (mv [1] = pmv [1] [ref_idx [1]] [mvp_idx [1]) included in the first motion information, the first motion information is restored. . At this time, the differential motion vector mvd [1] may be set to a zero vector and added to the predicted motion vector to restore the first motion vector. At this time, if the reference screen identified by ref_idx and the reference screen identified by ref_idx of the target block are different, the prediction motion information is based on the frame numbers of the encoding target screen and the two reference screens. May be subjected to scaling processing of motion vectors included in. That is, the motion vector included in the predicted motion information is scaled and converted into a motion vector that indicates the reference screen identified by the target block, and the converted motion vector is used as the predicted motion vector.

  FIG. 13 shows a flowchart of the 0th motion information restoring unit 212. First, in step S451, decoded data of 0th additional information (ref_idx [0], mvp_idx [0], and mvd [0]) is input. Next, in step S452, based on the input ref_idx [0] and mvp_idx [0], from the adjacent block motion information as shown in FIG. The motion information indicated by mvp_idx [0] is input, and a predicted motion vector (pmv [0] [ref_idx [0]] [mvp_idx [0]]) is derived. At this time, if the reference screen identified by ref_idx and the reference screen identified by ref_idx of the target block are different, they are included in the predicted motion information based on the frame numbers of the encoding target screen and the two reference screens. A motion vector scaling process may be performed. Thereafter, in step S453, the generated motion vector predictor and the difference motion vector are added to restore the 0th motion vector (mv [0] = pmv [0] [ref_idx [0]] [mvp_idx [0] ] + Mvd [0]). Finally, in step S454, the 0th motion information (ref_idx [0] and mv [0]) is output to the motion compensator 207 and the motion information memory 113, and the process is terminated.

  FIG. 12 shows a flowchart of the first motion information restoring unit 211. First, decoded data of the first additional information (ref_idx [1] and mvp_idx [1]) is input in step S401, and the vector value of the difference motion vector (mvd [1]) is set to 0 in step S402. . Next, in step S403, based on the input ref_idx [1] and mvp_idx [1], as shown in FIG. 7, from the decoded motion information (n = 4 can be included), mvp_idx [1 ] Is input, and a predicted motion vector (pmv [1] [ref_idx [1]] [mvp_idx [1]]) is derived. At this time, if the reference screen identified by ref_idx and the reference screen identified by ref_idx of the target block are different, they are included in the predicted motion information based on the frame numbers of the encoding target screen and the two reference screens. A motion vector scaling process may be performed. Thereafter, in step S404, the generated motion vector predictor and the difference motion vector are added to restore the first motion vector (mv [1] = pmv [1] [ref_idx [1]] [mvp_idx [1] ] + Mvd [1]). Finally, in step S405, the first motion information (ref_idx [1] and mv [1]) is output to the motion compensator 207 and the motion information memory 113, and the process ends. Since the vector value of mvd [1] is always 0, step S402 may be omitted and the predicted motion vector may be set as a motion vector in step S404 (mv [1] = pmv [1] [ref_idx [1]] [mvp_idx [1]]).

  Next, a moving picture predictive decoding method in the moving picture predictive decoding apparatus 200 shown in FIG. 10 will be described using FIG. First, compressed data is input via the input terminal 201 (step S201). Then, the decoder 202 analyzes the data of the compressed data, performs entropy decoding, and decodes the 0th additional information and the first additional information and the quantized transform coefficient necessary for generating the bi-predictive signal (step) S202).

  Next, the 0th motion information restoring unit 212 constituting the motion information restoring unit 208 uses the 0th additional information and the motion information of the adjacent block stored in the motion information memory 113 to generate the 0th The motion information is restored (step S250). Details of this step have already been described with reference to FIG.

  Subsequently, in the first motion information restoring unit 211 constituting the motion information restoring unit 208, the first additional information and the motion information of the adjacent block stored in the motion information memory 113 are used to The motion information is restored (step S200). Details of this step have already been described with reference to FIG.

  Next, based on the restored motion information, the motion compensator 207 generates a bi-prediction signal for the decoding target block, and stores the motion information in the motion information memory 113 (S207).

  The quantized transform coefficient decoded by the decoder 202 is inversely quantized by the inverse quantizer 203 and inversely transformed by the inverse transformer 204 to generate a reproduction residual signal (S208). Then, the generated bi-prediction signal and the reproduction residual signal are added to generate a reproduction signal, and this reproduction signal is stored in the frame memory 104 to reproduce the next decoding target block (step S209). ). If there is the next compressed data, the process from S202 to S209 is repeated (S210), and all the data is processed to the end.

  Next, a method for selectively using a plurality of bi-prediction methods including the present invention will be described. Bi-prediction (referred to as bi-prediction type 2) that encodes only one differential motion vector described above is a bi-prediction (referred to as bi-prediction type 1) that encodes two conventional differential motion vectors; It can be used in combination with uni-prediction adaptively. These prediction methods can be used by switching on a screen unit, a slice unit including a plurality of blocks, or a block unit. For the switching process, a method of encoding the switching information or a method of determining on the decoding side based on the frame number of the reference screen can be used.

  In the process of switching between bi-prediction type 1 and bi-prediction type 2, the first motion information estimator 121 in FIG. 2 subtracts the prediction vector from the searched motion vector and the block matching function to calculate a differential motion vector. This can be realized by adding a function and adding a function of acquiring a differential motion vector from the decoder to the first motion information decompressor of FIG. A specific method will be described below.

(Screen / slice unit switching, bi-prediction type switching information encoding)
When switching between bi-prediction type 1 and bi-prediction type 2 on a screen / slice basis, bi-prediction type switching information (for example, BiPred_type) is included in the screen or slice header and encoded.

  As shown in FIG. 5A, when all the reference screen candidates are past screens in the display order from the encoding target screen, the bi-prediction type 2 is valid. Select the use of 2. Then, the encoding apparatus encodes instruction information indicating bi-prediction type 2 (for example, BiPred_type = 1) using screen or slice header information. On the other hand, as shown in FIG. 5B, when the future screen is included in the reference screen candidates with respect to the encoding target screen in the display order, the encoding device selects bi-prediction type 1 and displays the screen. Alternatively, instruction information (for example, BiPred_type = 0) indicating bi-prediction type 1 is encoded with the header information of the slice.

  When bi-prediction type 1 is used, a motion vector mv [1] is searched when a block in a screen or a slice is encoded. As with the 0th additional information (encoding ref_idx [0], mvd [0], and mvp_idx [0]), mvd [1] is included in the first additional information, along with ref_idx [1] and mvp_idx [1] Encode. When bi-prediction type 2 is used, ref_idx [0], mvd [0], and mvp_idx [0] are encoded as the 0th additional information, and ref_idx [1] and mvp_idx [1] as the 1st additional information Is encoded.

  On the decoding side, the mvd [1] restoration method is switched when each block in the screen or slice is decoded based on the instruction information decoded by the header information of the screen or slice. That is, when instruction information indicating bi-prediction type 1 (for example, BiPred_type = 0) is decoded, mvd [1] is included in the first additional information when each block in the screen or slice is decoded. And ref_idx [1] and mvp_idx [1]. When instruction information indicating bi-prediction type 2 (for example, BiPred_type = 1) is decoded, when decoding each block in the screen or slice, ref_idx [1] and mvp_idx [1 are used as the first additional information. ] And set the horizontal and vertical vector values of mvd [1] to 0. As the 0th additional information, ref_idx [0], mvd [0], and mvp_idx [0] are decoded regardless of the value of the instruction information.

  Note that the switching method between bi-prediction type 1 and bi-prediction type 2 on the encoding side does not depend on the method shown here. For example, bi-prediction type 2 may be used when all reference screens are future screens relative to the encoding target screen in the display order. Further, the coding efficiency (for example, the evaluation value obtained by converting the square sum of the coding error signal + the number of coding bits) when bi-prediction type 1 and bi-prediction type 2 are actually applied is investigated and coded. A method of selecting the one with higher efficiency can be adopted.

  Further, in the encoding method that adaptively uses a plurality of block sizes as shown in FIG. 8, instruction information for switching between bi-prediction type 1 and bi-prediction type 2 is displayed for different block sizes on the screen or slice header. You may make it send separately. For example, when the block size is 64 × 64, 32 × 32, 16 × 16, or 8 × 8, four pieces of instruction information are encoded.

  When both uni-prediction and bi-prediction are used in the block prediction method, a prediction type (uni-prediction or bi-prediction) is selected on the encoding side, and is encoded by including in additional information. Then, the switching process between the bi-prediction type 1 and the bi-prediction type 2 is performed only in the block for which bi-prediction is selected. On the decoding side, the switching process between the bi-prediction type 1 and the bi-prediction type 2 is performed only with a block obtained by decoding information indicating bi-prediction as a prediction type.

(Screen / slice unit switching, determination based on frame number of reference screen)
The reference screen candidates shown in FIGS. 5A and 5B are the same on the encoding side and the decoding side. Therefore, it is possible to determine which of bi-prediction type 1 and bi-prediction type 2 is used based on the frame number of the reference screen candidate and the frame number of the encoding target screen. That is, when all the reference screen candidates are past screens in the display order from the encoding target screen, the bi-prediction type 2 is used, and the future screen is referred to the encoding target screen in the display order. Bi-prediction type 1 is used. In this method, there is no need to send instruction information.

  In addition, when all the reference screen candidates are future screens in the display order with respect to the encoding target screen, the bi-prediction type 2 may be used.

(Block unit switching, switching information encoding)
When switching between bi-prediction type 1 and bi-prediction type 2 in block units, bi-prediction type switching information (for example, BiPred_block_type) is included in the block additional information and encoded.

  Since the bi-prediction type 2 is valid when the two reference screens of bi-prediction are past screens in the display order from the encoding target screen, the use of bi-prediction type 2 is selected on the encoding side. Then, the encoding apparatus includes the instruction information indicating bi-prediction type 2 (for example, BiPred_block_type = 1) in the additional information of the block, and includes ref_idx [0], mvd [0], mvp_idx [0, which are the 0th additional information. ] And the first additional information ref_idx [1] and mvp_idx [1]. On the other hand, when the two reference screens of bi-prediction include future screens relative to the encoding target screen in the display order, bi-prediction type 1 is selected. Then, the encoding apparatus includes instruction information indicating bi-prediction type 1 (for example, BiPred_block_type = 0) in the additional information of the block, and includes ref_idx [0], mvd [0], and mvp_idx [0 as the 0th additional information. ], The first additional information ref_idx [1], mvd [1] (mvd [1] is included in the first additional information), and mvp_idx [1].

  On the decoding side, decoding is performed by including bi-prediction type switching information (for example, BiPred_block_type) in the block additional information, and the restoration method of mvd [1] is switched based on the decoded value. That is, when instruction information indicating bi-prediction type 1 (for example, BiPred_block_type = 0) is decoded as additional information of a block, ref_idx [0], mvd [0], and mvp_idx [0 are used as the 0th additional information. ] And ref_idx [1], mvd [1], and mvp_idx [1] as first additional information. On the other hand, when instruction information indicating bi-prediction type 2 (for example, BiPred_block_type = 1) is decoded as additional information of a block, ref_idx [0], mvd [0], and mvp_idx [0] are added as 0th additional information. And ref_idx [1] and mvp_idx [1] are decoded as the first additional information, and the horizontal and vertical vector values of mvd [1] are set to 0.

  Note that the switching method between bi-prediction type 1 and bi-prediction type 2 on the encoding side does not depend on the method shown here. For example, when both of the two reference screens are future screens relative to the encoding target screen in the display order, bi-prediction type 2 may be used. Also, the coding efficiency (for example, the evaluation value obtained by converting the square sum of the coding error signal + the number of coding bits) when bi-prediction type 1 and bi-prediction type 2 are actually applied is investigated for each block. A method of selecting the one with higher encoding efficiency can be adopted.

  Also, when both uni-prediction and bi-prediction are used in the block prediction method, a prediction type (uni-prediction or bi-prediction) is selected on the encoding side, and encoding is included in the additional information. Then, the switching process between the bi-prediction type 1 and the bi-prediction type 2 is performed only in the block for which bi-prediction is selected. On the decoding side, the switching process between the bi-prediction type 1 and the bi-prediction type 2 is performed only with a block obtained by decoding information indicating bi-prediction as a prediction type.

(Block unit switching, determination based on frame number of reference screen)
The reference screen candidates shown in FIGS. 5A and 5B are the same on the encoding side and the decoding side. Therefore, whether to use bi-prediction type 1 or bi-prediction type 2 based on the frame number of the two reference screens used for bi-prediction to be encoded / decoded as additional information of the block and the frame number of the encoding target screen. Can be determined. That is, if the two reference screens used for bi-prediction are both past in the display order from the encoding target screen, the bi-prediction type 2 is used, and one or both of the two reference screens are encoded in the display order. In the case of a future screen with respect to the target screen, bi-prediction type 1 is used. This method does not require sending instruction information. Note that bi-prediction type 2 may be used when both of the two reference screens are future screens relative to the encoding target screen in the display order.

(Combination of screen / slice unit switching and block unit switching)
The instruction information indicating whether the bi-prediction type is switched in units of screens / slices or in units of blocks is encoded / decoded in units of screens / slices.

  When bi-prediction type switching is performed on a screen / slice basis, in addition, as described above, bi-prediction type switching information (for example, BiPred_type) is included in the screen or slice header to encode / Decrypt. At this time, in the encoding method that adaptively uses a plurality of block sizes as shown in FIG. 8, instruction information for switching between bi-prediction type 1 and bi-prediction type 2 for different block sizes on the screen or slice header. May be sent individually. For example, when the block size is 64 × 64, 32 × 32, 16 × 16, or 8 × 8, four pieces of instruction information are encoded.

  When bi-prediction type switching is performed in units of block screens, in addition, as described above, bi-prediction type switching information (for example, BiPred_block_type) is included in the additional information of the block for each block. Convert / decrypt. At this time, as shown in FIG. 8, in the encoding method using a plurality of block sizes adaptively, switching between bi-prediction type 1 and bi-prediction type 2 is performed for different block sizes on the screen or slice header. Instruction information for instructing whether or not to perform processing may be individually sent. For example, when the block size is 64 × 64, 32 × 32, 16 × 16, or 8 × 8, four pieces of instruction information are encoded.

  In addition, instruction information for instructing whether to use only bi-prediction type 1 or only bi-prediction type 2 or to switch between two bi-prediction types for each block in the screen / slice in screen / slice units. May be encoded / decoded. At this time, in the encoding method using a plurality of block sizes adaptively as shown in FIG. 8, the instruction information may be individually sent for different block sizes.

  Alternatively, the instruction information indicating whether or not to use bi-prediction type 2 is encoded / decoded for each screen / slice. For screens / slices that use bi-prediction type 2, switch between bi-prediction type 1 and bi-prediction type 2 for each block in the screen / slice, or use bi-prediction type 2 for all blocks in the screen / slice. It is also possible to encode / decode instruction information instructing whether or not. At this time, in the encoding method that adaptively uses a plurality of block sizes as shown in FIG. 8, the instruction information may be individually sent for different block sizes on the screen or the slice header.

  In the above description, the 0th motion information of the target block 400 is included in the first predicted motion vector candidates shown in FIG. However, the 0th motion information is used as the first predicted motion information, and the predicted motion information is used as the first reference screen index included in the 0th motion information and the first reference screen index included in the first motion information. A method of scaling based on this and using it as the first motion information may be separately prepared as bi-prediction type 3. That is, the scaling is performed so that the motion vector included in the 0th motion information becomes the motion vector of the reference screen indicated by the reference screen index of the first motion information. In this case, the first motion prediction information candidate in the bi-prediction type 2 may not include the 0th motion information of the target block (n = 4 in FIG. 7 is excluded from the candidates). When bi-prediction type 3 is used, the first prediction motion information index is not encoded / decoded. That is, when the first additional information is encoded / decoded, the first reference screen index (ref_idx [1]) is included, the difference motion vector (mvd [1]) and the first predicted motion information index ( mvp_idx [1]) should not be included.

  As a method of using bi-prediction type 3, instruction information for switching between bi-prediction type 1, bi-prediction type 2, and bi-prediction type 3 may be encoded / decoded on a screen / slice basis or a block basis. In addition, instruction information for switching between bi-prediction type 1 and bi-prediction type 3 or bi-prediction type 2 and bi-prediction type 3 may be encoded / decoded in units of screens / slices or blocks.

  In addition, a method that can be used as a substitute for bi-prediction type 1 is also conceivable. That is, when a plurality of reference screen candidates include future screens with respect to the encoding target screen in the display order, the bi-prediction type 3 is used, and all the plurality of reference screen candidates are displayed in the display order. On the other hand, in the case of a past screen, bi-prediction type 2 may be used.

  Alternatively, when the reference screen indicated by the reference screen index included in the 0th motion information is different from the reference screen indicated by the reference screen index of the first motion information, the bi-prediction type 2 instead of the bi-prediction type 2 3 (the encoding / decoding of the first prediction motion information index is unnecessary), while the reference screen indicated by the reference screen index included in the 0th motion information is the reference screen index of the first motion information In the case of the same reference screen as instructed, bi-prediction type 2 may be used (the first prediction motion information index is encoded / decoded).

  Note that the bi-prediction type switching information is not limited to identification information such as BiPred_type and BiPred_block_type. Any information indicating whether or not to encode / decode a differential motion vector may be used. For example, flag information indicating whether mvd is included in the first additional information or the zeroth additional information, such as mvd_list1_zero_flag and mvd_list0_zero_flag, may be encoded / decoded. When switching between bi-prediction type 1, bi-prediction type 2 and bi-prediction type 3 in screen / slice units or block units, when using either bi-prediction type 2 or bi-prediction type 3, mvd Send a flag indicating that it does not contain Switching between bi-prediction type 2 and bi-prediction type 3 may be performed based on the reference screen index as described above (the reference screen indicated by the reference screen index included in the 0th motion information is the first screen). When the reference screen index of the motion information is the same as the reference screen instructed, bi-prediction type 2 is used), and the selection information may be encoded / decoded.

  FIG. 15 is a block diagram illustrating modules of a program that can execute the video predictive coding method. As shown in FIG. 15A, the moving picture predictive encoding program P100 includes a block division module P101, a prediction signal generation module P103, a storage module P104, a subtraction module P105, a conversion module P106, a quantization module P107, and an inverse quantization module. P108, an inverse transform module P109, an addition module P110, an encoding module P111, and a motion information storage module P113 are provided. Further, as shown in FIG. 15B, the prediction signal generation module P103 includes a first motion information estimation module P121, a zeroth motion information estimation module P122, and a prediction signal synthesis module P123. The functions realized by executing the respective modules are the same as the functions of the moving picture predictive coding apparatus 100 described above. That is, the function of each module of the moving picture predictive coding program P100 includes a block divider 102, a prediction signal generator 103, a frame memory 104, a subtractor 105, a converter 106, a quantizer 107, an inverse quantizer 108, The functions of the inverse transformer 109, adder 110, encoder 111, motion information memory 113, first motion information estimator 121, zeroth motion information estimator 122, and prediction signal synthesizer 123 are the same.

  FIG. 16 is a block diagram showing modules of a program that can execute the moving picture predictive decoding method. As shown in FIG. 16A, the video predictive decoding program P200 includes a decoding module P201, a motion information decoding module P202, a motion compensation module P203, a motion information storage module P204, an inverse quantization module P205, an inverse transform module P206, and addition. A module P207 and a storage module P104 are provided. Further, as shown in FIG. 16B, the motion information decoding module P202 includes a first motion information restoration module P211 and a 0th motion information restoration module P212.

  The functions realized by executing the modules are the same as those of the components of the moving picture predictive decoding apparatus 200 described above. That is, the functions of each module of the moving picture predictive decoding program P200 are as follows: the decoder 202, the motion information restorer 208, the motion compensator 207, the motion information memory 113, the inverse quantizer 203, the inverse transformer 204, and the adder 205. The functions of the frame memory 104, the first motion information restoring unit 211, and the zeroth motion information restoring unit 212 are the same.

  The moving picture predictive coding program P100 or the moving picture predictive decoding program P200 configured as described above is stored in the recording medium 10 shown in FIGS. 17 and 18 described later, and is executed by a computer described later.

  FIG. 17 is a diagram illustrating a hardware configuration of a computer for executing a program recorded in a recording medium, and FIG. 18 is a perspective view of the computer for executing a program stored in the recording medium. Note that a program that executes a program stored in a recording medium is not limited to a computer, and may be a DVD player, a set-top box, a mobile phone, or the like that includes a CPU and performs processing and control by software.

  As shown in FIG. 17, the computer 30 includes a reading device 12 such as a flexible disk drive device, a CD-ROM drive device, and a DVD drive device, a working memory (RAM) 14 in which an operating system is resident, and a recording medium 10. A memory 16 for storing programs stored therein, a display device 18 such as a display, a mouse 20 and a keyboard 22 as input devices, a communication device 24 for transmitting and receiving data and the like, and a CPU 26 for controlling execution of the programs. And. When the recording medium 10 is inserted into the reading device 12, the computer 30 can access the moving image predictive encoding / decoding program stored in the recording medium 10 from the reading device 12. Thus, it becomes possible to operate as a video encoding device or a video decoding device according to the present embodiment.

  As shown in FIG. 18, the moving picture predictive encoding program and the moving picture decoding program may be provided as a computer data signal 40 superimposed on a carrier wave via a network. In this case, the computer 30 can store the video predictive encoding program or video decoding program received by the communication device 24 in the memory 16 and execute the video predictive encoding program or video predictive decoding program. .

  In the present invention, the following modifications are possible.

(1) Relationship between 0th Additional Information and 1st Additional Information In the above, the 1st additional information includes ref_idx [1] and mvp_idx [1], and the 0th additional information includes ref_idx [0] and mvd [ 0] and mvp_idx [0] are included, but this configuration may be reversed. That is, the first additional information includes ref_idx [1], mvd [1], and mvp_idx [1], and the zeroth additional information includes ref_idx [0] and mvp_idx [0]. In this case, the 0th prediction motion information is generated from the motion information of the adjacent block and the 0th prediction signal is generated. Subsequently, the absolute value sum of the bi-prediction signal and the original signal of the target block + the additional information The first motion information that minimizes the evaluation value is searched.

  Also, the first additional information may include ref_idx [1] and mvp_idx [1], and the zeroth additional information may include ref_idx [0] and mvp_idx [0]. That is, the difference motion vector is not encoded, and the horizontal and vertical vector values are set to 0 on the decoding side of the two difference motion vectors. Alternatively, the zeroth motion vector and the first motion vector are set to the zeroth motion vector and the first motion vector, respectively.

(2) Reference screen In the above description, the reference screen index is included in the first additional information for encoding / decoding. However, the reference screen index is determined based on the reference screen index attached to the motion information of the adjacent block indicated by mvp_idx. (For example, for the adjacent blocks 401 to 404 in FIG. 7, the reference screen index of the attached List 1 is used as it is, and for the adjacent blocks 410 to 415, the reference screen index of the target block and the adjacent block indicates. Based on the difference in the reference screen, a scaling process is performed on the motion vector included in the adjacent block). Also, a reference screen index for the first motion information may be determined in advance. In these cases, since the decoding side can uniquely restore, it is not necessary to encode the reference screen index ref_idx by including it in the first additional information.

  Selection of the reference screen and its list are not limited to the examples of FIGS. The reference screen list may be encoded on a screen / slice basis, and the number of reference screens is not four as shown in FIG. 5, and the present invention can be implemented with any other number.

(3) Predicted motion information In the above, a plurality of encoded / decoded motion information is used as predicted motion information candidates as shown in FIG. 7, but the number and the position of adjacent blocks to be used are not limited.

  The 0th motion vector of the adjacent block may be included in the first motion vector predictor candidate, or the first motion vector of the adjacent block may be included in the 0th motion vector predictor candidate. Can be implemented. The 0th motion information previously encoded / decoded may be added to the first predicted motion information candidate.

  Whether or not the motion vector scaling processing is performed when the reference screen index included in the encoded / decoded motion information is different from the reference screen index included in the motion information associated with the target block is not limited in the present invention.

  When there is one candidate motion vector predictor, it is not necessary to encode the motion information index mvp_idx included in the first additional information.

  The method for generating a predicted motion vector is not limited in the present invention. For example, the median value of the odd number of motion vector predictor candidates may be adopted as the motion vector predictor. Also in this case, since the prediction motion vector is uniquely determined on the decoding side, it is not necessary to encode the prediction motion information index mvp_idx by including it in the first additional information.

  As shown in FIG. 8, even in the case of an encoding / decoding scheme using a plurality of block sizes, bi-prediction according to the present invention can be implemented if a method for determining prediction motion information candidates is defined. For example, in the example of FIG. 8, the blocks 421 to 428 adjacent to the target block 400 may be candidates for prediction information, and the numbering method of adjacent blocks is determined in advance and designated in units of screens or slices. You may make it set the motion information of several adjacent blocks to the prediction motion vector candidate.

(4) 0th motion information and additional information The configurations of the 0th motion information and additional information may be fixed values without encoding the reference screen index and the predicted motion information index, which are not limited in the present invention. However, it may be derived by a predetermined method.

  Also, for the 0th motion information and additional information, as with the first motion information and additional information, mvd [0] is not encoded, and mv [0] = pmv [0] ( Alternatively, mvd [0] = 0) may be set.

(5) Frame number In the above description, the frame number (frame_num) is used to identify the reference screen and the encoding target screen. However, other information can be used as long as it is information that can identify the reference screen. There is no impact on implementation.

(6) Restoration of first differential motion vector mvd [1] In the bi-prediction of the present invention described above, the vector value of the first differential motion vector mvd [1] is 0. For this reason, mvd [1] is not encoded, and the vector value of mvd [1] is set to 0 or the motion vector mv [1] is set to pmv [1] on the decoding side. As another embodiment, it can be said that a method in which mvd [1] = 0 is included in the first additional information and 0 value is efficiently entropy-coded is also effective.

  When arithmetic coding is used for entropy coding, for example, the 0th differential motion vector and the first differential motion vector are encoded / decoded with different probability models. For example, two probability models with different probability frequencies of the difference motion vector vector value 0 are prepared for encoding / decoding the difference motion vector. When encoding / decoding the first difference motion vector, the second probability model in which the frequency of the vector value 0 of the difference motion vector is set higher is used to encode the zeroth difference motion vector. In the case of decoding, another first probability model is used. Further, a probability model may be separately prepared for the horizontal and vertical vector values of the difference motion vector.

  The second probability model is used only for the first difference motion vector of the block to which bi-prediction type 2 is applied, and the first probability model is used for encoding / decoding of the other difference motion vectors. It may be.

  When variable length coding is used, the 0th differential motion vector and the first differential motion vector are encoded / decoded using different variable length tables. For example, two variable length tables having different code lengths assigned to the vector value 0 of the differential motion vector are prepared. Then, when encoding / decoding the first differential motion vector, the second variable length table in which the code length assigned to the vector value 0 of the differential motion vector is set shorter is used, and the zeroth differential motion vector is used. Is encoded / decoded, another first variable length table is used. Furthermore, a variable length table may be separately prepared for the horizontal and vertical vector values of the difference motion vector.

  Note that the second variable length table is used only for the first differential motion vector of a block to which bi-prediction type 2 is applied, and the first variable length table is used for encoding / decoding of the other differential motion vectors. You may do it.

(7) N prediction In the above, the prediction types of inter-screen prediction are uni-prediction and bi-prediction, but the present invention can also be applied to a prediction method that combines three or more prediction signals. When a prediction signal is generated by synthesizing three or more prediction signals, the number of additional information not including mvd may be any number as long as it is one or more.

(8) Transformer, Inverse Transformer The residual signal conversion process may be performed with a fixed block size, or the target area may be subdivided according to the partial area and the conversion process may be performed.

(9) Color signal In the above description, the color format is not particularly described, but the prediction signal may be generated separately from the luminance signal for the color signal or the color difference signal. Further, it may be performed in conjunction with the luminance signal processing.

  The present invention has been described in detail based on the embodiments. However, the present invention is not limited to the above embodiment. The present invention can be variously modified without departing from the gist thereof.

  DESCRIPTION OF SYMBOLS 100 ... Moving image predictive coding apparatus, 101 ... Input terminal, 102 ... Block divider, 103 ... Prediction signal generator, 104 ... Frame memory, 105 ... Subtractor, 106 ... Converter, 107 ... Quantizer, 108 ... Inverse quantizer 109 ... Inverse transformer 110 ... Adder 111 ... Encoder 112 ... Output terminal 113 ... Memory for motion information 121 ... First motion information estimator 122 ... 0th motion Information estimator, 123 ... Predictive signal synthesizer, 201 ... Input terminal, 202 ... Decoder, 203 ... Inverse quantizer, 204 ... Inverse transformer, 205 ... Adder, 206 ... Output terminal, 207 ... Motion compensator, 208... Motion information restorer, 211... First motion information restorer, 212.

Claims (6)

Decoding means for decoding the additional information of the decoding target region to be decoded and the compressed data of the residual signal from among the compressed data encoded by dividing into a plurality of regions;
Motion information restoring means for restoring a motion vector required for generating a prediction signal of the target region from the additional information;
Motion information recording means for storing the motion vector;
Motion compensation means for generating a prediction signal of the target region based on the motion vector;
Residual signal restoration means for restoring the reproduction residual signal of the target area from the compressed data of the residual signal;
Image recording means for restoring the pixel signal of the decoding target area by adding the prediction signal and the reproduction residual signal, and storing the restored pixel signal as a previously reproduced image;
Comprising
The decoding means decodes the 0th additional information and the first additional information,
The 0th additional information is a 0th differential motion vector and a 0th motion vector for identifying one selected from a plurality of motion vectors stored in the motion information recording means as a 0th predicted motion vector. And a predicted motion information index of
The first additional information includes a first predicted motion information index that identifies one selected from a plurality of motion vectors stored in the motion information recording unit as a first predicted motion vector,
The movement information restoring means is
A 0th motion vector predictor is generated based on the 0th motion vector predictor index, and the 0th motion vector is restored by adding the generated 0th motion vector predictor and the 0th differential motion vector. Zeroth motion information restoring means;
First motion information restoring means for generating a first predicted motion vector based on the first predicted motion information index and restoring the generated first predicted motion vector as a first motion vector;
The motion compensation means generates a prediction signal of the target region by combining two signals acquired from the already reproduced image based on the 0th motion vector and the first motion vector,
The decoding means further decodes instruction information indicating whether or not the first additional information includes a difference motion vector for each screen or slice,
When the instruction information indicates that the first additional information includes a first differential motion vector, the differential motion vector is decoded as the first additional information;
The first motion information restoration means includes
When the instruction information indicates that the first additional information does not include the first differential motion vector, a first predicted motion vector is generated based on the first predicted motion information index, and the generated first Restore one predicted motion vector as the first motion vector;
When the instruction information indicates that the first additional information includes a difference vector, a first predicted motion vector is generated based on the first predicted motion information index, and the generated first predicted motion vector And a decoded differential motion vector to generate and restore a first motion vector.   The decoding means decodes the instruction information included in a screen header or a slice header,
  The moving picture predictive decoding apparatus according to claim 1. A video predictive decoding method executed by a video predictive decoding device,
A decoding step of decoding the additional information of the decoding target area to be decoded and the compressed data of the residual signal from the compressed data encoded by dividing into a plurality of areas;
A motion information restoring step for restoring a motion vector required for generating a prediction signal of the target region from the additional information;
A motion information recording step for storing the motion vector in a motion information recording means;
A motion compensation step of generating a prediction signal of the target region based on the motion vector;
A residual signal restoration step of restoring the reproduction residual signal of the target area from the compressed data of the residual signal;
An image recording step of restoring the pixel signal in the decoding target area by adding the prediction signal and the reproduction residual signal, and storing the restored pixel signal as a previously reproduced image;
Comprising
In the decoding step, the video predictive decoding device decodes the 0th additional information and the first additional information,
The 0th additional information is a 0th differential motion vector and a 0th motion vector for identifying one selected from a plurality of motion vectors stored in the motion information recording means as a 0th predicted motion vector. And a predicted motion information index of
The first additional information includes a first predicted motion information index that identifies one selected from a plurality of motion vectors stored in the motion information recording unit as a first predicted motion vector,
The motion information restoration step comprises:
A 0th motion vector predictor is generated based on the 0th motion vector predictor index, and the 0th motion vector is restored by adding the generated 0th motion vector predictor and the 0th differential motion vector. A 0th motion information restoration step;
Generating a first predicted motion vector based on the first predicted motion information index and restoring the generated first predicted motion vector as a first motion vector; and
In the motion compensation step, the video predictive decoding apparatus combines the two signals acquired from the already reproduced image based on the 0th motion vector and the first motion vector to predict the target region. Generate a signal,
In the decoding step, the video predictive decoding apparatus further decodes instruction information indicating whether or not the first additional information includes a differential motion vector for each screen or slice,
When the instruction information indicates that the first additional information includes a first differential motion vector, the video predictive decoding device decodes the differential motion vector as the first additional information,
In the first motion information restoration step, the video predictive decoding device includes:
When the instruction information indicates that the first additional information does not include the first differential motion vector, a first predicted motion vector is generated based on the first predicted motion information index, and the generated first Restore one predicted motion vector as the first motion vector;
When the instruction information indicates that the first additional information includes a difference vector, a first predicted motion vector is generated based on the first predicted motion information index, and the generated first predicted motion vector And a decoded differential motion vector to generate and restore a first motion vector.   In the decoding step, the video predictive decoding device decodes the instruction information included in a screen header or a slice header.
  The moving image predictive decoding method according to claim 3. Computer
Decoding means for decoding the additional information of the decoding target region to be decoded and the compressed data of the residual signal from among the compressed data encoded by dividing into a plurality of regions;
Motion information restoring means for restoring a motion vector required for generating a prediction signal of the target region from the additional information;
Motion information recording means for storing the motion vector;
Motion compensation means for generating a prediction signal of the target region based on the motion vector;
Residual signal restoration means for restoring the reproduction residual signal of the target area from the compressed data of the residual signal;
Image recording means for restoring the pixel signal of the decoding target area by adding the prediction signal and the reproduction residual signal, and storing the restored pixel signal as a previously reproduced image;
Is a moving picture predictive decoding program to function as
The decoding means decodes the 0th additional information and the first additional information,
The 0th additional information is a 0th differential motion vector and a 0th motion vector for identifying one selected from a plurality of motion vectors stored in the motion information recording means as a 0th predicted motion vector. And a predicted motion information index of
The first additional information includes a first predicted motion information index that identifies one selected from a plurality of motion vectors stored in the motion information recording unit as a first predicted motion vector,
The movement information restoring means is
A 0th motion vector predictor is generated based on the 0th motion vector predictor index, and the 0th motion vector is restored by adding the generated 0th motion vector predictor and the 0th differential motion vector. Zeroth motion information restoring means;
First motion information restoring means for generating a first predicted motion vector based on the first predicted motion information index and restoring the generated first predicted motion vector as a first motion vector;
The motion compensation means generates a prediction signal of the target region by combining two signals acquired from the already reproduced image based on the 0th motion vector and the first motion vector,
The decoding means further decodes instruction information indicating whether or not the first additional information includes a difference motion vector for each screen or slice,
When the instruction information indicates that the first additional information includes a first differential motion vector, the differential motion vector is decoded as the first additional information;
The first motion information restoration means includes
When the instruction information indicates that the first additional information does not include the first differential motion vector, a first predicted motion vector is generated based on the first predicted motion information index, and the generated first Restore one predicted motion vector as the first motion vector;
When the instruction information indicates that the first additional information includes a difference vector, a first predicted motion vector is generated based on the first predicted motion information index, and the generated first predicted motion vector And a decoded difference motion vector to generate and restore a first motion vector.   The decoding means decodes the instruction information included in a screen header or a slice header,
  The moving picture predictive decoding program according to claim 5.

Images