Fast mode decision-making method for multi-view video coding

Abstract

The invention discloses a fast mode decision-making method for multi-view video coding, which comprises the following steps of: S1, dividing coding blocks by taking 16 * 16 pixels as units, performing coarse intra-frame prediction, coarse inter-frame prediction and coarse inter-view prediction, and storing the obtained three types of cost in a data structure of a corresponding frame; s2, extracting continuous N frames from any viewpoint of the current video sequence, extracting the motion vector of the coarse inter-frame prediction for each coding block in the N frames, clustering the motion vector, and calculating the Euclidean distance of the clustering center after the clustering is completed; s3, coding the reference viewpoint, executing a Skip mode on the current coding block, if the code block mark CBF after execution is equal to 0, setting the optimal mode of the coding block as the Skip mode, and ending the coding of the current coding block; otherwise, executing S4; and S4, based on the obtained cost of the coarse intra-frame prediction, the coarse inter-frame prediction and the coarse inter-viewpoint prediction, checking whether the coarse intra-frame cost of the current coding block is the optimal cost.

Description

Technical Field

[0001] The invention belongs to the technical field of video encoding, and in particular relates to a fast mode decision method for multi-viewpoint video encoding. Background Art

[0002] In recent years, with the rapid development of multimedia technology, people are no longer satisfied with simple high-definition video applications, but hope to receive and experience video content in a multi-faceted manner through their own senses. This demand has promoted the development of ordinary single-view video content applications to multi-view video applications. Multi-view video refers to video that is recorded by multiple cameras at different angles at the same time. In this type of video application, multi-view video compression is an important research and component. The international standards group has expanded on the basis of the single-view compression standard and released standards for multi-view video. The most advanced one is the MV-HEVC standard based on the HEVC standard extension. Compared with multi-channel simulcast encoding, the technology used in the MV-HEVC standard can reduce the bit rate by 20% to 30% under the same picture quality.

[0003] The prediction technology used by MV-HEVC in the process of predictive coding is basically the same as that of HEVC, including 35 angle predictions for intra-frame prediction; AMVP, Merge, and Skip modes for inter-frame prediction. Among them, Merge and Skip modes perform inter-frame prediction based on adjacent encoded MVs, and do not need to search for MVs by themselves. The AMVP mode searches for the optimal MV in a specified area when there is no suitable adjacent encoded MV, and finds a suitable MV by itself. Similarly, in the process of rate-distortion optimization, the CTU (Coding Tree Unit) needs to be recursively divided into 64×64, 32×32, 16×16, and 8×8. However, the increase in the number of viewpoints to be compressed also brings about a more complex encoding process. In addition to the intra-frame prediction and inter-frame prediction of the newly added viewpoints themselves, different viewpoints in the same POC (Picture Order Count) will use a new prediction coding method-inter-view prediction to pursue smaller residuals. The principle of inter-view prediction is basically the same as that of inter-frame prediction, and the main difference lies in the different reference frames. The reference frame of inter-frame prediction is the adjacent frame in the time domain, and the motion estimation performed on this frame is called ME (Motion Estimation). The reference frame of inter-view prediction is the adjacent viewpoint frame at the same time, and the motion estimation performed on this frame is called disparity estimation DE (Disparity Estimation).

[0004] In the reference viewpoints of non-key frames in multi-view videos, due to the existence of reference viewpoints and reference frames in the time domain, the proportion of intra-frame mode selected as the optimal prediction mode is less than 5%. Traditional MV-HEVC still needs to perform intra-frame prediction, which increases the encoding complexity of multi-view videos. In addition, the selection of inter-frame prediction and inter-view prediction with similar execution principles is often distributed regionally. If a certain area tends to choose inter-view prediction, then the execution of inter-frame prediction in this area is redundant, which will also affect the encoding complexity. Summary of the invention

[0005] In view of the above problems, the present invention provides a fast mode decision method for multi-view video coding, which is used to obtain the estimated mode of each block to be encoded based on the cost of coarse intra-frame prediction and inter-frame prediction in the pre-analysis stage to skip the execution of redundant intra-frame prediction modes.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0007] A fast mode decision method for multi-view video coding comprises the following steps:

[0008] S1, based on the pre-analysis lookahead process, select a GOP number of frames to be encoded into the lookahead sequence, reduce the resolution of these encoded frames to 1 / 2 of the original storage, divide the encoding blocks into 16*16 pixels, perform coarse intra-frame prediction, coarse inter-frame prediction and coarse inter-viewpoint prediction, and store the three obtained costs into the data structure of the corresponding frame;

[0009] S2, extract N consecutive frames from any viewpoint of the current video sequence, extract the coarse inter-frame predicted motion vector for each coding block in the N frames, cluster the motion vectors using the K-Means algorithm, and after clustering, calculate the Euclidean distance Distance(i,j) of the cluster center:

[0010]

[0011] Where i and j represent the i-th and j-th cluster centers respectively. Assuming the judgment threshold sh is 5.0, the current video sequence type Type is determined by the following formula:

[0012]

[0013] S3, encode the reference viewpoint, execute the Skip mode for the current coding block, if the coding block flag CBF after execution is 0, set the optimal mode of the coding block to the Skip mode, and end the encoding of the current coding block; otherwise, execute S4;

[0014] S4, based on the obtained costs of coarse intra-frame prediction, coarse inter-frame prediction and coarse inter-viewpoint prediction, check whether the coarse intra-frame cost of the current coding block is the optimal cost; if so, further check the current frame type, if the current frame is a P frame, or the current frame is a B frame, but the cost difference percentage between the coarse intra-frame prediction and the coarse inter-frame prediction and coarse inter-viewpoint prediction is greater than 30%, then the current coding block executes the intra-frame prediction mode, and the optimal prediction mode of the coding block is set to the intra-frame prediction mode, and the encoding of the current coding block is ended; otherwise, execute S5;

[0015] S5, calculating the cost difference percentage between the coarse inter-frame prediction and the coarse inter-viewpoint prediction of the current coding block, if the percentage is less than 20%, executing S6; otherwise, if the cost of the coarse inter-frame prediction is smaller, executing S8; if the cost of the coarse inter-viewpoint prediction is smaller, executing S9;

[0016] S6, obtaining the prediction mode distribution selected by the adjacent coded coding blocks of the current coding block, if the percentage of inter-frame prediction is high, executing S8; if the percentage of inter-view prediction is high, executing S9; otherwise, executing S7;

[0017] S7, if the current video type is a translation type, execute S8; if the current video type is an irregular type, execute S9;

[0018] S8, according to the adjacent coded blocks of the current coding block, the coding blocks at the corresponding position of the reference frame and all adjacent coding blocks, for the inter-frame prediction of the referenced view, these coding blocks only perform the optimal prediction mode, and store the corresponding optimal cost as the cost of the coding block; let the number of Merge modes selected by the 9 coding blocks in the referenced frame be C RMerge , then the probability of selecting the Merge mode is:

[0019]

[0020] The probability of selecting AMVP mode is:

[0021]

[0022] For the current coding block, there are five adjacent blocks that have been coded. In the special case where the coding block on the lower left is in the same CTU as the current coding block, it will be pre-coded. The probability P of selecting the Merge mode is also calculated. CM erge and the probability P of selecting AMVP mode CA mvp; then the prediction probabilities of the current coding block selecting the Merge mode and the AMVP mode are:

[0023] Pre Merge =α1PRMerge +α2P CMerge

[0024] Pre Amvp =α1P RAmvp +α2P CAmvp

[0025] Among them, α1 and α2 are weight coefficients, and their calculation formula is:

[0026]

[0027] α2=1-α1

[0028] Among them, POCdiff is the absolute value of the frame number difference between the reference frame and the reference frame; the value of α1 is inversely proportional to the absolute value, which means that the farther the reference frame is, the smaller its influence is; finally, compare Pre Merge and Pre Amvp , the mode with a higher probability is the current coding block C cur The optimal mode is executed, and the encoding of the current encoding block ends;

[0029] S9, when the reference frame of the inter-view prediction is a non-main view frame, the selection probability of the Merge mode and the AMVP mode is also calculated, and the prediction probability is calculated accordingly, and the mode with a higher prediction probability is selected as the optimal mode of the current coding block; if the reference frame is a main view frame, the coding block of the main view frame does not use the fast mode decision algorithm, so these coding blocks all save the cost after the execution of all prediction modes; extract the sum of the costs of these coding blocks when executing the Merge mode and the AMVP mode, and calculate their proportional coefficient β:

[0030]

[0031] Then, calculate the probability P that the adjacent coded blocks of the current block to be coded select the Merge and AMVP modes 2)*+,* and P 2-. / 0 , compare β×P 2)*+,* and P 2-. / 0 The mode corresponding to the larger value is selected as the optimal mode for the current coding block. After executing the mode, the coding of the current coding block is completed, and the coding of the current coding block ends.

[0032] In one possible implementation, in S1, the coarse intra-frame prediction refers to calculating only a small number of SATDs of prediction modes compared to conventional intra-frame prediction, including 8 modes in total, namely DC, PLANAR, ANGLE5, 10, 15, 20, 25 and 30. Then, the mode with the best SATD is selected to perform SATD calculations at two prediction angles with a step size of 2. Finally, the best SATD is selected as the cost and its corresponding prediction mode is selected as the optimal prediction angle of the coarse intra-frame prediction mode.

[0033] In a possible implementation, in S1, the coarse inter-frame prediction refers to performing a motion search in the reduced reference frame to find the best matching block, and calculating a SATD value of the residual as the cost of the coarse inter-frame prediction.

[0034] In a possible implementation, in S1, the coarse inter-viewpoint prediction refers to performing motion search using a referenced viewpoint frame at the same POC as a reference frame to find the best matching block, and calculating a residual SATD value as the cost of the coarse inter-viewpoint prediction.

[0035] In one possible implementation, the criterion for checking whether the coarse intra-frame cost of the current coding block is the optimal cost in S4 is: comparing the three costs obtained by coarse intra-frame prediction, coarse inter-frame prediction and coarse inter-viewpoint prediction. If the coarse intra-frame cost is the smallest, the coarse intra-frame cost is the optimal cost.

[0036] The present invention has the following beneficial effects:

[0037] (1) In existing multi-view video coding standards, such as the MV-HEVC standard, the prediction mode performed on the reference viewpoint is consistent with the prediction mode of the referenced viewpoint. Based on statistical observations, it is found that, except for the reference viewpoint of the key frame, the reference viewpoints of other types of frames (P, B frames) rarely or never select intra-frame prediction as the optimal prediction mode. In this case, executing redundant intra-frame prediction modes will increase the complexity of coding to a certain extent. In order to avoid this problem, the technical solution of the method of the present invention obtains the estimated mode of each block to be coded based on the cost of coarse intra-frame prediction and inter-frame prediction in the pre-analysis stage to skip the execution of redundant intra-frame prediction modes. In addition, the statistical information also shows the regional distribution characteristics of inter-frame prediction and inter-view prediction. In order to further improve the coding efficiency, the method of the present invention also makes further optimization decisions on the coding mode of the reference viewpoint based on a large amount of statistical information according to the temporal and viewpoint change characteristics of the video sequence, so as to select the optimal prediction mode with as few coding modes as possible, so as to reduce the overall coding complexity.

[0038] (2) Using the encoder's pre-analysis information (including SATD, mode, frame type, etc. for intra-frame prediction and inter-frame prediction based on 16*16CU), the prediction mode of each block to be encoded of the reference viewpoint is estimated to determine whether the intra-frame prediction mode needs to be skipped.

[0039] (3) A comprehensive decision is made using the prediction mode around the block to be encoded of the reference viewpoint, as well as the pre-analysis information, the corresponding reference viewpoint division information, the change type information of the video sequence, etc., to determine the optimal mode to be selected. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A flowchart of a fast mode decision method for multi-view video coding according to an embodiment of the present invention;

[0041] Figure 2 A schematic diagram of adjacent coded blocks of the current coded block;

[0042] Figure 3 Schematic diagram of the coding block and all adjacent coding blocks at the corresponding position of the reference frame. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] Reference Figure 1 , which is a flowchart of a fast mode decision method for multi-view video coding according to an embodiment of the present invention, comprising the following steps:

[0045] S1, based on the pre-analysis lookahead process, select a GOP number of frames to be encoded into the lookahead sequence, reduce the resolution of these encoded frames to 1 / 2 of the original storage, divide the encoding blocks into 16*16 pixels, perform coarse intra-frame prediction, coarse inter-frame prediction and coarse inter-viewpoint prediction, and store the three obtained costs into the data structure of the corresponding frame;

[0046] S2, extract N consecutive frames from any viewpoint of the current video sequence, N is determined by the frame rate of the video, and the video frames contained in 5 seconds are taken as the best. For each coding block in these N frames, extract the motion vector of the coarse inter-frame prediction, cluster the motion vector using the K-Means algorithm, and set the initial value of the number of cluster center clusters to 2. The number of cluster clusters can be continuously tried according to whether the actual clustering results are appropriate or not. The criterion for judging suitability is that almost all motion vectors are obviously clustered around a certain cluster center. After clustering is completed, calculate the Euclidean distance Distance(i,j) of the cluster center:

[0047]

[0048] Where i and j represent the i-th and j-th cluster centers respectively. Assuming the judgment threshold sh is 5.0, the current video sequence type Type is determined by the following formula:

[0049]

[0050] S3, encode the reference viewpoint, where it is assumed that the encoding of the reference viewpoint has been completed. Execute the Skip mode for the current coding block. If the coding block flag CBF after execution is 0, set the optimal mode of the coding block to the Skip mode and end the encoding of the current coding block; otherwise, execute S4;

[0051] S4, based on the obtained costs of coarse intra-frame prediction, coarse inter-frame prediction and coarse inter-viewpoint prediction, check whether the coarse intra-frame cost of the current coding block is the optimal cost; if so, further check the current frame type, if the current frame is a P frame, or the current frame is a B frame, but the cost difference percentage between the coarse intra-frame prediction and the coarse inter-frame prediction and coarse inter-viewpoint prediction is greater than 30%, then the current coding block executes the intra-frame prediction mode, and the optimal prediction mode of the coding block is set to the intra-frame prediction mode, and the encoding of the current coding block is ended; otherwise, execute S5;

[0052] S5, calculating the cost difference percentage between the coarse inter-frame prediction and the coarse inter-viewpoint prediction of the current coding block, if the percentage is less than 20%, executing S6; otherwise, if the cost of the coarse inter-frame prediction is smaller, executing S8; if the cost of the coarse inter-viewpoint prediction is smaller, executing S9;

[0053] S6, obtain the prediction mode distribution of the adjacent coded blocks of the current coding block. If the percentage of inter-frame prediction is high, execute S8; if the percentage of inter-viewpoint prediction is high, execute S9; otherwise execute S7; the high percentage of inter-frame prediction is the number of inter-frame prediction modes in all adjacent coded blocks. For example, if there are 5 adjacent coded blocks, 3 of which select inter-frame prediction, then the percentage of inter-frame prediction is 60%, and if there are 2 inter-viewpoint predictions, then the percentage of inter-viewpoint prediction is 40%.

[0054] S7, if the current video type is a translation type, execute S8; if the current video type is an irregular type, execute S9;

[0055] S8, Figure 2 and Figure 3 The adjacent coded blocks of the current coding block, as well as the coding block at the corresponding position of the reference frame and all adjacent coded blocks are shown respectively. Figure 2 The current encoding block is C cur , the adjacent coded block is C tl , C top , C tr , C left and C bl ; The reference coding block of the current block to be coded in the reference frame is R cur , whose adjacent coded reference coding block is R tl , R top , R tr , R left , R right , R bl , R bottom and R br According to the adjacent coded blocks of the current coding block, the coding blocks at the corresponding position of the reference frame and all adjacent coding blocks, for the inter-frame prediction of the referenced view, these coding blocks only perform the optimal prediction mode and store the corresponding optimal cost as the cost of the coding block; let the number of Merge modes selected by the 9 coding blocks in the referenced frame be C RMerge , then the probability of selecting the Merge mode is:

[0056]

[0057] The probability of selecting AMVP mode is:

[0058]

[0059] For the current coding block, there are five adjacent blocks that have been coded. In the special case where the coding block on the lower left is in the same CTU as the current coding block, it will be pre-coded. The probability P of selecting the Merge mode is also calculated. CMergeand the probability P of selecting the AMVP mode CAmvp ; Then the prediction probabilities of selecting Merge mode and AMVP mode for the current coding block are:

[0060] Pre Merge =α1P RMerge +α2P CMerge

[0061] Pre Amvp =α1P RAmvp +α2P CAmvp

[0062] Among them, α1 and α2 are weight coefficients, and their calculation formula is:

[0063]

[0064] Among them, POCdiff is the absolute value of the frame number difference between the reference frame and the reference frame; the value of α1 is inversely proportional to the absolute value, which means that the farther the reference frame is, the smaller its influence is; finally, compare Pre Merge and Pre Amvp , the mode with a higher probability is the current coding block C cur The optimal mode is executed, and the encoding of the current encoding block ends;

[0065] S9, when the reference frame of the inter-view prediction is a non-main view frame, the selection probability of the Merge mode and the AMVP mode is also calculated, and the prediction probability is calculated accordingly, and the mode with a higher prediction probability is selected as the optimal mode of the current coding block; if the reference frame is a main view frame, the coding block of the main view frame does not use the fast mode decision algorithm, so these coding blocks all save the cost after the execution of all prediction modes; extract the sum of the costs of these coding blocks when executing the Merge mode and the AMVP mode, and calculate their proportional coefficient β:

[0066]

[0067] Then, calculate the probability P that the adjacent coded blocks of the current block to be coded select the Merge and AMVP modes 2)*+,* and P 2-. / 0 , compare β×P 2)*+,* and P 2-. / 0 The mode corresponding to the larger value is selected as the optimal mode for the current coding block. After executing the mode, the coding of the current coding block is completed, and the coding of the current coding block ends.

[0068] Specifically, in S1, coarse intra-frame prediction means that compared with conventional intra-frame prediction, only a small number of prediction modes' SATD are calculated, including 8 types, namely DC, PLANAR, ANGLE5, 10, 15, 20, 25 and 30. Then, the mode with the best SATD is selected to perform SATD calculations for two prediction angles with a step size of 2. Finally, the best SATD is selected as the cost and its corresponding prediction mode is used as the optimal prediction angle of the coarse intra-frame prediction mode. Coarse inter-frame prediction refers to performing motion search in the reduced reference frame to find the best matching block, and calculating the SATD value of the residual as the cost of coarse inter-frame prediction. Coarse inter-viewpoint prediction refers to performing motion search in the referenced viewpoint frame at the same POC as the reference frame to find the best matching block, and calculating the SATD value of the residual as the cost of coarse inter-viewpoint prediction.

[0069] Specifically, the criterion for checking whether the coarse intra-frame cost of the current coding block is the optimal cost in S4 is: comparing the three costs obtained by coarse intra-frame prediction, coarse inter-frame prediction and coarse inter-viewpoint prediction. If the coarse intra-frame cost is the smallest, the coarse intra-frame cost is the optimal cost.

[0070] In a specific application example, Table 1 shows the objective quality change and encoding speed change of the method of the present invention compared with the original encoder under three 1080p and three 720p 3D sequence tests. It can be seen that the method of the present invention can reduce the encoding complexity by about 35% on average, and the performance loss is 0.09% to 0.38%. Based on the motion type of the video (irregular and translational), the performance loss of irregular video sequences is lower.

[0071] Table 1 Objective quality change and encoding speed change of the method of the embodiment of the present invention compared with the original encoder

[0072]

[0073] It should be understood that the exemplary embodiments described herein are illustrative rather than restrictive. Although one or more embodiments of the present invention have been described in conjunction with the accompanying drawings, it should be understood by those skilled in the art that various changes in form and detail may be made without departing from the spirit and scope of the present invention as defined by the appended claims.

Claims

1. A fast mode decision method for multi-view video coding, characterized in that: The following steps are involved: S1, based on the pre-analysis lookahead process, select a GOP number of frames to be encoded into the lookahead sequence, reduce the resolution of these encoded frames to 1 / 2 of the original storage, divide the encoding blocks into 16*16 pixels, perform coarse intra-frame prediction, coarse inter-frame prediction and coarse inter-viewpoint prediction, and store the three obtained costs into the data structure of the corresponding frame; S2, extract N consecutive frames from any viewpoint of the current video sequence, extract the coarse inter-frame predicted motion vector for each coding block in the N frames, cluster the motion vectors using the K-Means algorithm, and after clustering, calculate the Euclidean distance Distance(i,j) of the cluster center: Where i and j represent the i-th and j-th cluster centers respectively. Assuming the judgment threshold sh is 5.0, the current video sequence type Type is determined by the following formula: S3, encode the reference viewpoint, execute the Skip mode for the current coding block, if the coding block flag CBF after execution is 0, set the optimal mode of the coding block to the Skip mode, and end the encoding of the current coding block; Otherwise, execute S4; S4, based on the obtained costs of coarse intra-frame prediction, coarse inter-frame prediction and coarse inter-viewpoint prediction, check whether the coarse intra-frame cost of the current coding block is the optimal cost; if so, further check the current frame type, if the current frame is a P frame, or the current frame is a B frame, but the cost difference percentages of the coarse intra-frame prediction and the coarse inter-frame prediction and coarse inter-viewpoint prediction are greater than 30%, then the current coding block executes the intra-frame prediction mode, and the optimal prediction mode of the coding block is set to the intra-frame prediction mode, and the encoding of the current coding block is ended; Otherwise, execute S5; S5, calculating the cost difference percentage between the coarse inter-frame prediction and the coarse inter-viewpoint prediction of the current coding block, if the percentage is less than 20%, executing S6; otherwise, if the cost of the coarse inter-frame prediction is smaller, executing S8; if the cost of the coarse inter-viewpoint prediction is smaller, executing S9; S6, obtaining the prediction mode distribution selected by the adjacent coded coding blocks of the current coding block, if the percentage of inter-frame prediction is high, executing S8; if the percentage of inter-view prediction is high, executing S9; otherwise, executing S7; S7, if the current video type is a translation type, execute S8; if the current video type is an irregular type, execute S9; S8, according to the adjacent coded blocks of the current coding block, the coding blocks at the corresponding position of the reference frame and all adjacent coding blocks, for the inter-frame prediction of the referenced view, these coding blocks only perform the optimal prediction mode, and store the corresponding optimal cost as the cost of the coding block; let the number of Merge modes selected by the 9 coding blocks in the referenced frame be C RMerge , then the probability of selecting the Merge mode is: The probability of selecting AMVP mode is: For the current coding block, there are five adjacent blocks that have been coded. In the special case where the coding block on the lower left is in the same CTU as the current coding block, it will be pre-coded. The probability P of selecting the Merge mode is also calculated. CMerge and the probability P of selecting the AMVP mode CAmvp ; Then the prediction probabilities of selecting Merge mode and AMVP mode for the current coding block are: Before )*+,* =α1P ()*+,* +a # P 2)*+,* Before -. / 0 =α1P (-. / 0 +a # P 2-. / 0 Among them, α1 and α # is the weight coefficient, and its calculation formula is: α # ＝1-α1 Among them, POCdiff is the absolute value of the frame number difference between the reference frame and the reference frame; the value of α1 is inversely proportional to the absolute value, which means that the farther the reference frame is, the smaller its influence is; finally, compare Pre )*+,* and Pre -. / 0 , the mode with a higher probability is the current coding block C cur The optimal mode is executed, and the encoding of the current encoding block ends; S9, when the reference frame of the inter-view prediction is a non-main view frame, the selection probability of the Merge mode and the AMVP mode is also calculated, and the prediction probability is calculated accordingly, and the mode with a higher prediction probability is selected as the optimal mode of the current coding block; if the reference frame is a main view frame, the coding block of the main view frame does not use the fast mode decision algorithm, so these coding blocks save the cost after the execution of all prediction modes; extract the sum of the costs of these coding blocks when executing the Merge mode and the AMVP mode, and calculate their proportional coefficient β: Then, calculate the probability P that the adjacent coded blocks of the current block to be coded select the Merge and AMVP modes 2)*+,* and P 2-. / 0 , compare β×P 2)*+,* and P 2-. / 0 The mode corresponding to the larger value is selected as the optimal mode for the current coding block. After executing the mode, the coding of the current coding block is completed, and the coding of the current coding block ends.

2. The fast mode decision method for multi-view video coding according to claim 1, characterized in that: In S1, the coarse intra-frame prediction means that compared with the conventional intra-frame prediction, only a small number of prediction modes of SATD are calculated, including 8 types including DC, PLANAR, ANGLE5, 10, 15, 20, 25 and 30. Then, the mode with the best SATD is selected to perform SATD calculations for two prediction angles with a step size of 2. Finally, the best SATD is selected as the cost and its corresponding prediction mode is selected as the optimal prediction angle of the coarse intra-frame prediction mode.

3. The fast mode decision method for multi-view video coding according to claim 1, characterized in that: In S1, the coarse inter-frame prediction refers to performing motion search in the reduced reference frame to find the best matching block, and calculating the SATD value of the residual as the cost of the coarse inter-frame prediction.

4. The fast mode decision method for multi-view video coding according to claim 1, characterized in that: In S1, the coarse inter-viewpoint prediction refers to performing motion search on the referenced viewpoint frame in the same POC as a reference frame to find the best matching block, and calculating the SATD value of the residual as the cost of the coarse inter-viewpoint prediction.

5. The fast mode decision method for multi-view video coding according to any one of claims 1 to 4, characterized in that: In S4, the judgment criterion for checking whether the coarse intra-frame cost of the current coding block is the optimal cost is: comparing the three costs obtained by coarse intra-frame prediction, coarse inter-frame prediction and coarse inter-viewpoint prediction. If the coarse intra-frame cost is the smallest, the coarse intra-frame cost is the optimal cost.

Images