CN105491390B - Intra-frame prediction method in hybrid video coding standard - Google Patents

Abstract

The invention relates to an intra-frame prediction method in a hybrid video coding standard, which belongs to the field of video coding. The purpose of the present invention is to effectively deal with complex blocks existing in video sequences, such as video blur caused by object or camera motion, multi-directional complex blocks, etc., and propose a hybrid video coding standard Intra-frame prediction method, to Further improve the performance of video encoding. In this intra prediction method, two different prediction modes are used to obtain two different prediction values. A new prediction for the current coded block is obtained by weighting these two predictors. Obtain intra-frame coding mode information of several adjacent coded blocks around the current coding block, select one of the modes as mode 1; on the basis of mode 1, select another intra-frame mode as mode 2. The complex block in the video sequence can be processed by using the predicted value synthesized by two different prediction modes, so that the coding efficiency is further improved.

Description

Intra Prediction Method in Hybrid Video Coding Standard

technical field

The invention relates to an intra-frame prediction method in a mixed video coding standard, which belongs to the field of video coding.

Background technique

With the improvement of people's requirements for video display quality, new video application forms such as high-definition and ultra-high-definition video have emerged. In the situation that the high-resolution and high-quality video appreciation is more and more widely used, how to enhance the video compression efficiency becomes crucial. During the digitization of images and videos, a large amount of data redundancy is generated, which makes video compression technology possible. Generally speaking, redundancy types include at least spatial redundancy, time redundancy, and information entropy redundancy. For the elimination of spatial redundancy, a method based on prediction is generally used, that is, intra-frame prediction coding. The basic idea is to use the reconstructed pixel values around the current coding block to generate the prediction value of the current block through direction-based interpolation. After the predicted block is obtained, the difference between the current block and the predicted block, that is, the residual block, is easier to code than the original coded block. Intra-frame prediction effectively reduces spatial redundancy in video coding. Due to the unidirectional interpolation prediction adopted by the intra prediction in the existing video coding standards, this method cannot predict complex blocks.

To handle complex coded blocks in video sequences, Y.Ye and M.Karczewicz, “Improved H.264 intra coding based on bi-directional intra prediction, directional transform, and adaptive coefficient scanning,” in Proc.IEEE Int.Conf.Image Process., Oct.2008, pp.2116–2119. A bidirectional intra-frame prediction coding method is proposed. The method is based on nine prediction modes in the H.264/AVC video coding standard, and selects a certain number of combinations of the two modes. For each combination, an offline-trained weight table is used to weight-average the predictions produced by the two modalities. There is still the problem of poor encoding performance of the video.

Contents of the invention

The purpose of the present invention is to effectively process complex blocks in a video sequence, and propose an intra-frame prediction method in a hybrid video coding standard, so as to further improve video coding performance.

The technical scheme that the present invention takes for solving the problems of the technologies described above is:

An intra-frame prediction method in a hybrid video coding standard, the prediction method is used to describe complex coding blocks existing in a video sequence, and the implementation process of the prediction method is as follows:

Step 1: Obtain the intra-frame coding modes of several adjacent coded blocks around the current coded block. The size of the current coded block is W*H, W is the width of the current coded block, and H is the height of the current coded block; Adjacent coded blocks are called adjacent coded blocks;

Step 2: Obtain a set of coding modes 1 of the current coding block according to the intra-frame coding modes of adjacent coding blocks obtained in step 1;

Step 3: Obtain the corresponding mode 2 according to each mode 1 in the set of coding mode 1: select one of the other two modes closest to mode 1 in the direction as mode 2, or select the mode 2 with the minimum value after combining with mode 1 The mode of prediction distortion is mode 2,

According to the set of encoding mode 1 obtained in step 2, obtain another set of encoding modes of the current encoding block for each mode in the set of encoding mode 1, that is, the set of encoding mode 2; merge the set of mode 1 and the set of mode 2 , to get a set of two-tuples, each two-tuple contains two related patterns 1 and 2;

Step 4: For each mode combination in the 2-tuple set generated in step 3, interpolate adjacent pixels around the current block to obtain two different prediction blocks; a bidirectional prediction result of the current coding block is the two different predictions The weighted average block of the block; select the optimal combination of mode 1 and mode 2 to predict the current block;

Step 5: Select the optimal prediction mode for the luma block and the chrominance block in the coding unit respectively;

Step 6: Encode the encoding modes of the luma block and the chrominance block in the coding unit respectively.

In step 1, the adjacent coding blocks are intra-frame coding blocks that have been coded on the left, top, bottom left, and top right of the current coding block;

The collection principle for obtaining the coding mode 1 of the current coding block described in step 2 is:

Select the modes with the most usage modes of the adjacent coding blocks obtained in step 1 as mode 1, or select the mode of the adjacent coding blocks on the left and above of the current coding block as mode 1, or select the mode of any one of these adjacent blocks Mode 1, or select a subset of the modes of these adjacent blocks as mode 1, or specify a weight for each adjacent coding block, accumulate the weights of adjacent coding blocks with the same intra-frame coding mode, and select step 1 The modes with the largest weights of the adjacent coding blocks acquired in are mode one.

In step 3, the implementation process of acquiring mode 2 of the current coding block for each mode 1 is:

Select the two encoding modes closest to the current mode 1 direction as mode 2, the specific process is: mode 1 is represented by mode1, if the mode of mode 1 is between 3 and 33, mode 2 is selected as mode1-1 and mode1+ 1; if the value of mode1 is 2 or 34, mode 2 is selected as 3 and 33; if the value of mode1 is DC mode or PLANAR mode, mode 2 is selected as 10 (horizontal mode) and 26 (vertical mode);

Or select the mode with the smallest prediction distortion after combining with mode 1 as mode 2, the implementation process is: for each mode 1, obtain the predicted values corresponding to all remaining intra-frame coding modes, and then compare the predicted values of mode 1 with each The predictions corresponding to the remaining intra-frame modes are weighted and averaged, and the coding mode with the smallest distortion after the weighted average with mode 1 and the current coding block is selected as mode 2; the criterion for the distortion between the coding block and the prediction block can be: minimum mean square error, minimum Hadamard error or rate-distortion optimization criterion.

In step 4, when performing a coding test on each coding mode group in the binary group set, select the optimal mode group to predict the current block; the selection of the optimal mode group can use the minimum mean square error, minimum Hadamard Error or rate-distortion optimization criteria.

In step 4, the process of weighting the predictions produced by two different prediction modes is: different weights are given to the prediction blocks of different prediction modes; the weighted average can use the same Weights, that is, average them to obtain the prediction block of the current coding block; or assign different weights according to the importance of different prediction modes, or assign different weights according to the accuracy of predictions generated by different prediction modes, or set Set some weights with high probability, and use the search traversal method to obtain the best weights.

In step five, the process of selecting the best prediction mode for the luma block and chrominance block in the coding unit is: for the luma block, the optimal prediction mode is selected from the original unidirectional prediction mode and bidirectional prediction mode , the selected criterion is the minimum rate-distortion criterion; and for a chroma block, if its corresponding luma block selects bidirectional prediction as the best prediction mode, then the optimal prediction mode of the current chroma block is to select the two prediction modes of its corresponding luma block. predictive mode.

In step six: respectively encode the encoding modes of the luma block and the chrominance block in the coding unit, and the specific process is as follows:

If the current intra-frame coding mode is bi-directional prediction, two coding modes in the bi-directional prediction, ie, mode 1 and mode 2, need to be coded; mode 1 comes from adjacent coding blocks, and directly codes the index of the selected neighboring block.

For luminance blocks, when mode 1 is selected from the left or top of the current block, a 1-bit symbol can be used to indicate that the selected mode is from the left or the top; coding mode 2 is obtained based on mode 1, similarly, A one-bit sign can be used to indicate which of the mode-adjacent modes the selected mode is;

For the chroma block, if the current luma block selects bidirectional prediction, the prediction mode of the chroma block will be set to bidirectional prediction mode, and the corresponding two prediction modes directly come from the luma block, and there is no need to encode the prediction mode; if The current luma block selects the original unidirectional prediction, and the chroma block will select from the original five prediction modes.

The beneficial effects of the present invention are:

The prediction method of the present invention can effectively deal with complex blocks existing in video sequences, such as video blur caused by object or camera motion, multi-directional complex blocks, and the like. The present invention uses the intra-frame mode information of adjacent coded blocks to obtain two modes of the current coded block, and bidirectional prediction based on these two modes can predict complex blocks in video sequences, such as blocks with multiple directions, objects and blurred blocks caused by camera motion, so that the performance of intra-frame prediction is improved, and the coding efficiency is further improved.

In this intra prediction method, two different prediction modes are used to obtain two different prediction values. A new prediction for the current coded block is obtained by weighting these two predictors. Obtain intra-frame coding mode information of several adjacent coded blocks around the current coding block, select one of the modes as mode 1; on the basis of mode 1, select another intra-frame mode as mode 2. The complex block in the video sequence can be processed by using the predicted value synthesized by two different prediction modes, so that the coding efficiency is further improved.

Different from the methods proposed before, the solution of the present invention does not need to train and save the weight table when performing bidirectional prediction. The two different modes in the bidirectional prediction mode of the solution of the present invention are obtained based on the adjacent blocks of the current block, so the bit overhead for encoding these two modes is small. In addition, the solution of the present invention requires fewer combinations of different modes to be tested, because the encoding complexity is lower.

Description of drawings

FIG. 1 is a positional diagram of the current block (C), the left adjacent block (L) and the upper adjacent block (A) in Embodiment 2 of the present invention.

FIG. 2 is a relationship diagram between coding mode 1 and candidate coding mode 2 and mode 1 for the current coding block in Embodiment 3 of the present invention. In the figure, mode 1 is 3, and mode 2 is mode 2 or 4 with the closest angle.

Detailed ways

Specific Embodiment 1: The intra prediction method in the hybrid video coding standard described in this embodiment is used to predict the complex coding blocks existing in the video sequence, and the prediction method is based on the direction-based intra prediction method in the original coding standard ( The prediction method is realized based on a unidirectional intra-frame prediction algorithm),

The prediction method described above is called a bidirectional prediction method based on adjacent coding modes, or bidirectional prediction for short. The bidirectional prediction method has two encoding modes, namely mode one and mode two; the implementation process of the prediction method is:

Step 1: Obtain the intra-frame coding modes of several adjacent coded blocks around the current coded block. The size of the current coded block is W*H, W is the width of the current coded block, and H is the height of the current coded block; Adjacent coded blocks are called adjacent coded blocks;

Step 2: Obtain a set of coding modes 1 of the current coding block according to the intra-frame coding modes of adjacent coding blocks obtained in step 1;

Step 3: Obtain the corresponding mode 2 according to each mode 1 in the set of coding mode 1: select one of the other two modes closest to mode 1 in the direction as mode 2, or select the mode 2 with the minimum value after combining with mode 1 The mode of predictive distortion is mode 2; according to the set of encoding mode 1 obtained in step 2, for each mode in the set of encoding mode 1, another set of encoding modes of the current encoding block is obtained, that is, the set of encoding mode 2; the merge mode The set of one and the set of mode two, get a set of two-tuples, each two-tuple contains two related modes one and two;

Step 4: For each mode combination in the 2-tuple set generated in step 3, interpolate adjacent pixels around the current block to obtain two different prediction blocks; a bidirectional prediction result of the current coding block is the two different predictions The weighted average block of the block; select the optimal combination of mode 1 and mode 2 to predict the current block;

Step 5: Select the optimal prediction mode for the luma block and the chrominance block in the coding unit respectively;

Step 6: Encode the encoding modes of the luma block and the chrominance block in the coding unit respectively.

Specific implementation mode two: the intra-frame prediction method in the hybrid video coding standard described in this implementation mode is characterized in that:

In step 1, the adjacent coding blocks are intra-frame coding blocks that have been coded on the left, top, bottom left, and top right of the current coding block. As shown in Figure 1, we can select the left block L and the top block A of the current block;

It also supports selecting a larger number of adjacent blocks to obtain the intra prediction mode 1 of the current coding block, except for the adjacent blocks on the left, upper, lower left, and upper right, adjacent blocks in other positions are also supported;

The collection principle for obtaining the coding mode 1 of the current coding block described in step 2 is:

Select the modes with the most usage modes of the adjacent coding blocks obtained in step 1 as mode 1, or select the mode of the adjacent coding blocks on the left and above of the current coding block as mode 1, or select the mode of any one of these adjacent blocks Mode 1, or select a subset of the modes of these adjacent blocks as mode 1, or specify a weight for each adjacent coding block, accumulate the weights of adjacent coding blocks with the same intra-frame coding mode, and select step 1 Mode 1 is the mode with the largest weight value of adjacent coding blocks obtained in ; as shown in Figure 1, the intra prediction mode corresponding to the left block L and the upper block A of the current block can be selected as the intra coding mode 1 of the current block.

Other steps are the same as in the first embodiment.

Specific embodiment 3: As shown in FIG. 2 , in the intra prediction method in the hybrid video coding standard described in this embodiment, in step 3, the process of obtaining the intra frame coding mode 2 of the current coding block is:

Select the two encoding modes closest to the current mode one direction as mode two. The implementation process is, if the mode of mode 1 (mode1) is between 3 and 33, mode 2 is selected as mode1-1 and mode1+1; if the value of mode1 is 2 or 34, mode 2 is selected as 3 and 33 ; If the value of mode1 is DC mode or PLANAR mode, mode 2 is selected as 10 (horizontal mode) and 26 (vertical mode).

Or, in step 3, the process of obtaining the intra-frame coding mode 2 of the current coding block is:

The implementation process of selecting the mode with the smallest prediction distortion after combining with mode 1 as mode 2 is: for each mode 1, obtain the predicted values corresponding to all remaining intra coding modes, and then combine the predicted values of mode 1 with each remaining The prediction corresponding to the intra mode is weighted and averaged, and the coding mode with the least distortion of the current coding block after the weighted average with mode 1 is selected as mode 2. Here, the criterion for the distortion of the coding block and the prediction block may be: minimum mean square error, minimum Hadamard error or rate-distortion optimization criterion.

Other steps are the same as those in Embodiment 1 or 2.

Embodiment 4: In the intra-frame prediction method in the hybrid video coding standard described in this embodiment, in step 4, when performing a coding test on each coding mode group in the binary group set, the optimal mode group is selected to perform the coding test. The current block is predicted. The optimal mode set can be selected by: minimum mean square error, minimum Hadamard error or rate-distortion optimization criteria. Other steps are the same as those in Embodiment 1, 2 or 3.

Embodiment 5: In the method for intra-frame prediction in the hybrid video coding standard described in this embodiment, in step 4, the process of weighting the predictions generated by two different prediction modes is: giving prediction blocks of different prediction modes Unused weights. The weighted average here can adopt the same weight for these two different prediction blocks, that is, average them to obtain the prediction block of the current coding block. Or assign different weights according to the importance of different prediction modes, or assign different weights according to the accuracy of predictions generated by different prediction modes, or set some weights with higher probability, and use search traversal to obtain the best weight. Other steps are the same as those in Embodiment 1, 2, 3 or 4.

Specific embodiment six: In the intra prediction method in the hybrid video coding standard described in this embodiment, in step five, the process of selecting the best prediction mode for the luma block and chrominance block in the coding unit is: for the luma block, The optimal prediction mode is selected from the original unidirectional prediction mode and the bidirectional prediction mode provided by the present invention, and the selected criterion is the minimum rate-distortion criterion. For a chroma block, if the corresponding luma block selects bidirectional prediction as the best prediction mode, then the optimal prediction mode of the current chroma block is to select the two prediction modes of its corresponding luma block. Other steps are the same as those in Embodiment 1, 2, 3, 4 or 5.

Embodiment 7: In the intra-frame prediction method in the hybrid video coding standard described in this embodiment, the intra-frame mode coding method is: if the current intra-frame coding mode is bidirectional prediction, the two coding modes in the bidirectional prediction, namely mode One and two need to be encoded. Mode one is derived from adjacent coded blocks, and directly encodes the index of the selected adjacent block. Other steps are the same as those in Embodiment 1, 2, 3, 4, 5 or 6.

Embodiment 8: In this embodiment, mode 1 is selected from the left or top of the current block, and a 1-bit symbol can be used to indicate whether the selected mode is from the left or the top. Coding mode 2 is obtained based on mode 1. Similarly, a symbol of one bit can be used to indicate which mode is selected to be one of the neighboring modes of mode 1. For the chroma block, if the current luma block selects bidirectional prediction, the prediction mode of the chroma block will be set to the bidirectional prediction mode, and the corresponding two prediction modes directly come from the luma block. Therefore there is no need to encode the prediction mode. If the current luma block selects the original unidirectional prediction, the chroma block will choose from the original five prediction modes. Other steps are the same as those in Embodiment 1, 2, 3, 4, 5, 6 or 7.

Example

Embodiment one:

The specific implementation steps of the intra prediction method in the hybrid video coding standard are given:

Step 1: Obtain the coding modes modeL and modeA of the left adjacent block and the upper adjacent block of the current coding block (the size is W*H, W is the width of the current coding block, and H is the height of the current coding block);

Step 2: Obtain mode 1 of the current coding block according to the coding modes of adjacent coding blocks obtained in step 1. If modeL is equal to modeA, then the mode one set of the current coding block is {modeL}; if modeL is not equal to modeA, then the mode one set of the current coding block is {modeL, modeA};

Step 3: Obtain mode 2 of the current coding block according to the mode 1 set of the current coding block acquired in step 2. For each mode modei in the mode one set, if modei is between 3 and 33, mode two is selected as modei-1 and modei+1; if the value of modei is 2 or 34, mode two is selected as 3 and 33; if the value of modei is DC mode or PLANAR mode, mode 2 is selected as 10 (horizontal mode) and 26 (vertical mode). For each mode in the mode 1 set, select its corresponding mode 2 set, and you can get a 2-tuple set combining mode 1 and mode 2. Each high-end element in the set is composed of the corresponding mode 1 and mode 2. That is (mode1, mode2).

Step 4: For each mode combination (mode1, mode2) in the 2-tuple set generated in step 3, interpolate adjacent pixels around the current block to obtain two different prediction blocks, namely pred1 and pred2. A bidirectional prediction result pred of the current coding block is the average of the two different prediction blocks, that is, pred=(pred1+pred2+1)>>1, and the optimal combination of mode 1 and mode 2 is selected by using rate-distortion optimization. The current block is predicted.

Step 5: Select the optimal prediction mode for the luma block and the chrominance block in the coding unit respectively. For the luma block, the optimal prediction mode is selected from the original unidirectional prediction mode and the bidirectional prediction mode provided by the present invention, and the selected criterion is the minimum rate-distortion criterion. For a chroma block, if the corresponding luma block selects bidirectional prediction as the best prediction mode, then the optimal prediction mode of the current chroma block is to select the two prediction modes of its corresponding luma block.

Step 6: Encode the encoding modes of the luma block and the chrominance block in the coding unit respectively. If the current intra-frame coding mode is bi-directional prediction, two coding modes in the bi-directional prediction, ie mode 1 and mode 2, need to be coded. Mode one is derived from adjacent coded blocks, and directly encodes the index of the selected adjacent block. For example, mode 1 is selected from the left or top of the current block, and a 1-bit symbol can be used to indicate whether the selected mode is from the left or the top. Coding mode 2 is obtained based on mode 1. Similarly, a symbol of one bit can be used to indicate which mode is selected to be one of the neighboring modes of mode 1. For the chroma block, if the current luma block selects bidirectional prediction, the prediction mode of the chroma block will be set to the bidirectional prediction mode, and the corresponding two prediction modes directly come from the luma block. Therefore there is no need to encode the prediction mode. If the current luma block selects the original unidirectional prediction, the chroma block will choose from the original five prediction modes.

The above content is a further detailed description of the present invention in conjunction with specific preferred embodiments, and it cannot be assumed that the specific implementation of the present invention is limited to these descriptions. The scope of the invention is preferably referred to in the appended claims. For those of ordinary skill in this technical field, without departing from the concept of the present invention, some simple deduction or replacement can also be made, which should be regarded as belonging to the scope of patent protection determined by the claims submitted by the present invention .

Embodiment 1 is implemented on VC-0.4 (a test model with some technologies added to the HEVC test model HM12.0), and is tested according to the VC266 pass-through test conditions. For VC266 pass-test conditions, refer to VC266 Study Group, "Test condition and evaluation methodology" ,VC-02-N005,VC266 2th Meeting:Suzhou,Mar.2015.

The experimental results of Example 1 are shown in Table 1. It can be seen from Table 1 that compared with VC-0.4, under the All Intra Main_HighBitrate (AI-HR) configuration condition, there are 0.8% and 0.6% on average for Y, U and V components. % and 1.1% BD bitrate savings, with an average BD bitrate savings of 0.7%, 0.4% and 0.6% for Y, U and V components under All Intra Main_LowBitrate (AI-LR) configuration. The BD bit rate indicates the bit rate savings of the two methods under the same objective quality, refer to G. "Calculation of average PSNR differences between RD-Curves," ITU-T SG16 Q.6 Document, VCEG-M33, Austin, US, April 2001.

Table 1. Embodiment 1 relative to the BD bit rate performance of VC-0.4

Claims (8)

1. intra-frame prediction method in a kind of hybrid video coding standard, the prediction technique is for describing present in video sequence Complicated encoding block, which is characterized in that the realization process of the prediction technique is：

Step 1：The intra-frame encoding mode of several adjacent coded blocks around present encoding block is obtained, present encoding block Size is W*H, and W is the width of present encoding block, and H is the height of present encoding block；Several adjacent coded blocks of surrounding are referred to as neighbouring Encoding block；

Step 2：The coding mode of present encoding block is obtained according to the intra-frame encoding mode of the neighbouring encoding block of step 1 acquisition One set；

Step 3：Corresponding pattern two is obtained according to each pattern one in the set of coding mode one：It chooses and exists from pattern one One on direction in other two nearest pattern is pattern two, or is chosen after pattern one is combined with minimum prediction The pattern of distortion is pattern two,

According to the set for the coding mode one that step 2 obtains, each pattern acquiring present encoding in gathering coding mode one The set of another coding mode of block, the i.e. set of coding mode two；The set of the set and pattern two of merging patterns one, Two tuple-sets are obtained, each two tuple includes two relevant patterns one and pattern two；

Step 4：Each mode combinations in two tuple-sets generated for step 3, with pixel adjacent around current block Interpolation obtains two different prediction blocks；One bi-directional predicted result of present encoding block be this two different prediction blocks plus Weight average block；The combination of optimal pattern one and pattern two is selected to predict current block；

Step 5：To in coding unit luminance block and chrominance block carry out the selections of optimal prediction modes respectively；

Step 6：The coding mode of luminance block and chrominance block in coding unit is encoded respectively.

2. intra-frame prediction method in hybrid video coding standard according to claim 1, it is characterised in that：

In step 1, the neighbouring encoding block is the left side of present encoding block, top, lower-left, in the coded frame of upper right Encoding block；

The aggregation of coding mode one that present encoding block is obtained described in step 2 is：

The several patterns of the neighbouring encoding block use pattern obtained in selecting step one at most are pattern one, or the current volume of selection The pattern of the neighbouring encoding block on the code block left side and top is pattern one, or selects these adjacent to the mould of any one block in the block Formula is pattern one, either selects a subset of the pattern of these contiguous blocks for pattern one or refers to for each neighbouring encoding block A fixed weights, the weights of the neighbouring encoding block with identical intra-frame encoding mode are added up, the neighbour obtained in selecting step one Several patterns of the maximum weight of nearly encoding block are pattern one.

3. intra-frame prediction method in hybrid video coding standard according to claim 1, it is characterised in that：In step 3 In, the realization process that the pattern two of present encoding block is obtained for each pattern one is：

It is pattern two to choose with two nearest coding modes of one direction of present mode, and detailed process is：The mode1 tables of pattern one Show, if the pattern of pattern one is between 3 and 33, pattern two is chosen for mode1-1 and mode1+1；If the value of mode1 It is 2 or 34, pattern two is chosen for 3 and 33；If the value of mode1 is DC patterns or PLANAR patterns, pattern two is selected It is taken as 10 and 26；

Or the pattern chosen with minimum predicted distortion after pattern one is combined is pattern two, realization process is：For every A pattern one obtains the corresponding predicted value of all remaining intra-frame encoding modes, then by the predicted value of pattern one and each residue The corresponding prediction of frame mode is weighted average, selection and volume minimum with the distortion of present encoding block after one weighted average of pattern Pattern is pattern two；The criterion of the distortion of encoding block and prediction block can be：Least mean-square error, minimum Hadamard errors Or rate-distortion optimization criterion.

4. intra-frame prediction method in hybrid video coding standard according to claim 1, it is characterised in that：In step 4 In, in two tuple-sets each coding mode group carry out encoded test when, select optimal modal sets come to current block into Row prediction；Select optimal modal sets that least mean-square error, minimum Hadamard errors or rate-distortion optimization criterion can be used.

5. intra-frame prediction method in hybrid video coding standard according to claim 1, it is characterised in that：In step 4 In, the prediction of the prediction mode different to two generation, which is weighted process, is：The prediction block of different prediction modes is given Different weights；The weighted average, which may be used, gives both identical weights of different prediction blocks, that is, to they into Row averagely obtains the prediction block of present encoding block；Or different weights are assigned according to the importance of different prediction modes, or The accuracy that person generates prediction according to different prediction modes assigns different weights, or sets the higher weights of some probability, Best weights are obtained with the mode of search spread.

6. intra-frame prediction method in hybrid video coding standard according to claim 5, it is characterised in that：In step 5 In, the process for luminance block and chrominance block selection optimum prediction mode in coding unit is：It is optimal for luminance block Prediction mode is chosen from original one-direction prediction modes and bi-predictive mode, and the criterion of selection is minimum rate distortion Criterion；And for chrominance block, if its corresponding luminance block select it is bi-directional predicted for optimum prediction mode, current chroma block Optimal prediction modes are two prediction modes for selecting its corresponding brightness block.

7. intra-frame prediction method in hybrid video coding standard according to claim 1 or 6, it is characterised in that：

In step 6：The coding mode of luminance block and chrominance block in coding unit is encoded respectively, detailed process For：

If current intra-frame encoding mode is bi-directional predicted, two coding modes during this is bi-directional predicted, i.e. pattern one and pattern Two are encoded；Pattern one comes from neighbouring encoding block, the index of the selected contiguous block of direct coding.

8. intra-frame prediction method in hybrid video coding standard according to claim 7, it is characterised in that：

For luminance block, when pattern one is to choose to obtain from the left side of current block or top, then the symbol of 1 bit For indicating that the pattern chosen comes from the left side or top；Coding mode based on pattern one second is that obtained, similarly, a ratio Special symbol can be used for indicating to be selected pattern is which of one adjacent modes of pattern；

For chrominance block, if the selection of present intensity block is bi-directional predicted, the prediction mode of chrominance block will be set to bi-directional predicted Pattern, corresponding two prediction modes directly luminance block, woth no need to be encoded to prediction mode；If present intensity Block selects original single directional prediction, chrominance block that will be chosen from five original prediction modes.