WO2009089779A1

WO2009089779A1 - An intra frame predicting method and apparatus

Info

Publication number: WO2009089779A1
Application number: PCT/CN2009/070057
Authority: WO
Inventors: Yingjia Liu; Sixin Lin; Shan Gao; Jiantong Zhou; Jiali Fu; Qingpeng Xie; Xiaosen Zhang; Lianhuan Xiong
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2008-01-07
Filing date: 2009-01-07
Publication date: 2009-07-23
Also published as: CN101483780A; CN101483780B

Abstract

An intra frame predicting method is provided. It includes obtaining a current block in the frame of which the pixel values are to be predicted, and determining the pixel predicting values of the current block in the texture direction of the current block according to the predicting mode direction of a neighbor block of the current block and the reference pixel value of the neighbor block. An intra frame predicting apparatus is also provided. With the application of the invention, the continuous characteristic of the texture distribution of the current block is maintained when the pixel values of the current block are predicted, the veracity of the prediction is ensured, and it is easy to be carried out.

Description

Method and device for intra prediction

The present application claims priority to Chinese Patent Application No. 200810025656.3, entitled "A Method and Apparatus for In-Frame Prediction", filed on January 7, 2008, the entire contents of which is incorporated herein by reference. In the application. Technical field

The present invention relates to the field of communications, and in particular, to a method for intra prediction and an intra prediction device. Background technique

In video coding, the intraframe coding compression rate is low, and the intra coding does not depend on the previous coded frame, so that decoding can be performed without obtaining the previous frame. The intra-coded block uses the surrounding information for prediction. The luminance sub-blocks of (4 pixels x 4 pixels) or the luminance blocks of (8 pixels x 8 pixels) each have 9 different prediction modes, specifically 8 different types. Prediction mode direction prediction and a direct current (DC) prediction. The luminance macroblocks of size (16 pixels x 16 pixels) have 4 different prediction modes, which are vertical, horizontal, DC and plane prediction modes. The prediction can be made using the pixels around the currently encoded sub-block/block/macroblock as reference pixels.

In H.264 video coding standard/H.264/Advanced Video Coding (H.264/AVC), intra prediction uses adjacent rows and adjacent column pixels of decoded blocks for intra prediction. . From these pixels, different directions of intra prediction mode or DC prediction mode can be selected. H.264/AVC includes 9 intra prediction modes based on blocks/subblocks of size (4 pixels x 4 pixels) or (8 pixels x 8 pixels), and macros of size (16 pixels x 16 pixels) Four intra prediction modes of the block.

The left part of Fig. 1 shows a sub-block of size (4 pixels X 4 pixels) and its reference pixel AM, and the right part of Figure 2 shows 8 different intra prediction modes of ap pixels in the sub-block. The mode direction is predicted, and FIG. 2 shows nine prediction modes of the intra prediction mode: vertical prediction mode, horizontal prediction mode, DC prediction mode, and Diagonal Down-left prediction. Mode, Diagonal Down-right prediction mode, Right-right prediction mode, Horizontal-down prediction mode, Vertical-left prediction mode, Upper horizontal ( Horizontal -up ) Prediction mode.

Since the prediction of the DC in the prior art (the mode in FIG. 2 is equal to 2) is to use the average values of the reference pixels A, B, C, D and I, J, K, L to predict the pixel values of the current block, only Single The reference pixels (A, B, C, D) on the left side of the current block and the reference pixels (1, J, K, L) on the upper side are averaged to obtain a pixel prediction value, which destroys the continuous characteristic of the texture distribution of the current block. , reducing the accuracy of the prediction. Summary of the invention

The embodiment of the invention provides a method for intra prediction and an intra prediction device, which can maintain the continuous characteristic of the texture distribution of the current block when performing current block pixel value prediction, and ensure the accuracy of prediction.

The embodiment of the invention provides a method for intra prediction, which includes:

Obtaining a current block in the frame for performing pixel value prediction;

Determining, according to a prediction mode direction of the neighboring block of the current block and a reference pixel value in the neighboring block, a pixel prediction value of the current block in the current block texture direction.

Correspondingly, an embodiment of the present invention further provides an intra prediction apparatus, including:

Obtaining a unit, acquiring a current block in the frame for performing pixel value prediction;

a prediction unit, determining a texture direction of the current block according to a prediction mode direction of a neighboring block of the current block and a reference pixel value in the neighboring block, and determining the current according to a texture direction of the current block The pixel prediction value of the block.

The embodiment of the present invention performs pixel value prediction by using a current block in a frame, and determines a texture direction of the current block according to a prediction mode direction of a neighboring block of the current block and a reference pixel value in the adjacent block, and Further determining a pixel prediction value of the current block according to a texture direction of the current block, thereby maintaining a continuous characteristic of a texture distribution of the current block when performing current block pixel value prediction, and ensuring accuracy of prediction. BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are incorporated in the claims Other drawings may also be obtained from these drawings without the inventive labor.

1 is a schematic diagram of a method for intra prediction according to an embodiment of the present invention;

2 is a schematic diagram of still another embodiment of a method for intra prediction of the present invention;

3 is a schematic diagram showing the positional relationship between a current sub-block and its neighboring sub-blocks in the embodiment of the present invention; FIG. 4 is a schematic diagram showing four types of distribution of a macroblock to which a current sub-block belongs in the embodiment of the present invention; FIG. 5 is a schematic diagram of a macroblock C to which a current sub-block belongs according to an embodiment of the present invention; FIG.

6 is a schematic structural diagram of an intra prediction apparatus according to an embodiment of the present invention;

FIG. 7 is a schematic structural diagram of still another embodiment of the intra prediction apparatus of the present invention. detailed description

The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

The intra prediction method in this embodiment includes:

101. Obtain a current block in a frame for performing pixel value prediction.

102. Determine, according to a prediction mode direction of a neighboring block of the current block and a reference pixel value in the neighboring block, a pixel prediction value of the current block in a current block texture direction.

A method for performing intra prediction according to an embodiment of the present invention can maintain the continuous feature of the texture distribution of the current block when performing current block pixel value prediction, and ensure prediction accuracy.

2 is a schematic diagram of still another embodiment of a method for intra prediction in the present invention;

In this method, it is assumed that the bit depth is 8 bits, and a fixed value of 2 ' ― ¹ is 128. Other values may be specifically adopted according to actual conditions, and are not limited thereto. Referring to FIG. 2, the method mainly includes:

201. Obtain a current subblock (current Subblock) in the frame for performing pixel value prediction.

Wherein the sub-block adjacent to the current block of adjacent sub positional relationship shown in FIG. 3, wherein adjacent sub-blocks may include blocks adjacent to the left of the current block (left), upper block _(U p), a combination of one or more of the upper right block (up_right) and the upper left block (up_left), and eight, B, C, D, E, F, G, H, I, J, K, L, X are adjacent The reference pixel adjacent to the current sub-block (having a corresponding reference pixel value), for convenience of description, the reference pixel value corresponding to each reference pixel is also used A, B, C, D, E, F, G, H , I, J, K, L, X indicate that the intra prediction mode of each adjacent sub-block is also represented by using left, up, up_right, upjeft, and the SO mentioned below is the average of I, J, K, and L. , SI is the average value of A, B, C, D, 82 is the average of 5, F, G, H, and S3 is X;

202, in combination with the distribution of the current sub-block in a frame, when a combination of one or more of the left block, the upper block, the upper right block, and the upper left block exists, according to the neighboring sub-blocks Obtaining a reconstructed pixel value of each adjacent sub-block by using a reference pixel value adjacent to the current sub-block, and specifically referring to the following process: In combination with the four distributions of the macroblock to which the current sub-block belongs as shown in FIG. 2, where C represents the macroblock to which the current sub-block belongs, A represents the adjacent macroblock encoded above C, and L represents the left of C. Coding adjacent macroblocks, and the current sub-block we mentioned may be as shown in the position 0 of the macroblock C to which the current sub-block belongs as shown in FIG. 4, and the macroblock C adopts the specification of (16 pixels x 16 pixels). The current sub-block adopts the specification of (4 pixels and 4 pixels);

In the first distribution case, the macroblock C is distributed as shown in (a) of FIG. 4, and the macroblock C is located at the first upper left corner of the frame, according to the definition of the adjacent sub-block of the current sub-block in the macroblock C, In this case, there is no relevant reference pixel around the current sub-block, so it can be determined that the pixel prediction value of the current sub-block is 128, at this time:

When the left block, the upper block, the upper right block, and the upper left block of the adjacent sub-blocks are not present, it is determined that the pixel prediction value of the current block is a fixed value 128, and the situation may be expressed as follows:

S0=128, Sl=128, S2=128, S3=128, and the intra prediction mode value ipmode_x=-l,

Ipmode_x takes a value of -1 to indicate that there is no corresponding left block, upper block, upper right block, and upper left block, and ipmode_x may take any one of 10 integers between -1 and 8 in different cases, where ipmode_x takes 0 to The values of the 9 integers between 8 represent 9 intra prediction modes; the second distribution case, the distribution of macroblocks is as shown in (b) of Fig. 4, and the macroblock C is located in the first line of the frame, according to the above pair of macros The definition of the adjacent sub-block of the current sub-block in block C. In this distribution case, the current sub-block (in the case of sub-blocks 0, 1, 4, 5 as shown in FIG. 5) has four left sides. Reference pixel (1, J, K, L), at this time:

When the left block of the adjacent sub-block exists, and the upper block, the upper right block, and the upper left block do not exist, the following manner is expressed:

S0=mean(U,K,L), Sl=128, S2=mean(S0,Sl), S3=128; ipmode_left=M0,

Where mean means averaging, for example S0=mean(U,K,L) means SO is the average of I, J, K, L, and M0 means the frame of the left block when the left block of the adjacent sub-block exists The intra prediction mode value is M0, where the value of S2 is artificially set/calculated due to the absence of the upper right block, that is, S2=mean(S0, S1) is an average value of SO and S1 based on the empirical value S2, of course. In the actual case, S2 can also select an appropriate value in the range of 0 to 255 pixels; in the third distribution case, the macroblock is distributed as shown in (c) of FIG. 4, and the macroblock C is located in the first column of the frame, according to the above The definition of the adjacent sub-block of the current sub-block in the macro block C. In this distribution case, the current sub-block (as in the case of sub-blocks 0, 2, 8, and 10 shown in FIG. 5) has an upper side and an upper right side. 8 directions Reference pixels (A, B, C, D, E, F, G, H), at this time:

When the upper block and the upper right block of the adjacent sub-block exist, and the left block and the upper left block do not exist, the following manner is expressed:

S0=128, Sl=mean(A,B,C,D), S2=mean(E,F,G,H), S3=128; ipmode_up=Ml, ipmode_up_right=M2, ipmode_left=-l, ipmode_up_left=- l , where M1 indicates that when the upper block of the adjacent sub-block exists, the intra prediction mode value of the upper block is M1; M2 indicates that when the upper right block of the adjacent sub-block exists, the intra prediction mode value of the upper right block is M2 ;

In the fourth distribution case, the macroblock is distributed as shown in (d) of FIG. 4, and the macroblock C is in the middle of the frame. According to the above definition of the adjacent sub-block of the current sub-block in the macroblock C, in the case of such a distribution , the current sub-block has a complete reference phase point around it, at this time:

When the left block, the upper block, the upper right block, and the upper left block of the adjacent sub-blocks all exist, the following representation is performed:

S0=mean(U,K,L) , Sl=mean(A,B,C,D) , S2=mean(E,F,G,H) , S3=X ; ipmode_left=M0 ,

, wherein the intra prediction mode value of the upper left block is Μ3;

In addition, the fifth distribution case is that the macroblock C is in the last column of the frame. According to the definition of the adjacent sub-block of the current sub-block in the macroblock C, in the case of the distribution, only the upper right block of the current sub-block does not exist. , then the values of E, F, G, and H can be manually set/calculated as E=mean(A,I), F=mean(B,J), G=mean(C,K), according to experience. H=mean(D,L), and the intra prediction mode value of the upper right block can be set to M3, at this time:

When the left block, the upper block, the upper right block, and the upper left block of the adjacent sub-blocks all exist (where the upper right block information is artificially set), the following representation is as follows:

S0=mean(U,K,L) , Sl=mean(A,B,C,D) , S2=mean(E,F,G,H) , S3=X ; ipmode_left=M0 ,

, where the values of E, F, G, and H show that the value of S2 is actually the average value of SO and SI;

This fifth distribution can be attributed to the fourth distribution described above.

It is to be noted that, according to the reference pixel values adjacent to the current sub-block in each adjacent sub-block, the reconstructed pixel value of each adjacent sub-block is obtained, which may be a corresponding reference. The pixel value is used as the reconstructed pixel value of each adjacent sub-block, and is not necessarily the reference pixel mean value (obtained by the mean step);

203. Determine a texture direction of the current sub-block according to a prediction mode direction of each adjacent sub-block, and further determine the current sub-block according to a relationship between the reconstructed pixel values of the adjacent sub-blocks. The texture direction may be specifically combined with the four distributions of the macroblock to which the current sub-block belongs as shown in FIG. 4, and the following process is performed:

For the first distribution case described above, since there is no relevant reference pixel around the current sub-block, the pixel prediction value of the current sub-block is 128;

For the second distribution case described above, since only the left block of the adjacent sub-block exists, the prediction mode direction (ipmode) of the left block and the reconstructed pixel value of each adjacent sub-block may be used as reference information, according to different The prediction mode direction and the relationship between SO, SI, S2 determine the texture direction of the current sub-block, and the approximate texture direction of the current sub-block can be determined according to the intra prediction mode value M0 of the left block, and then according to S0, SI Between the size or the relationship between the absolute difference between SO and S1 and a preset threshold, a more accurate texture direction is determined from the approximate texture direction;

For the third distribution case described above, due to the existence of the upper block and the upper right block of the adjacent sub-block, the prediction mode direction (ipmode) of the upper block and the upper right block and the reconstructed pixel value of each adjacent sub-block may be used. As the reference information, the texture direction of the current sub-block is determined according to the relationship between the different prediction mode directions and SO, SI, S2, and the approximate sub-block is determined according to the intra prediction mode values M1 and M2 of the upper block and the upper right block. The texture direction, and then according to the size between Sl, S2 or according to the relationship between the absolute difference between SI and S2 and the preset threshold, a more accurate texture direction is determined from the approximate texture direction;

For the fourth distribution case mentioned above (including the fifth distribution case), since there are left block, upper block, upper right block, and upper left block of adjacent sub-blocks, the left block, the upper block, the upper right block, and the upper block may be The prediction mode direction (ipmode) of the upper left block and the reconstructed pixel value of each adjacent sub-block are used as reference information, and the texture direction of the current sub-block is determined according to the relationship between different prediction mode directions and SO, SI, S2, and S3. Determining the approximate texture direction of the current sub-block according to the intra prediction mode values M0, M1, M2, M3 of the left block, the upper block, the upper right block, and the upper left block, and then according to SO and S1, SO and S2, SO and S3, Size relationship between SI and S2, SI and S3 or S2 and S3, or according to SO and Sl, SO and S2, SO and S3, SI and S2, SI and S3 or S2 and S3 absolute difference and preset threshold a relationship that determines a more accurate texture direction from the approximate texture direction;

In addition, according to the correlation between the adjacent sub-block and the current sub-block, the following rules may be considered in the selection of the texture direction: the prediction mode direction of the left block is optimal, followed by the prediction mode direction of the upper block, and the third The prediction mode direction of the upper right block, and finally the prediction mode direction of the upper left block;

204. Determine, according to the further determined texture direction, a pixel prediction value of the current sub-block by the reconstructed pixel value of each adjacent sub-block, that is, pre-predicting a pixel of the current sub-block according to the predicted texture distribution. The measured value is equal to one of the reconstructed pixel values of the adjacent sub-blocks or a combination thereof, and the content described in the above 303 is specifically as follows, so that the pixel prediction value of the current sub-block is equal to the respective One of the reconstructed pixel values of adjacent sub-blocks is an example but is not limited to this:

For the above second distribution case and considering the corresponding content in 203, since only the left block of the adjacent sub-block exists, when M0 is horizontal, and SO is smaller than S1, or absolute difference abs(s0, sl) is smaller than some a set threshold, the pixel prediction value of the current sub-block is SO;

For the above third distribution case and considering the corresponding content in 203, since there is the existence of the upper block and the upper right block of the adjacent sub-block, when M1 is equal to the lower left diagonal direction, M2 is equal to the upper horizontal direction, abs(S1, S2) is greater than a certain threshold, then the pixel prediction value of the current sub-block is S2;

For the fourth distribution case mentioned above (including the fifth distribution case) and considering the corresponding content in 203, since there are left block, upper block, upper right block, and upper left block of adjacent sub-blocks, when M0, Ml, M2 exist And M3 is a DC prediction value. If abs(S0, S3) is less than a certain threshold and abs(S1, S2) is less than the threshold, the pixel prediction value of the current sub-block is S3.

Performing, by performing the intra prediction method of the embodiment of the present invention as shown in FIG. 2, performing pixel value prediction on a current sub-block in the acquisition frame, when one of the left block, the upper block, the upper right block, and the upper left block is used. Or a plurality of combinations, according to the reference pixel values adjacent to the current block in each neighboring block, obtaining reconstructed pixel values of the neighboring blocks, and according to the adjacent sub-blocks The prediction mode direction determines a texture direction of the current sub-block, and further determines a texture direction of the current sub-block according to a relationship between the reconstructed pixel values of the adjacent sub-blocks, and finally, according to the further Determining the texture direction, determining a pixel prediction value of the current sub-block by the reconstructed pixel value of each adjacent sub-block, so that the reference pixel around the current sub-block and the prediction mode information of the adjacent sub-block may be fully utilized, The texture distribution characteristics of the current sub-block are maintained, and the accuracy of the current sub-block pixel prediction is improved without increasing the complexity, and the coding efficiency is effectively improved.

As an implementation manner, when there is no prediction mode direction of the neighboring block, for example, when the prediction mode of the neighboring block is all DC prediction mode or the neighboring block is an inter block, the foregoing 203, 204 Can be replaced by the following steps:

Determining a texture direction of the current block according to a relationship between reconstructed pixel values of the neighboring blocks; determining, according to the determined texture direction, the current block by reconstructed pixel values of the neighboring blocks a pixel prediction value, specifically, determining, by the reconstructed pixel value of each neighboring block, a pixel prediction value of the current block may be such that a pixel prediction value of the current sub-block is equal to a reconstruction of the neighboring sub-blocks In the pixel value The mean of one or a combination thereof.

As an implementation manner, when the encoding end obtains the true pixel value of the current sub-block (instead of the predicted pixel value), one of the reconstructed pixel values of the plurality of adjacent sub-blocks obtained by the above method or The predicted pixel value closest to the real pixel value is selected from the combined mean values. At this time, a corresponding identifier needs to be added at the encoding end, for example, a DC_sd_Idx identifier, and the following macroblock prediction pixel value conversion syntax modification is shown in Table 1. Show:

Table 1

If the pixel prediction value of the current sub-block is a pixel prediction value corresponding to the left block,

DC_sel_Idx=0; if the pixel prediction value of the current sub-block is a pixel prediction value corresponding to the upper block, DC_sel_Idx=l; if the pixel prediction value of the current sub-block is a pixel prediction value corresponding to the upper right block, DC_sel_Idx= 2; if the pixel prediction value of the current sub-block is a pixel prediction value corresponding to the upper left block, DC_sel_Idx=3 and so on.

As an implementation manner, when the encoding end obtains the pixel prediction value of the current sub-block by the above method, and When the current pixel value of the current sub-block is obtained, the difference between the pixel prediction value of the current sub-block and the real pixel value may be further obtained, indicated and transmitted by the corresponding identifier in the code table, and the current The pixel prediction value of the sub-block, the decoded end can be restored to the real pixel value by the received difference value and the pixel prediction value.

6 is a schematic structural diagram of an intra prediction apparatus according to an embodiment of the present invention. Referring to the figure, the apparatus mainly includes:

The obtaining unit 61 is configured to acquire a current block in the frame for performing pixel value prediction;

The prediction unit 62 determines a pixel prediction value of the current block in the current block texture direction according to a prediction mode direction of a neighboring block of the current block and a reference pixel value in the neighboring block.

The intra prediction apparatus provided by the embodiment of the present invention can maintain the continuous characteristic of the texture distribution of the current block when performing the current block pixel value prediction, and ensure the accuracy of the prediction.

FIG. 7 is a schematic structural diagram of still another embodiment of an intra prediction apparatus according to the present invention.

In this embodiment, the device includes:

The obtaining unit 71, the prediction unit 72, wherein the prediction unit 72 includes a pixel prediction value obtaining unit 721, a texture direction determining unit 722, and a determining unit 723, the determining unit 723 further includes a first determining unit 7231 and a second determining unit 7232, wherein The function of each unit is as follows:

The obtaining unit 71 is configured to acquire a current block in the frame for performing pixel value prediction;

a pixel prediction value obtaining unit 721, in combination with the distribution of the current block in a frame, when a neighboring block adjacent to the current block: a combination of one or more of a left block, an upper block, an upper right block, and an upper left block When present, the reconstructed pixel values of the neighboring blocks are obtained according to reference pixel values adjacent to the current block in the neighboring blocks, where the positional relationship between the neighboring block and the current block can be referred to FIG. As shown, the processing of the various distribution cases is mentioned in the above corresponding content, and details are not described herein again;

The texture direction determining unit 722 determines the texture direction of the current block according to the prediction mode direction of each neighboring block, and further determines the current according to the relationship between the reconstructed pixel values of the adjacent blocks. Specifically, the texture direction of the block may be determined according to the prediction mode direction of each neighboring block, and then according to the relationship between the reconstructed pixel values of the adjacent blocks. Determining a more accurate texture direction from the approximate texture direction, wherein a relationship between reconstructed pixel values of each adjacent block is an absolute difference of pixel prediction values corresponding to at least one pair of adjacent blocks a relationship between a value and a preset threshold, or a magnitude relationship of pixel prediction values corresponding to at least one pair of adjacent blocks; a first determining unit 7231, according to the further determined texture direction, by the neighboring blocks Reconstructing a pixel value to determine a pixel prediction value of the current block;

The second determining unit 7232 determines that the pixel prediction value of the current block is a fixed value when the left block, the upper block, the upper right block, and the upper left block are not present, and the fixed value may be an experience in the prior art. The value of 128, the second determining unit 7232 can be selected according to the actual situation.

As an implementation manner, when there is no prediction mode direction of the neighboring block, for example, the prediction mode of the neighboring block is all DC prediction mode or the neighboring block is an inter-subblock/block/macroblock The texture direction determining unit 722 is further configured to determine a texture direction of the current block according to a relationship between the reconstructed pixel values of the neighboring blocks, where the determining unit 723 is further configured to be used according to the determining a texture direction, the pixel prediction value of the current block is determined by reconstructed pixel values of the neighboring blocks.

The following points need to be explained:

1. The setting of the reference pixel of the adjacent block may be selected according to actual conditions, and is not limited to adopting a pixel point adjacent to the current block in the adjacent block;

2. The current block may be a specification of (4 pixels X 4 pixels), (8 pixels χ 8 pixels), or (16 pixels X 16 pixels).

In addition, one of ordinary skill in the art can understand that all or part of the process in implementing the foregoing embodiments may be completed by a program instructing related hardware, and the program may be stored in a computer readable storage medium. The program, when executed, may include the flow of an embodiment of the methods as described above. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

The above is a specific embodiment of the present invention. It should be noted that those skilled in the art can also make several improvements and retouchings without departing from the principles of the present invention. These improvements and retouchings are also considered. It is the scope of protection of the present invention.

Claims

Rights request

A method for intra prediction, comprising:

Obtaining a current block in the frame for performing pixel value prediction;

2. The method of intra prediction according to claim 1, wherein the neighboring block comprises one or more of a left block, an upper block, an upper right block, and an upper left block adjacent to the current block. And determining, according to a prediction mode direction of a neighboring block of the current block and a reference pixel value in the neighboring block, a texture side of the current block, determining the The pixel prediction value of the current block is specifically:

When a combination of one or more of the left block, the upper block, the upper right block, and the upper left block exists, obtaining the respective reference values according to reference pixel values adjacent to the current block in each adjacent block Reconstructed pixel values of adjacent blocks;

Determining a texture direction of the current block according to a prediction mode direction of each neighboring block, and further determining a texture direction of the current block according to a relationship between the reconstructed pixel values of the neighboring blocks; Determining the determined texture direction, determining a pixel prediction value of the current block from reconstructed pixel values of the neighboring blocks.

3. The method of intra prediction according to claim 1, wherein the neighboring block comprises one or more of a left block, an upper block, an upper right block, and an upper left block adjacent to the current block. Combination, when there is no prediction mode direction of the neighboring block, determining the current block texture according to a prediction mode direction of a neighboring block of the current block and a reference pixel value in the neighboring block The pixel prediction value of the current block in the direction is specifically:

Determining a texture direction of the current block according to a relationship between reconstructed pixel values of the neighboring blocks; determining, according to the determined texture direction, the current block by reconstructed pixel values of the neighboring blocks Pixel prediction

The method of intra prediction according to claim 2 or 3, wherein the method further comprises: determining, when the left block, the upper block, the upper right block, and the upper left block are not present, determining the current block The pixel prediction value is a fixed value.

The method of intra prediction according to claim 2 or 3, wherein the obtaining the neighboring blocks according to reference pixel values adjacent to the current block in each neighboring block The reconstructed pixel value is specifically:

When the left block exists, the upper block, the upper right block, and the upper left block do not exist, an average value of reference pixel values adjacent to the current block in the left block is calculated, and the average value is used as the left a reconstructed pixel value of the block; a fixed value as a reconstructed pixel value of the upper block; an average value of the reconstructed pixel value of the left block and a reconstructed pixel value of the upper block as the upper right a reconstructed pixel value of the block; using the fixed value as a reconstructed pixel value of the upper left block;

When the upper block and the upper right block exist, and the left block and the lower left block do not exist, the fixed value is used as a reconstructed pixel value of the left block; and the upper block is adjacent to the current block. An average value of reference pixel values, using the average value as a reconstructed pixel value of the upper block; calculating an average value of reference pixel values adjacent to the current block in the upper right block, using the average value as a Reconstructing pixel values corresponding to the upper right block; using the fixed value as the reconstructed pixel value of the upper left block;

When the left block, the upper block, the upper right block, and the upper left block are both present, an average value of reference pixel values adjacent to the current block in the left block is calculated, and the average value is used as the weight of the left block Constructing a pixel value; calculating an average value of reference pixel values adjacent to the current block in the upper block, using the average value as a reconstructed pixel value of the upper block; calculating the current upper right block and the current An average value of the reference pixel values adjacent to the block, the average value being used as the reconstructed pixel value of the upper right block; and a reference pixel value adjacent to the current block in the upper left block as the upper left block Reconstructed pixel value;

When the left block, the upper block, and the upper left block exist, and the upper right block does not exist, an average value of reference pixel values adjacent to the current block in the left block is calculated, and the average value is used as the left Reconstructing a pixel value of the block; calculating an average value of the reference pixel values adjacent to the current block in the upper block, using the average value as a reconstructed pixel value of the upper block; Constructing an average value of the pixel value and the reconstructed pixel value of the upper block as a reconstructed pixel value of the upper right block; in the upper left block and the current block A reference pixel value of the neighbor is used as the reconstructed pixel value of the upper left block.

The method of intra prediction according to claim 2, wherein the determining, according to the relationship between the reconstructed pixel values of the adjacent blocks, the texture direction of the current block is specifically:

And determining, according to a prediction mode direction of each neighboring block, a relationship between an absolute difference of reconstructed pixel values of the at least one pair of adjacent blocks and a preset threshold. Further determining the texture direction of the current block, or

Determining the texture of the current block according to the magnitude relationship of the reconstructed pixel values of the at least one pair of neighboring blocks on the basis of determining the texture direction of the current block according to the prediction mode direction of each neighboring block. direction.

7. The intra prediction method according to claim 4, wherein the fixed value is 2, wherein bitdepth represents a bit depth.

The method of intra prediction according to any one of claims 1 to 3, wherein the current block is (4 pixels X 4 pixels), (8 pixels X 8 pixels) or (16 pixels X 16 pixels) specification.

9. An intra prediction apparatus, comprising:

And a prediction unit, determining, according to a prediction mode direction of the neighboring block of the current block and a reference pixel value in the neighboring block, a pixel prediction value of the current block in the current block texture direction.

10. The intra prediction apparatus according to claim 9, wherein the neighboring block includes one or more of a left block, an upper block, an upper right block, and an upper left block adjacent to the current block. The combination of the prediction unit includes:

a first pixel prediction value obtaining unit, when a combination of one or more of the left block, the upper block, the upper right block, and the upper left block exists, according to the neighboring block adjacent to the current block Referring to the pixel values, obtaining reconstructed pixel values of the neighboring blocks;

a first texture direction determining unit, determining a texture direction of the current block according to a prediction mode direction of each adjacent block, and further determining a relationship according to a relationship between reconstructed pixel values of the adjacent blocks Describe the texture direction of the current block;

The first determining unit determines a pixel prediction value of the current block from reconstructed pixel values of the neighboring blocks according to the further determined texture direction.

11. The intra prediction apparatus according to claim 9, wherein the neighboring block includes one or more of a left block, an upper block, an upper right block, and an upper left block adjacent to the current block. The combination unit, when there is no prediction mode direction of the neighboring block, the prediction unit includes:

a second pixel prediction value obtaining unit, when a combination of one or more of the left block, the upper block, the upper right block, and the upper left block exists, according to the neighboring block adjacent to the current block Referring to the pixel values, obtaining reconstructed pixel values of the neighboring blocks;

a second texture direction determining unit, determining a texture direction of the current block according to a relationship between reconstructed pixel values of the neighboring blocks;

The second determining unit determines, according to the determined texture direction, a pixel prediction value of the current block by the reconstructed pixel value of each neighboring block.

The intra prediction apparatus according to claim 10 or 11, wherein the determining unit comprises:

a first determining unit, determining, according to the further determined texture direction, a pixel prediction value of the current block by a reconstructed pixel value of each neighboring block;

And a second determining unit, when none of the left block, the upper block, the upper right block, and the upper left block is present, determining that the pixel prediction value of the current block is a fixed value.

13. The intra prediction apparatus according to claim 12, wherein the fixed value is 2, wherein bitdepth represents a bit depth.

The intra prediction apparatus according to any one of claims 9 to 11, wherein the current block is (4 pixels X 4 pixels), (8 pixels X 8 pixels) or (16 pixels X 16 Pixel) specifications.