WO2012097675A1

WO2012097675A1 - Encoding/decoding method and codec based on intra-prediction mode of image block

Info

Publication number: WO2012097675A1
Application number: PCT/CN2011/085195
Authority: WO
Inventors: 曹潇然; 赖昌材; 何芸
Original assignee: 华为技术有限公司; 清华大学
Priority date: 2011-01-18
Filing date: 2011-12-31
Publication date: 2012-07-26
Also published as: CN102724493A; CN102724493B

Abstract

Disclosed are an encoding/decoding method and a codec based on an intra-prediction mode of an image block. The method includes: obtaining a prediction mode of a first sub-block in a current image block and a prediction mode of a second sub-block in the current image block, wherein the first sub-block and the second sub-block are adjacent to each other and are arranged in the horizontal direction or the vertical direction; taking the prediction mode of the first sub-block as the most probable prediction mode of the second sub-block; and determining an identifier for indicating the prediction mode of the second sub-block, and writing the same into a code stream, according to the prediction mode of the second sub-block and the most probable prediction mode of the second sub-block. By adopting the technical solution, the most probable prediction mode of the image block can be effectively predicted, so that the overhead of the code stream can be saved.

Description

Encoding and decoding method and codec based on image block-based intra prediction mode The present application claims to be filed on January 18, 2011, the Chinese Patent Office, application number 201110020668.9, and the invention name is "image block-based intra prediction mode. The priority of the Chinese Patent Application for the Decoding Method and the Codec, the entire contents of which are incorporated herein by reference.

Technical field

The present invention relates to the field of image processing technologies, and in particular, to a codec method and a codec based on an intra block prediction mode of an image block. Background technique

In video coding compression processing, the basic principle of an intra coding is to use spatial correlation, such as using intra prediction, to remove spatial redundancy. Intra coding does not require a reference frame and does not involve interpolating the reference frame. When performing intra-frame coding, the prediction in the division mode may be performed according to a certain division manner, thereby obtaining a residual block corresponding to the division size, and transforming, quantizing, and entropy coding the residual block. At the time of decoding, the division mode of the current coding object and the corresponding division size are first obtained, and then the coefficients in the code stream are decoded, and the prediction and reconstruction under the corresponding block division mode are performed to generate a decoded image. When performing the prediction of the upper image block, it is necessary to first obtain the prediction direction of the current block (also referred to as a prediction mode), and then obtain reference pixels from the pixels of the adjacent blocks according to the obtained prediction direction, as in the prediction. The predicted pixel value of the current block in the direction.

Figure 1 is a schematic diagram of sub-block partitioning in a 16x16 image block. A total of 16 4x4 blocks are included in the current 16x16 block, and each of the 16 4x4 blocks is a sub-block of the current 16x16 block. It can be understood that a block and a sub-block are a relative concept. As a single block, a sub-block may also include other sub-blocks therein; the sub-blocks within a block may also be flexible, for example, sub-blocks may be It is a horizontal rectangle or a vertical rectangle, not just a square.

2 is a schematic diagram of an intra prediction direction of an image block, in which an 8x8 block is identified by a thick line, and the arrow in the figure points to a possible prediction direction of one of the 8x8 blocks. Generally speaking, When intra prediction is performed on the current sub-block, an optimal prediction direction is determined from a plurality of possible prediction directions as a prediction mode of the current sub-block, and the current sub-block is predictively coded according to the prediction mode, and The prediction mode is written in the code stream, and the residual is encoded.

In order to reduce the bits occupied by the prediction mode in the code stream, the prior art introduces a most probable mode (MPM), also called the most probable prediction direction. The actual prediction mode of the sub-block is not directly written in the code stream, but the most probable prediction mode of the current sub-block is obtained first, and then the actual prediction mode of the sub-block is compared with the most probable prediction mode of the sub-block. And obtaining identification information that needs to be written into the code stream according to the obtained size relationship; if the actual prediction mode of the sub-block is the same as the most probable prediction mode of the sub-block, the identifier can be saved in the code stream. Occupied bits. Therefore, how to determine the most probable prediction mode of an image block as close as possible to the actual prediction mode of the image block becomes a problem. In the conventional block-based coding structure, when predicting the most probable prediction mode of the current block, the actual prediction mode of the neighboring block with the smallest prediction mode value in the upper neighboring block and the left neighboring block of the current block is used as The most probable prediction mode of the current block, so that the correlation between the current block and its left neighboring block and the upper neighboring block can be utilized to improve the accuracy of predicting the most probable prediction mode of the current block. However, for a block that uses irregular division, a sub-block of one block may not be a block, but a rectangular block or the like, so the left adjacent block and the upper adjacent block of the current block may have different divisions, if still in prediction mode. The actual prediction mode of the neighboring block with the smallest value is the most probable prediction mode of the current block, and the most probable prediction mode obtained is not necessarily accurate. Therefore, how to effectively predict the most probable prediction mode of a block and make this prediction method applicable not only to block division but also to irregular block division becomes a problem.

Summary of the invention

Embodiments of the present invention provide a codec method and a codec based on an intra prediction mode of an image block to effectively predict a most probable prediction mode of an image block. According to an embodiment of the present invention, an encoding method based on an intra block prediction mode of an image block is provided, including:

Obtaining a prediction mode of the first sub-block in the current image block and a prediction mode of the second sub-block in the current image block, the first sub-block being adjacent to the second sub-block and in a horizontal direction or a vertical direction arrangement; Determining, by the prediction mode of the first sub-block, a most probable prediction mode of the second sub-block; determining, according to the prediction mode of the second sub-block and the most probable prediction mode of the second sub-block, for indicating the second sub-block The identifier of the prediction mode is written and written into the code stream.

According to another embodiment of the present invention, a decoding method based on an intra block prediction mode of an image block is provided, including:

Obtaining an identifier for indicating a prediction mode of the second sub-block in the current image block from the code stream; using a prediction mode of the first sub-block in the current image block as a most probable prediction mode of the second sub-block, The first sub-block is adjacent to the second sub-block and arranged in a horizontal direction or a vertical direction;

A prediction mode of the second sub-block is determined based on the identification and a most probable prediction mode of the second sub-block.

According to another embodiment of the present invention, an encoder is provided, including:

An acquiring unit, configured to acquire a prediction mode of a first sub-block in a current image block and a prediction mode of a second sub-block in the current image block, where the first sub-block is adjacent to the second sub-block and is horizontal Arranged in the direction or in the vertical direction;

a determining unit, configured to use a prediction mode of the first sub-block as a most probable prediction mode of the second sub-block;

a code stream writing unit, configured to determine, according to a prediction mode of the second sub-block and a most probable prediction mode of the second sub-block, an identifier for indicating a prediction mode of the second sub-block, and write the identifier into the code stream in.

According to another embodiment of the present invention, a decoder is provided, including:

a code stream reading unit, configured to acquire, from the code stream, an identifier for indicating a prediction mode of the second sub-block in the current image block;

a determining unit, configured to use a prediction mode of the first sub-block in the current image block as a most probable prediction mode of the second sub-block, where the first sub-block is adjacent to the second sub-block and is horizontal or vertical Straight in the direction;

a decoding unit, configured to determine, according to the identifier and a most probable prediction mode of the second sub-block The prediction mode of the two sub-blocks.

By adopting the above technical solution, the embodiment of the present invention can better utilize the correlation between different neighboring blocks arranged in a certain direction to predict the most probable prediction direction of a block, so that the most probable prediction direction of the block is close. The actual prediction direction is beneficial to save the number of bits occupied by the identifier for indicating the actual prediction direction of the block, thereby saving code stream overhead.

According to another embodiment of the present invention, an encoding method based on an intra block prediction mode of an image block is provided, including:

Obtaining a prediction mode of a first sub-block in a current image block, the first sub-block including a plurality of sub-blocks arranged in a horizontal direction or a vertical direction;

Determining a second sub-block, the prediction mode of the second sub-block as a most probable prediction mode of the first sub-block, the second sub-block being adjacent to the first sub-block and including horizontally or vertically Arranging multiple sub-blocks;

And determining, according to the prediction mode of the first sub-block and the most probable prediction mode of the first sub-block, an identifier for indicating a prediction mode of the first sub-block, and writing the identifier into the code stream.

Obtaining, from the code stream, an identifier indicating a prediction mode of the first sub-block in the current image block, the first sub-block including a plurality of sub-blocks arranged in a horizontal direction or a vertical direction;

Determining a prediction mode of the first sub-block based on the identifier and a most probable prediction mode of the first sub-block.

An obtaining unit, configured to acquire a prediction mode of a first sub-block in a current image block, where the first sub-block a plurality of sub-blocks arranged in a horizontal direction or a vertical direction;

a determining unit, configured to determine a second sub-block, the prediction mode of the second sub-block as a most probable prediction mode of the first sub-block, the second sub-block being adjacent to the first sub-block and including horizontally a plurality of sub-blocks arranged in a direction or a vertical direction;

a code stream writing unit, configured to determine, according to a prediction mode of the first sub-block and a most probable prediction mode of the first sub-block, an identifier for indicating a prediction mode of the first sub-block, and write the identifier into the code stream in.

a code stream reading unit, configured to acquire, from the code stream, an identifier for indicating a prediction mode of the first sub-block in the current image block, where the first sub-block includes a plurality of sub-blocks arranged in a horizontal direction or a vertical direction; a determining unit, configured to determine a second sub-block, the prediction mode of the second sub-block as a most probable prediction mode of the first sub-block, the second sub-block being adjacent to the first sub-block and including horizontally a plurality of sub-blocks arranged in a direction or a vertical direction;

And a decoding unit, configured to determine a prediction mode of the first sub-block according to the identifier and a most probable prediction mode of the first sub-block. The embodiment of the present invention is used to determine the most probable prediction mode of two adjacent sub-blocks with similar sub-block division manner in an image block, so that the actual prediction mode of one sub-block is the most of all sub-blocks in another sub-block. The possible prediction mode can effectively predict the most probable prediction mode, which is beneficial to save the number of bits occupied by the identifier for indicating the actual prediction direction of the block, thereby saving code stream overhead. DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings to be used in the embodiments will be briefly described below. It is obvious that the drawings in the following description are only some embodiments of the present invention, which are common in the art. For the skilled person, other drawings can be obtained from these drawings without any creative work.

1 is a schematic diagram of sub-block partitioning in a 16x16 image block;

2 is a schematic diagram of an intra prediction direction of an image block; FIG. 3 is a schematic diagram of an encoding method of an intra block prediction mode based on an image block according to an embodiment of the present invention; FIG.

FIG. 3b is a schematic diagram of a decoding method of an intra block prediction mode based on an image block according to an embodiment of the present invention; FIG.

FIG. 3c is a schematic flow chart of a method for encoding and decoding an image block prediction direction according to an embodiment of the present invention;

4 is a schematic diagram of an image block division structure according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of another image block division structure according to an embodiment of the present disclosure;

FIG. 6 is a schematic diagram of another encoding method of an intra block based intra prediction mode according to an embodiment of the present invention; FIG.

FIG. 7 is a schematic diagram of an 8×8 intra-subblock partitioning according to an embodiment of the present invention; FIG. 8 is a schematic diagram of another image block-based intra prediction mode decoding method according to an embodiment of the present invention;

FIG. 9 is a schematic structural diagram of an encoder according to an embodiment of the present disclosure;

FIG. 10 is a schematic structural diagram of a decoder according to an embodiment of the present disclosure;

FIG. 11 is a schematic structural diagram of another encoder according to an embodiment of the present disclosure;

FIG. 12 is a schematic structural diagram of another decoder according to an embodiment of the present invention.

detailed description

BRIEF DESCRIPTION OF THE DRAWINGS The technical solutions in the embodiments of the present invention will be described in detail below with reference to the accompanying drawings. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without the creative work are all within the scope of the present invention.

FIG. 3a is a schematic diagram of an encoding method of an image block-based intra prediction mode according to an embodiment of the present invention, where the method includes:

S3 la: acquiring a prediction mode of the first sub-block in the current image block and a second sub-module in the current image block a prediction mode of the block, the first sub-block being adjacent to the second sub-block and arranged in a horizontal direction or a vertical direction.

S32a: The prediction mode of the first sub-block is taken as the most probable prediction mode of the second sub-block.

S33a: Determine an identifier for indicating a prediction mode of the second sub-block according to a prediction mode of the second sub-block and a most probable prediction mode of the second sub-block, and write the identifier into the code stream.

In this embodiment and subsequent embodiments, the prediction mode is also referred to as the prediction direction, and the two expressions are different but represent the same meaning. The method provided by the above embodiments can effectively predict the most probable prediction direction of an image block, thereby determining the value of the identifier of the write code stream. This method is not only suitable for direction prediction of block division, but also for direction prediction of non-blocks such as horizontal rectangles or vertical rectangle blocks. In order to predict that the most probable prediction direction of the current block is different by selecting one block as the reference block in the left neighboring block and the upper neighboring block, the present embodiment only uses one neighboring block of the current block to predict the most probable prediction direction of the current block. The adjacent block for predicting the most probable prediction direction of the current block is arranged horizontally or vertically with the current block, and the correlation between different adjacent blocks arranged in a certain direction can be better utilized, so that the predicted result is obtained. The most probable prediction direction of the current block is close to its actual prediction direction, and can be applied to the scene in which the sub-block division manner is irregular in the image block, which is advantageous for saving the number of bits for indicating the actual prediction direction of the current block, when the coding end When the coding information is transmitted by using the method provided in this embodiment, the code stream overhead can be saved as much as possible, and the overall coding performance is improved.

FIG. 3b is a schematic diagram of a decoding method of an intra block based intra prediction mode according to an embodiment of the present invention, where the method includes:

S31b: Obtain an identifier for indicating a prediction mode of the second sub-block in the current image block from the code stream;

S32b: The prediction mode of the first sub-block in the current image block is regarded as the most probable prediction mode of the second sub-block, where the first sub-block is adjacent to the second sub-block and is horizontal or vertical Arrange in direction

S33b: Determine a prediction mode of the second sub-block according to the identifier and a most probable prediction mode of the second sub-block.

The method provided in this embodiment can be applied to the decoding end, and can effectively predict the most probable prediction of the image block. Direction, in order to obtain an actual prediction direction of the image block, and predictively decode the image block according to the prediction direction, and accurately obtain the decoded image.

For ease of understanding, the present invention also provides an embodiment of a method for encoding and decoding an image block prediction direction, as shown in Figure 3c, the method includes:

S31c: Dividing the current image block to obtain a plurality of sub-image blocks, and writing the division mode information of the current image block into the code stream. Wherein the plurality of sub-image blocks are sequentially arranged in a horizontal or vertical direction, and the plurality of sub-image blocks include at least a first sub-image block at a boundary of the current image block, and the first sub-image block Adjacent second sub-image block.

A schematic diagram of a specific image block partitioning structure provided by an embodiment of the present invention is shown in FIG. 4. For a 32x32 block, four 16x16 blocks can be divided, and each 16x16 block can be divided into four 8x8 blocks. The 16x16 block in the lower left corner of the block is divided into four 4x16 blocks, and the same 8x8 block can be further divided into 4x4 blocks or 2x8 blocks, etc., and those skilled in the art can perform various kinds of image blocks according to actual needs. Flexible division, to get a variety of blocks of different sizes and shapes, these blocks can be square or other shape of the rectangle, a 4x16 block in Figure 4 is a vertically arranged rectangle, but in practice can also be divided into horizontal directions Arranged rectangles, such as a 16x16 block, can be divided into 4 vertical 16x4 blocks. The rectangular block may include various forms, and it is understood that a linear block such as 1x16 or 16x1 is also a type of rectangular block.

S32c: predicting the first sub-image block and the second sub-image block according to adjacent image blocks of the current image block from a plurality of predefined prediction directions, respectively obtaining the first sub-image block The prediction direction and the prediction direction of the second sub-image block. For a specific method for predicting the prediction direction of an image block, refer to the prior art, which is not specifically described in this embodiment.

S33c: Determine a most probable prediction direction of the first sub-image block according to a prediction direction of a neighboring image block of the current image block.

The process may adopt a prior art method of selecting a block with the smallest prediction mode value among the upper neighboring block and the left neighboring block of the current image block, and the prediction direction of the selected block is taken as the first in the current image block. The most likely predicted direction of a sub-image block. In practical applications, each prediction mode may correspond to a value, that is, each mode may correspond to a number, which is used to reflect the size of the mode value, and when the block is selected, the prediction mode may be selected according to the number. The block with the smallest value. Here, when one image block has 2 neighboring blocks, the block with the smallest prediction mode value may be selected as a reference to predict the most probable prediction direction of the current block or the subblock in the current block. Of course, if only one of the left neighboring block and the upper neighboring block of the current block is available, the predicted direction of the available neighboring block is used as the most probable prediction direction. If the left neighboring block and the upper neighboring block are not available, a predefined default value is used as the most probable prediction mode.

S34c: Determine a prediction direction of the first sub-image block as a most probable prediction direction of the second sub-image block.

S35c: determining, according to a prediction direction of the first sub-image block and a most probable prediction direction of the first sub-image block, a value that is required to be written into the code stream to identify a prediction direction of the first sub-image block; Determining, according to a prediction direction of the second sub-image block and a most probable prediction direction of the second sub-image block, a value of a prediction direction of the second sub-image block that needs to be written into the code stream.

The following is a brief introduction to the process of how to write the identity of an image block in the code stream. Assuming that the prediction mode of the current sub-block is ilntraDirLuma, the most probable prediction mode of the current sub-block is iMostProbable. Then, if ilntraDirLuma is equal to iMostProbable, a bit of the identifier is written in the code stream, for example, the identifier "1" can be used to indicate that the prediction mode of the current sub-block is the most probable prediction mode. If ilntraDirLuma is not equal to iMostProbable, the value of the above identifier is first identified as "0", that is, the prediction mode of the current sub-block is not equal to the most probable prediction mode, and further judgment is made. If the actual prediction direction of the current sub-block is smaller than the predicted most probable prediction direction, the actual prediction direction value ilntraDirLuma of the current sub-block is written in the code stream, if the actual prediction direction of the current sub-block is greater than the predicted most probable prediction direction. , then write ilntraDirLuma - 1 in the stream.

The method in this embodiment can be used for 4x16 blocks or 16x4 blocks in 16x16 block partitioning, or 2x8 blocks or 8x2 blocks in 8x8 block partitioning, and the like. Of course, for 8x32 in 32x32 squares Or the 32x8 partition, and the 1x16 partition in 4x16, or the 16x1 partition in 16x4 can also use the method in this embodiment.

S36c: After receiving the code stream, the decoding end parses and obtains the division mode information of the current image, the identification information of the prediction direction of the first sub-image block, and the prediction direction of the second sub-image block from the code stream. Identification information.

S37c: Divide the current image block according to the division mode information to obtain the plurality of sub image blocks.

S38c: Determine a most probable prediction direction of the first sub-image block according to a prediction direction of a neighboring image block of the current image block. The execution process of this step can be seen in S33.

S39c: obtaining, from the identification information of the prediction direction of the first sub-image block and the most likely prediction direction of the first sub-image block, a prediction direction of the first sub-image block.

S310c: Use a prediction direction of the first sub-image block as a most probable prediction direction of the second sub-image block.

S311 c: obtaining a prediction direction of the second sub-image block from the identification information of the prediction direction of the second sub-image block and the most probable prediction direction of the second sub-image block.

The process of determining the prediction direction of the image block based on the identifier in S39c and S311c is the inverse of S35c. Specifically, the most probable prediction direction of the current sub-block is obtained first, and an identifier that identifies whether the prediction direction of the current sub-block is the most likely prediction direction is parsed in the code stream. For example, if the identifier is indicated as "1", the prediction direction of the current sub-block is the most probable prediction direction of the current sub-block. If the identifier indicates "0", further parsing the code stream, obtaining a value of the prediction direction identifier of the current sub-block, comparing the value of the prediction direction identifier with the value of the most probable prediction direction of the current sub-block Relationship, if the value of the identifier is greater than or equal to the value of the most probable prediction direction, the actual prediction direction of the current sub-block is incremented by 1 for the value of the identifier, if the value of the identifier It can be understood that the sequence of some steps of the embodiment of the present invention can be adaptively adjusted. For example, the execution of step S33c may be after step S34c. By adopting the method of the embodiment of the present invention, the most probable prediction direction of a block can be effectively predicted, thereby determining an identifier of the write code stream for indicating the prediction direction of the image block, when the actual prediction direction of the image block is equal to its most likely When predicting the direction, only one bit of the indication flag needs to be written in the code stream, so when the most probable prediction direction of a block is closer to the actual prediction direction of the block, it will be more advantageous to save the code stream. When a current image block and its neighboring blocks are arranged in a horizontal direction or a vertical direction, only the actual prediction mode of the neighboring block may be utilized as the most probable prediction mode of the current image block, without having to compare the predictions of the two neighboring blocks of the current image block. Mode values for simple and effective prediction. In particular, when there are multiple shape partitions in a sub-block in an image block, the left adjacent block and the upper adjacent block of one sub-block have a large difference in the division manner, and different sub-blocks in a block that are close or identical by the division manner are used. Blocks to predict the most probable prediction mode can achieve better results than the prior art. FIG. 5 is a schematic diagram of another image block division structure according to an embodiment of the present invention. It is assumed that the current image block 50 is an 8x8 block including four 8x2 sub-blocks arranged in the horizontal direction, which are sub-block 51, sub-block 52, sub-block 53, and sub-block 54, respectively. The sub-block 51 is located at the boundary of the current image block 50, and its most probable prediction mode is obtained by the prior art, that is, from the left adjacent block of the current image block 50 and the upper adjacent block (for the left adjacent block and the upper adjacent The prediction mode of the block with the smallest prediction mode value is selected as the most probable prediction mode of the sub-block 51 in the block diagram; the most probable prediction mode of the sub-block 52 is the prediction mode of the sub-block 51; The most probable prediction mode is the prediction mode of sub-block 52; the most probable prediction mode of sub-block 54 is the prediction mode of sub-block 53. Since there is a certain correlation between the four sub-blocks, the prediction modes of the sub-blocks adjacent to each other may have the same prediction mode to some extent, so the prediction mode of the adjacent sub-block is taken as the current sub-block. The most likely prediction mode.

FIG. 6 is a schematic diagram of another encoding method of an intra block based intra prediction mode according to an embodiment of the present invention, where the method includes:

S61: Obtain a prediction mode of a first sub-block in a current image block, where the first sub-block includes a plurality of sub-blocks arranged in a horizontal direction or a vertical direction;

S62: determining a second sub-block, and using the prediction mode of the second sub-block as the most likely of the first sub-block a prediction mode, the second sub-block is adjacent to the first sub-block and includes a plurality of sub-blocks arranged in a horizontal direction or a vertical direction;

S63: Determine, according to the prediction mode of the first sub-block and the most probable prediction mode of the first sub-block, an identifier for indicating a prediction mode of the first sub-block, and write the identifier into the code stream.

In the embodiment of the present invention, the current image block includes a plurality of sub-blocks. If each sub-block includes a plurality of sub-blocks arranged in a horizontal direction or a vertical direction, the prediction mode of one of the sub-blocks may be used as the adjacent sub-block. The most probable prediction mode of the block, the most probable prediction mode of the adjacent sub-block is applicable to all sub-blocks in the adjacent sub-block, and it is not necessary to separately solve one most probable prediction mode for each sub-block in the adjacent sub-block. It can be understood that the order of S61 and S12 is not fixed and can be flexibly adjusted.

In the normal case, if there is a similar sub-block division between two neighboring blocks, the prediction direction of one of the blocks can be regarded as the most probable prediction direction of the other block, and the most probable prediction direction is used as the The most probable prediction direction of all sub-blocks in another block may cause the most probable prediction direction of all sub-blocks in the other block to be close to the true prediction direction, thereby reducing the number of identification bits indicating the prediction direction transmitted in the code stream. The division of the two blocks is similar, and the division manners may be identical, or the sub-blocks within one block may be the same size and shape as the sub-blocks in another block, but the sub-blocks are arranged in different directions. Generally, as long as two blocks include a plurality of sub-blocks arranged in the horizontal direction or the vertical direction, the two blocks can be considered to be arranged in a similar manner. The two adjacent blocks may each be a square, and the sub-blocks in each block may be non-square rectangular blocks and may be arranged in a horizontal or vertical direction, as shown in FIG.

In this embodiment, the second sub-block and the first sub-block may both be located in the current image block, and of course, the second sub-block is not located outside the current image block. The second sub-block may be an upper neighboring block or a left neighboring block of the first sub-block, or the second sub-block may be the block of the upper neighboring block and the left neighboring block of the first sub-block having the smallest prediction mode value. That is, only one neighboring block of the first sub-block that is similar to the first sub-block partitioning manner may be used as a reference for determining the most probable prediction mode of the first sub-block. Of course, if the left adjacent block and the upper adjacent block of the first sub-block have a similar division manner as the first sub-block, that is, each includes a plurality of sub-blocks arranged in the horizontal direction or the vertical direction, the prediction may be selected. The block with the smallest mode value is used to determine the first sub-block A reference to the possible prediction mode.

For the sake of understanding, the prediction method provided by the present embodiment is exemplified below, and the drawings and the description are only for convenience of understanding and are not intended to limit the present invention. FIG. 7 is a schematic diagram of sub-block partitioning within 8×8 blocks according to an embodiment of the present invention. The 8×8 block is divided into four 4×4 sub-blocks, which are identified by labels 1, 2, 3, and 4, respectively. A plurality of 1x4 blocks or 4x1 blocks may be further divided in each 4x4 block, wherein the sub-blocks 1 and the sub-blocks 4 have the same arrangement, and the sub-block 2 and the sub-block 3 have the same arrangement. Each of the 4x4 blocks needs to have a prediction direction, and each 1x4 block or 4x1 block in the 4x4 block does not have to separately encode the prediction direction, that is, the prediction direction is encoded in units of 4x4 sub-blocks. Before encoding the prediction direction of 4x4 blocks, it is necessary to obtain the most probable prediction direction for each 4x4 block.

For the first sub-block 1 , the prediction direction in which the prediction direction value is the smallest in the left adjacent block and the upper adjacent block of the current image block in which it is located may be taken as the most probable prediction direction of the first sub-block 1;

For the second sub-block 2, the most likely prediction direction is the prediction direction of the first sub-block 1; optionally, if the upper adjacent block of the second sub-block 2 (not shown) is also 1x4 or 4x1 The block composition is taken as the prediction direction with the smallest prediction direction value in the left adjacent block and the upper adjacent block as the most probable prediction direction of the second sub-block 2;

For the third sub-block 3, the most likely prediction direction is the prediction direction of the first sub-block 1; optionally, if the left adjacent block of the third sub-block 3 (not shown) is also 1x4 or 4x1 The block composition is taken as the prediction direction with the smallest prediction direction value in the left adjacent block and the upper adjacent block as the most probable prediction direction of the third sub-block 3; the small prediction direction is the most the fourth sub-block 4 It is possible to predict the direction.

Corresponding to the embodiment of FIG. 6, FIG. 8 is a schematic diagram of another decoding method of an intra block prediction mode based on an image block according to an embodiment of the present invention, where the method includes:

S81: Obtain an identifier for indicating a prediction mode of the first sub-block in the current image block from the code stream, where the first sub-block includes a plurality of sub-blocks arranged in a horizontal direction or a vertical direction; S82: determining a second sub-block, using a prediction mode of the second sub-block as a most probable prediction mode of the first sub-block, where the second sub-block is adjacent to the first sub-block and includes horizontally or vertically Multiple sub-blocks arranged in a straight direction;

S83: Determine a prediction mode of the first sub-block according to the identifier and a most probable prediction mode of the first sub-block.

Similar to the previous embodiment shown in FIG. 3, the code stream can not only convey the identification of the prediction mode of the image block, but also the division mode information of the image block, so that the decoding end can know how the coding end divides the image block.

The embodiment of the present invention further provides an apparatus for encoding and decoding a code stream by using the method of the previous embodiment. Specifically, FIG. 9 is a schematic structural diagram of an encoder according to an embodiment of the present invention, including: an obtaining unit 91, And acquiring a prediction mode of the first sub-block in the current image block and a prediction mode of the second sub-block in the current image block, the first sub-block being adjacent to the second sub-block and horizontally or vertically Arrange in direction

a determining unit 92, configured to use a prediction mode of the first sub-block as a most probable prediction mode of the second sub-block;

a code stream writing unit 93, configured to determine, according to a prediction mode of the second sub-block and a most probable prediction mode of the second sub-block, an identifier for indicating a prediction mode of the second sub-block, and write the identifier into the code In the stream.

If the first sub-block is located at the current image block boundary, the determining unit 92 is further configured to predict a block with the smallest prediction mode value among the upper adjacent block and the left adjacent block of the current image block. The mode is used as the most probable prediction mode of the first sub-block, and the code stream writing unit 93 is further configured to determine, according to the prediction mode of the first sub-block and the most probable prediction mode of the first sub-block, for indicating the first sub-block. Another identifier of the prediction mode is written and written into the code stream.

Further, the encoder may further include a block dividing unit 94, configured to divide the current image block to obtain a horizontal direction or a vertical direction including the first sub-block and the second sub-block. a plurality of sub-blocks arranged in a straight direction; the code stream writing unit 93 is further configured to divide a current image block The information is written into the code stream.

FIG. 10 is a schematic structural diagram of a decoder according to an embodiment of the present invention, including:

a code stream reading unit 101, configured to acquire, from the code stream, an identifier for indicating a prediction mode of the second sub-block in the current image block;

a determining unit 102, configured to use a prediction mode of the first sub-block in the current image block as a most probable prediction mode of the second sub-block, where the first sub-block is adjacent to the second sub-block and in a horizontal direction or Arranged in the vertical direction;

The decoding unit 103 is configured to determine a prediction mode of the second sub-block according to the identifier and a most probable prediction mode of the second sub-block.

If the first sub-block is located at the current image block boundary, the code stream readout unit 101 is further configured to acquire another identifier for indicating a prediction mode of the first sub-block from the code stream;

The determining unit 102 is further configured to use, as a most probable prediction mode of the first sub-block, a prediction mode of a block in which a prediction mode value is the smallest among the upper neighboring block and the left neighboring block of the current image block;

The decoding unit 103 is further configured to determine a prediction mode of the first sub-block according to the another identifier and a most probable prediction mode of the first sub-block.

Further, the code stream readout unit 101 is further configured to acquire, by using the code stream, the split mode information of the current image block. The decoder further includes a block partitioning unit 104, configured to perform the partitioning manner according to the partitioning manner. The current image block is divided to obtain a plurality of sub-blocks arranged in the horizontal direction or the vertical direction including the first sub-block and the second sub-block.

FIG. 11 is a schematic structural diagram of another encoder according to an embodiment of the present disclosure, where the encoder includes:

The obtaining unit 111 is configured to acquire a prediction mode of the first sub-block in the current image block, where the first sub-block includes a plurality of sub-blocks arranged in a horizontal direction or a vertical direction;

a determining unit 112, configured to determine a second sub-block, using a prediction mode of the second sub-block as a most probable prediction mode of the first sub-block, where the second sub-block is adjacent to the first sub-block and includes an edge Horizontal or vertical Multiple sub-blocks arranged in a straight direction;

a code stream writing unit 113, configured to determine, according to a prediction mode of the first sub-block and a most probable prediction mode of the first sub-block, an identifier for indicating a prediction mode of the first sub-block, and write the identifier into the code In the stream.

For the determining unit 112, if the upper adjacent block of the first sub-block includes a plurality of sub-blocks arranged in a horizontal direction or a vertical direction, determining the upper adjacent block as the second sub-block Block; or, the left neighboring block is determined to be the second sub-block; or if the upper neighboring block and the left neighboring block of the first sub-block both include a plurality of horizontally or vertically arranged a sub-block, the block having the smallest prediction mode value among the upper neighboring block and the left neighboring block is determined as the second sub-block.

If the second sub-block is located at the current image block boundary, the determining unit 112 is further configured to: use a block that minimizes a prediction mode value among the upper adjacent block and the left adjacent block of the current image block. The prediction mode is used as the most probable prediction mode of the second sub-block; the code stream writing unit 113 is further configured to determine, according to the prediction mode of the second sub-block and the most probable prediction mode of the second sub-block, Another identification of the prediction mode of the sub-block and writing the other identification into the code stream.

FIG. 12 is a schematic structural diagram of another decoder according to an embodiment of the present disclosure, where the decoder includes:

The code stream reading unit 121 is configured to acquire, from the code stream, an identifier for indicating a prediction mode of the first sub-block in the current image block, where the first sub-block includes a plurality of sub-blocks arranged in a horizontal direction or a vertical direction. a determining unit 122, configured to determine a second sub-block, using a prediction mode of the second sub-block as a most probable prediction mode of the first sub-block, where the second sub-block is adjacent to the first sub-block and includes a plurality of sub-blocks arranged in a horizontal direction or a vertical direction;

The decoding unit 123 is configured to determine a prediction mode of the first sub-block according to the identifier and a most probable prediction mode of the first sub-block.

For the determining unit 122, if the upper adjacent block of the first sub-block includes a plurality of sub-blocks arranged in a horizontal direction or a vertical direction, the upper adjacent block is determined as the second sub-block Or, The left adjacent block is determined as the second sub-block; or, if the upper adjacent block and the left adjacent block of the first sub-block each include a plurality of sub-blocks arranged in a horizontal direction or a vertical direction, A block having the smallest prediction mode value among the upper neighboring block and the left neighboring block is determined as the second sub-block.

If the second sub-block is located at the current image block boundary, the code stream readout unit 121 is further configured to acquire another identifier for indicating a prediction mode of the second sub-block from the code stream; The determining unit 122 is further configured to use, as a most probable prediction mode of the second sub-block, a prediction mode of a block that is the smallest prediction mode value among the upper neighboring block and the left neighboring block of the current image block; The unit 123 is further configured to determine a prediction mode of the second sub-block according to the another identifier and a most probable prediction mode of the second sub-block.

The codec provided by the embodiment of the present invention is respectively located at the encoding end and the decoding end for encoding and decoding the code stream, and by effectively predicting the most probable prediction mode of the image block, the most probable prediction mode of an image block is closer to its actual prediction. The mode reduces the number of bits in the image block prediction mode reflected in the code stream, and improves the codec performance.

A person skilled in the art can understand that all or part of the processes in the above method embodiments can be completed by a computer program to instruct related hardware, and the program can be stored in a computer readable storage medium, and the program is executed. At the time, the flow of the embodiment of each of the above methods may be included. 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 only a few embodiments of the present invention, and various modifications and changes may be made thereto without departing from the spirit and scope of the invention. In the case of the skill, it is possible to combine with each other to form a new embodiment.

Claims

Rights request

An encoding method of an intra prediction mode based on an image block, comprising: acquiring a prediction mode of a first sub-block in a current image block and a prediction mode of a second sub-block in the current image block, where The first sub-block is adjacent to the second sub-block and arranged in a horizontal direction or a vertical direction; using the prediction mode of the first sub-block as the most probable prediction mode of the second sub-block;

And determining, according to the prediction mode of the second sub-block and the most probable prediction mode of the second sub-block, an identifier for indicating a prediction mode of the second sub-block, and writing the identifier into the code stream.

The method of claim 1, wherein the first sub-block is located at a boundary of the current image block, the method further comprising:

The prediction mode of the block having the smallest prediction mode value among the upper neighboring block and the left neighboring block of the current image block is taken as the most probable prediction mode of the first sub-block;

Another identifier for indicating a prediction mode of the first sub-block is determined according to a prediction mode of the first sub-block and a most probable prediction mode of the first sub-block, and the another identifier is written into the code stream.

3. The method according to claim 1 or 2, further comprising:

Dividing the current image block to obtain a plurality of sub-blocks arranged in a horizontal direction or a vertical direction, including the first sub-block and the second sub-block;

The division mode information of the current image block is written into the code stream.

4. The method according to claim 3, wherein the first sub-block and the second sub-block are non-square rectangular blocks.

A decoding method of an intra prediction mode based on an image block, comprising: acquiring, from a code stream, an identifier for indicating a prediction mode of a second sub-block in a current image block; The prediction mode of the first sub-block is the most probable prediction mode of the second sub-block, and the first sub-block is adjacent to the second sub-block and arranged in a horizontal direction or a vertical direction;

Determining a prediction mode of the second sub-block according to the identifier and a most probable prediction mode of the second sub-block.

The method of claim 5, wherein the first sub-block is located at the boundary of the current image block, the method further includes:

Obtaining another identifier for indicating a prediction mode of the first sub-block from the code stream;

Determining a prediction mode of the first sub-block according to the another identifier and a most probable prediction mode of the first sub-block.

7. The method according to claim 5 or 6, further comprising:

Obtaining division mode information of the current image block from the code stream;

The current image block is divided according to the division manner, and a plurality of sub-blocks arranged in a horizontal direction or a vertical direction including the first sub-block and the second sub-block are obtained.

8. The method according to claim 7, wherein the first sub-block and the second sub-block are non-square rectangular blocks.

9. An encoding method of an intra prediction mode based on an image block, comprising: acquiring a prediction mode of a first sub-block in a current image block, the first sub-block comprising being arranged in a horizontal direction or a vertical direction Multiple sub-blocks;

The method of claim 9, wherein the determining the second sub-block comprises: a block determining the upper neighboring block as the second sub-block; or Blocking, determining the left neighboring block as the second sub-block; or

If the upper adjacent block and the left adjacent block of the first sub-block each include a plurality of sub-blocks arranged in a horizontal direction or a vertical direction, prediction is performed among the upper adjacent block and the left adjacent block The block with the smallest mode value is determined as the second sub-block.

The method according to claim 9, wherein the first sub-block and the second sub-block are each a square.

The method according to claim 9, wherein the plurality of sub-blocks in the second sub-block and the plurality of sub-blocks in the first sub-block are non-square rectangular blocks.

The method according to any one of claims 9 to 12, wherein the second sub-block is located at a boundary of the current image block, the method further comprising:

The prediction mode of the block having the smallest prediction mode value among the upper neighboring block and the left neighboring block of the current image block is taken as the most probable prediction mode of the second sub-block;

Another identifier for indicating a prediction mode of the second sub-block is determined according to a prediction mode of the second sub-block and a most probable prediction mode of the second sub-block, and the another identifier is written into the code stream.

A method for decoding an intra prediction mode based on an image block, comprising: acquiring, from a code stream, an identifier for indicating a prediction mode of a first sub-block in a current image block, the first sub-block Including a plurality of sub-blocks arranged in a horizontal direction or a vertical direction;

The method according to claim 14, wherein the determining the second sub-block comprises: a block determining the upper neighboring block as the second sub-block; or Blocking, determining the left neighboring block as the second sub-block; or

16. The method of claim 14, wherein the first sub-block and the second sub-block are all squares.

The method according to claim 14, wherein the plurality of sub-blocks in the second sub-block and the plurality of sub-blocks in the first sub-block are non-square rectangular blocks.

The method according to any one of claims 14 to 17, wherein the second sub-block is located at the boundary of the current image block, the method further comprising:

Obtaining another identifier for indicating a prediction mode of the second sub-block from the code stream;

A prediction mode of the second sub-block is determined based on the another identifier and a most probable prediction mode of the second sub-block.

19. An encoder, comprising:

The encoder according to claim 19, wherein the first sub-block is located in the Current image block boundary;

The determining unit is further configured to use, as a most probable prediction mode of the first sub-block, a prediction mode of a block in which a prediction mode value among the upper neighboring block and the left neighboring block of the current image block is the smallest;

The code stream writing unit is further configured to determine another identifier for indicating a prediction mode of the first sub-block according to a prediction mode of the first sub-block and a most probable prediction mode of the first sub-block, and Another identifier is written into the code stream.

The encoder according to claim 19 or 20, further comprising: a block dividing unit, configured to divide the current image block to obtain the first sub-block and the second sub- a plurality of sub-blocks arranged in a horizontal direction or a vertical direction inside the block;

The code stream writing unit is further configured to write the division mode information of the current image block into the code stream.

22. A decoder, comprising:

And a decoding unit, configured to determine a prediction mode of the second sub-block according to the identifier and a most probable prediction mode of the second sub-block.

The decoder according to claim 22, wherein the first sub-block is located at a boundary of the current image block;

The code stream reading unit is further configured to acquire another identifier for indicating a prediction mode of the first sub-block from the code stream;

The determining unit is further configured to use, as a most probable prediction mode of the first sub-block, a prediction mode of a block that has a minimum prediction mode value among the upper neighboring block and the left neighboring block of the current image block; The decoding unit is further configured to determine a prediction mode of the first sub-block according to the another identifier and a most probable prediction mode of the first sub-block.

The decoder according to claim 22 or 23, wherein the code stream reading unit is further configured to acquire, according to the code stream, the division mode information of the current image block;

The decoder further includes: a block dividing unit, configured to divide the current image block according to the dividing manner, to obtain a horizontal direction or the first sub-block and the second sub-block Multiple sub-blocks arranged in the vertical direction.

25. An encoder, comprising:

An obtaining unit, configured to acquire a prediction mode of a first sub-block in a current image block, where the first sub-block includes a plurality of sub-blocks arranged in a horizontal direction or a vertical direction;

The decoder according to claim 25, wherein the determining unit is specifically configured to: block, determine the upper neighboring block as the second sub-block; or block, the left phase a neighboring block is determined as the second sub-block; or

The decoder according to claim 25 or 26, wherein the second sub-block is located at a boundary of the current image block; The determining unit is further configured to use, as a most probable prediction mode of the second sub-block, a prediction mode of a block in which a prediction mode value is the smallest among the upper neighboring block and the left neighboring block of the current image block;

The code stream writing unit is further configured to determine another identifier for indicating a prediction mode of the second sub-block according to the prediction mode of the second sub-block and the most probable prediction mode of the second sub-block, and Another identifier is written into the code stream.

28. A decoder, comprising:

a code stream reading unit, configured to acquire, from the code stream, an identifier for indicating a prediction mode of the first sub-block in the current image block, where the first sub-block includes a plurality of sub-blocks arranged in a horizontal direction or a vertical direction;

And a decoding unit, configured to determine a prediction mode of the first sub-block according to the identifier and a most probable prediction mode of the first sub-block.

The decoder according to claim 28, wherein the determining unit is specifically configured to: block, determine the upper neighboring block as the second sub-block; or block, the left phase a neighboring block is determined as the second sub-block; or

The decoder according to claim 28 or 29, wherein the second sub-block is located at a boundary of the current image block;

The code stream reading unit is further configured to acquire another identifier for indicating a prediction mode of the second sub-block from the code stream; The determining unit is further configured to use, as a most probable prediction mode of the second sub-block, a prediction mode of a block in which a prediction mode value is the smallest among the upper neighboring block and the left neighboring block of the current image block;

The decoding unit is further configured to determine a prediction mode of the second sub-block according to the another identifier and a most probable prediction mode of the second sub-block.