WO2012097742A1

WO2012097742A1 - Method, module and encoding/decoding device for obtaining reference motion vector

Info

Publication number: WO2012097742A1
Application number: PCT/CN2012/070586
Authority: WO
Inventors: 郑萧桢
Original assignee: 华为技术有限公司
Priority date: 2011-01-19
Filing date: 2012-01-19
Publication date: 2012-07-26
Also published as: CN102611881A; CN102611881B

Abstract

Provided are a method, a module and an encoding/decoding device for obtaining a reference motion vector. The method includes: obtaining a division mode of a sub-image block in a current image block; determining a first candidate reference motion vector from candidate reference motion vectors of the sub-image block according to the division mode, wherein the candidate reference motion vector is a motion vector of an image block or sub-image block adjacent to the current image block or a motion vector obtained after performing an operation on the motion vector of the image block or the sub-image block adjacent to the current image block; numbering each candidate reference motion vector according to the first candidate reference motion vector; and selecting the reference motion vector from the numbered candidate reference motion vectors. The method, module and encoding/decoding device for obtaining a reference motion vector provided by the embodiments of the present invention can improve the number encoding efficiency and can further improve the video encoding/decoding efficiency since the encoding end carries out variable length encoding on the number.

Description

Reference motion vector acquisition method, module, and coding and decoding device

The present invention relates to the field of information technology, and in particular, to a reference motion vector acquisition method, a module, and a coding and decoding apparatus. Background technique

In the existing video codec standard, an image is usually divided into several image blocks for encoding and decoding, and one image block can be further divided into a plurality of sub-image blocks. The encoding end usually searches for a neighboring encoded sub-image block that is most similar (ie, matched) to the current image block or sub-image block, and subtracts the pixel value corresponding to the current image block or sub-image block from the matched sub-image block by the residual. Then, the residual is subjected to transform and quantization, entropy coding, and finally the bit stream obtained by entropy coding is written into the coded stream. Among them, adjacent coded sub-image blocks are referred to as sub-reference image blocks. The offset between the current image block or sub-image block and the matched sub-image block is called a motion vector, and the encoding end encodes the motion vector and writes it into the encoded code stream to enable the decoding end to The motion vector knows the location of the matched sub-image block. In order to reduce the length of the byte occupied by the motion vector information in the code stream, a reference motion vector (Motion Vector Candidate) is usually used, and the difference between the motion vector and the reference motion vector is encoded and written into the coded stream. In the prior art, a method for obtaining a reference motion vector, the encoding end and the decoding end follow the same candidate reference motion vector table, where the candidate reference motion vector table includes motion vectors of several sub-image blocks adjacent to the current image block or the sub-image block. (as a candidate reference motion vector) and its corresponding number. The encoding end selects one of the candidate reference motion vector tables as the candidate reference motion vector, encodes the difference between the offset and the reference motion vector by using variable length coding, and numbers the reference motion vector by using variable length coding. Encoding is performed so that the decoding end knows which candidate reference motion vector is selected by the encoding end. However, the existing method of obtaining a reference motion vector does not consider the current image block or sub-image block. The degree of correlation with adjacent sub-image blocks, therefore, affects the coding efficiency of the motion vectors. Summary of the invention

The embodiment of the invention provides a reference motion vector acquisition method, a module, and a coding and decoding device, so as to solve the problem that the correlation between the current image block or the sub-image block and the adjacent sub-image block is not considered in the prior art, and the motion vector is affected. The problem of coding efficiency.

An embodiment of the present invention provides a reference motion vector acquisition method, including:

Obtaining a division manner of the sub image block in the current image block;

Determining, in the dividing manner, a first candidate reference motion vector in each candidate reference motion vector of the sub-image block, where the candidate reference motion vector is an motion of an image block or a sub-image block adjacent to the current image block. a vector, or a motion vector obtained by the motion vector of the adjacent image block or sub-image block;

Each candidate reference motion vector is numbered according to the first candidate reference motion vector; a reference motion vector is selected from each of the numbered candidate reference motion vectors.

An embodiment of the present invention further provides a reference motion vector acquisition module, including:

a identifying unit, configured to obtain a dividing manner of the sub image block in the current image block;

An analyzing unit, configured to determine, in the partitioning manner, a first candidate reference motion vector among each candidate reference motion vector of the sub-image block, where the candidate reference motion vector is an image block adjacent to the current image block or a motion vector of the sub-image block, or a motion vector obtained by the motion vector of the adjacent image block or the sub-image block;

a numbering unit, configured to number each candidate reference motion vector according to the first candidate reference motion vector;

And a selecting unit, configured to select a reference motion vector according to each of the candidate reference motion vectors after the numbering. An embodiment of the present invention further provides an encoding apparatus, including: a prediction module, a transform module, a quantization module, an entropy encoding module, and a reference motion vector acquiring module;

The reference motion vector acquisition module is configured to obtain a division manner of the sub image block in the current image block; Determining, in the dividing manner, a first candidate reference motion vector in each candidate reference motion vector of the sub-image block, where the candidate reference motion vector is an motion of an image block or a sub-image block adjacent to the current image block. a vector, or a motion vector obtained by the motion vector of the adjacent image block or sub-image block; and each candidate reference motion vector is numbered according to the first candidate reference motion vector; Selecting a reference motion vector in the motion vector;

The prediction module is configured to acquire a pixel residual between the sub-image block and the predicted image block, and obtain a difference between a motion vector of the predicted image block and the reference motion vector;

The transform module is configured to transform the residual obtained by the prediction module;

The quantization module is configured to quantize the residual obtained by the prediction module transformed by the transform module;

The entropy encoding module is configured to entropy encode the residual quantized by the quantization module, and perform variable length entropy coding on a difference between a motion vector of the predicted image block and the reference motion vector, where The number corresponding to the reference motion vector is variable length entropy encoded.

An embodiment of the present invention further provides a decoding apparatus, including: an entropy decoding module, an inverse quantization module, an inverse transform module, a prediction compensation module, and a reference motion vector acquisition module;

The entropy decoding module is configured to perform entropy decoding on the received encoded code stream, obtain a pixel residual between each sub-image block and the predicted image block in the current image block, and obtain a motion vector of the predicted image block and the reference The difference between the motion vectors and the number corresponding to the reference motion vector;

The inverse quantization module is configured to perform inverse quantization on the residual;

The inverse transform module is configured to perform inverse transform on a residual that is inversely quantized by the inverse quantization module, where the reference motion vector acquisition module is configured to obtain a partition mode of the sub image block in the current image block; And determining, in each candidate reference motion vector of the sub-image block, a first candidate reference motion vector, where the candidate reference motion vector is a motion vector of an image block or a sub-image block adjacent to the current image block, or a motion vector obtained by calculating a motion vector of an adjacent image block or a sub-image block; calculating each candidate reference motion vector according to the first candidate reference motion vector; selecting from each of the numbered candidate reference motion vectors Reference motion vector. The prediction compensation module is configured to acquire a corresponding reference motion vector according to the number, obtain the motion vector according to a difference between the motion vector and the reference motion vector, and determine the predicted image block according to the motion vector, And acquiring the current image block according to the predicted image block and the residual.

The reference motion vector obtaining method, the module, and the encoding and decoding apparatus provided by the embodiment of the present invention determine the first candidate reference motion vector of the sub-image block according to the division manner of the sub-image block in the current image block, and according to the first reference motion vector Each candidate reference motion vector is numbered so that the first candidate reference motion vector corresponds to a smaller number. Since the encoding end uses variable length coding, the coding efficiency of the number can be improved, thereby improving the video coding and decoding efficiency. .

DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings 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. One of ordinary skill in the art can also obtain other drawings based on these drawings without undue creative effort.

1 is a flowchart of a first embodiment of a reference motion vector acquisition method provided by the present invention; FIG. 2a is a schematic diagram of asymmetrically dividing a current image block into two left and right sub-image blocks; FIG. 2b is a schematic diagram of dividing a current image block into left and right. a schematic diagram of two sub-image blocks;

3a is a schematic diagram of asymmetrically dividing a current image block into upper and lower sub-image blocks; FIG. 3b is a schematic diagram of symmetrically dividing a current image block into upper and lower sub-image blocks;

4 is a schematic diagram of a sub-image block adjacent to a current image block;

5 is a schematic diagram of identifying a sub-image block in a current image block by a ratio of a dividing line in a current image block to a distance of a boundary line of a current image block or a boundary line extension line to an end point of the image block; FIG. A schematic diagram of the division of the sub-image block in the current image block by the angle and distance of the dividing line in the current image block;

FIG. 7 is a schematic structural diagram of a reference motion vector acquisition module according to a first embodiment of the present invention; FIG. 8 is a schematic structural diagram of a reference motion vector acquisition module according to a second embodiment of the present invention; FIG. 9 is a schematic structural diagram of a first embodiment of an encoding apparatus according to an embodiment of the present invention; FIG. 10 is a schematic structural diagram of a first embodiment of a decoding apparatus according to an embodiment of the present invention.

detailed description

The technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. It is a partial embodiment of the invention, and not all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

FIG. 1 is a flowchart of a first embodiment of a method for acquiring a reference motion vector according to the present invention. As shown in FIG. 1, the method includes:

5101. Obtain a division manner of the sub image block in the current image block.

The first candidate reference motion vector is determined in each candidate reference motion vector of the sub-image block according to the dividing manner, where the candidate reference motion vector is an image block or a sub-image block adjacent to the current image block. Motion vector, or a motion vector obtained by the motion vector of the adjacent image block or sub-image block after the operation;

5103. No., each candidate reference motion vector is numbered according to the first candidate reference motion vector;

5104. Select a reference motion vector from each of the candidate reference motion vectors after the numbering.

An image usually contains texture information in various directions. These textures are often not a rectangular area. An image usually contains multiple objects, and there is a clear boundary between the object and the image background. When performing video encoding, it is usually necessary to divide an image into a plurality of image blocks, which are M*N rectangular blocks, where M and N are integers greater than 0, and there is a multiple between M and N. relationship. Since an image block is a basic unit for video encoding and decoding, its area is small. Therefore, usually, an image block contains one or more objects, or includes an object and an image background, according to which the encoding end can be The boundary between the object and the object, or the boundary between the object and the background of the image, the current image block is divided into two sub-image blocks, which enables Dividing the boundary of the object between the image blocks, and the division is based on the actual boundary between the object and the object, or between the object and the background of the image. Therefore, the size of the two sub-image blocks obtained after the division It may or may not be the same shape. The positions of the two sub-image blocks may be symmetrical or asymmetric.

Since the area of an image block is small, the division of an image block into two sub-image blocks is usually divided into left and right, as shown in Fig. 2a and Fig. 2b, or divided up and down, as shown in Fig. 3a and Fig. 3b.

Based on the continuity of the video image, it can be known that the image block or sub-image block associated with the current image block is adjacent to the current image block. As shown in FIG. 4, these adjacent image blocks or sub-image blocks include: Image block or sub-image block AO to ΑηΑ-1 on the left side of the current module, image block or sub-picture block B0 to BnB-1 above the current image block, image block or sub-image block D on the upper left of the current image block, current image The image block or sub-image block C at the upper right of the block and the image block or sub-image block E at the lower left of the current image block.

The current image block is divided into sub-image blocks in different ways, and the correlation degree of each image block or sub-image block adjacent to the current image block is different, and the correlation indicates the sub-image block and phase in the current image block. The degree of correlation of image information contained in adjacent image blocks or sub-image blocks.

For example, if the current image block is divided into two sub-image blocks as shown in FIGS. 2a and 2b, it can be understood that, in this division mode, since the left sub-image block and the right sub-image block are divided. It is a different object, so the correlation between the left and right sub-image blocks is weak. Normally, the left sub-image block has a higher correlation with its left image block or sub-image block AO to AnA-1, and the right sub-image block is related to its upper right image block or sub-image block C. Higher degrees.

If the current image block is divided into upper and lower sub-image blocks as shown in FIGS. 3a and 3b, it can be understood that, in this division mode, since the upper sub-image block and the lower sub-image block are different. Therefore, the correlation between the upper and lower sub-image blocks is weak. Normally, the upper sub-image block has a higher correlation with the image block or sub-image blocks B0 to BnB-1 above it, and the right sub-image block and its left image block or sub-image block AO to AnA- The correlation of 1 is higher.

It can be seen that each sub-image can be analyzed according to the division manner of the sub-image blocks in the current image block. The degree to which a block is adjacent to an image block or sub-image block adjacent to the current image block.

The encoding end and the decoding end need to follow the same set of candidate reference motion vectors when encoding and decoding, and the candidate reference motion vectors may be motion vectors of several image blocks or sub-image blocks adjacent to the current image block, and may also include The motion vector obtained by the motion vector of several adjacent image blocks or sub-image blocks is calculated, and each candidate motion vector has its corresponding number. When the encoding end performs encoding, one of the candidate reference motion vectors is selected as the final reference motion vector by a certain preset rule, and the difference between the motion vector and the reference motion vector is encoded by the variable length coding method, and the variation is adopted. The long coding mode encodes the number corresponding to the reference motion vector, so that the decoding end knows which candidate reference motion vector is selected by the coding end.

Wherein, the reference motion vector number is encoded by using a variable length coding method, and different coding numbers are used to encode different numbers. Generally it can be: For the number 0, the coding symbol "1" can be used for coding; for the number 1, the coding symbol "01" can be used for coding; for the number 2, the coding symbol "10" can be used for coding: ... and so on, it can be known that the larger the number, the more symbols are included in the encoded symbol string.

This embodiment is described by taking the following set of candidate reference motion vectors as an example, but in fact, other types of motion vectors may be used as reference motion vectors, and this is not a limitation on the embodiments of the present invention. The set of candidate reference motion vectors may include:

The motion vector obtained by the median prediction method;

The motion vector of the image block or sub-image block on the left side of the current image block/sub-image block, that is, the first motion vector available from AO to AnA-1 shown in Fig. 4; wherein the available meaning is encoded.

The motion vector of the image block or sub-image block on the current image block/sub-image block, that is, the first one of B0 to BnA-1 shown in Fig. 4 can obtain the motion vector;

The motion vector of the image block or sub-image block on the upper left side of the current image block/sub-image block, that is, the motion vector of C shown in FIG. 4;

The motion vector of the image block or the sub-image block in the upper right side of the current image block/sub-image block, that is, the motion vector of D shown in FIG. 4; The motion vector of the image block or sub-image block at the lower right of the current image block/sub-image block, that is, the motion vector of E shown in FIG. 4;

The motion vector of the previous coded image or the next coded image whose current image block/sub-image block position is the same.

It should be noted that the image block or sub-image block corresponding to the candidate reference motion vector must be an already encoded image block or sub-image block, that is, the candidate reference motion vector must be available. The motion vector obtained by the median prediction method specifically obtains: an image block or a sub-image block motion vector adjacent to the current sub-image block, and according to the motion vectors, a specific rule, such as obtaining the median value and the average value of the motion vectors. Or acquiring the nearest neighbor image block or sub-picture block motion vector according to the current image block division manner to obtain a reference motion vector. This method is often referred to as Median Prediction.

The motion vector of the sub-image block on the left side of the current image block/sub-image block needs to determine whether the AO to AnA-1 is encoded according to the number of the lower right corner of the AO to AnA-1, and the first encoded motion will be encoded. The vector is used as the corresponding candidate motion vector; likewise, the motion vector of the sub-image block on the current image block/sub-image block needs to determine whether B0 to BnA-1 are encoded according to the number from the bottom right corner of B0 to BnA-1. The first encoded motion vector is used as the corresponding candidate motion vector.

Since the encoding end encodes the difference between the motion vector and the reference motion vector by using variable length coding, in order to improve the coding efficiency, the coding end usually selects the reference motion vector closest to the motion vector value, that is, the reference motion vector and motion. The smaller the difference between the vectors, the shorter the code string used and the higher the coding efficiency. Therefore, the closer the adjacent image block or sub-image block corresponding to the reference motion vector is to the sub-image block in the current image block, the higher the coding efficiency.

While encoding the motion vector and the reference motion vector difference, the corresponding number of the selected reference motion vector is also encoded, so that the decoding end knows which candidate reference motion vector is selected by the encoding end. Since the encoding of the number is also a variable length encoding method, it is preferred to determine a first candidate reference motion vector from each candidate reference motion vector, and the adjacent image block or sub-image block corresponding to the first candidate reference motion vector Has a higher or highest correlation with the current image block. And then, according to the first candidate reference motion vector, each candidate reference motion vector is numbered, so that the first candidate parameter The number corresponding to the motion vector is as small as possible, for example: number 1, or the number corresponding to the smallest candidate reference motion vector, for example: number 0.

Since the higher the degree of correlation, the probability that the adjacent image block or the sub-image block is finally selected as the reference motion vector is the largest, and the adjacent image block or sub-image block corresponding to the first candidate reference motion vector has the current image block and the current image block. Higher or highest correlation, so the number corresponding to the first candidate reference motion vector is as small as possible, for example: number 0 or number 1, so that encoding number 0 or number 1 only requires encoding symbol "1" or " 01 " , therefore, the coding efficiency of the number can be improved, thereby improving the video coding efficiency.

After numbering each candidate reference motion vector according to the first candidate reference motion vector, selecting one reference motion vector from each of the numbered candidate reference motion vectors as the final reference motion vector, using the variable length coding method for the motion vector and the reference The difference of the motion vector is encoded, and the number corresponding to the reference motion vector is encoded by variable length coding.

The reference motion vector acquiring method provided by the embodiment of the present invention determines a first candidate reference motion vector of the sub-image block according to a division manner of the sub-image block in the current image block, and performs each candidate reference motion vector according to the first reference motion vector. The numbering is such that the first candidate reference motion vector corresponds to a smaller number. Since the encoding end uses variable length coding for the numbering, the coding efficiency of the number can be improved, thereby improving the video encoding and decoding efficiency.

On the basis of the first embodiment of the reference motion vector acquisition method, the present invention provides several preferred embodiments for identifying a division manner of a sub-image block in a current image block, including:

The manner of dividing the sub-image blocks in the current image block may be identified according to the encoding mode identifier. For example, if the coding mode identifier is SIZE_2NxN or SIZE_Nx2N, it is usually indicated that the division mode is upper-left symmetric division or left-right symmetric division; if the coding mode identifier is SIZE_2NxnU or SIZE_2NxnD, etc., it usually indicates that the division mode is upper and lower asymmetric division; SIZE_nLx2N or SIZE_nRx2N, etc., usually means that the division mode is left and right asymmetric division.

Alternatively, the sub-image block in the current image block may be identified according to a ratio of the dividing line in the current image block to the distance between the two boundary lines of the current image block or the distance from the focus of the boundary line extension line to the end point of the image block. The split mode, for example: As shown in Figure 5, the division mode can be identified by the proportional relationship between BO and AO. Alternatively, the division manner of the sub-image block in the current image block may be identified according to the angle and distance of the dividing line in the current image block. As shown in FIG. 6, the division method is identified by identifying the angle Θ of the dividing line by the geometric partitioning method (Geometric Partition).

Further, the embodiment further provides a method for determining a first candidate reference motion vector of the sub-image block in each candidate reference motion vector according to the dividing manner, including:

If the current image block is divided into two sub-image blocks, the motion vector of the image block or the sub-image block adjacent to the left side of the current image block is the first candidate reference motion vector of the left sub-image block in the current image block. The motion vector of the image block or the sub-image block adjacent to the upper right of the current image block is the first candidate reference motion vector of the right sub-image block in the current image block;

If the current image block is divided into upper and lower sub-image blocks, the motion vector of the adjacent image block or sub-image block of the current image block is the first candidate reference motion of the upper sub-image block in the current image block. Vector, the motion vector of the adjacent image block or sub-image block of the left side of the current image block is the first candidate reference motion vector of the lower sub-image block in the current image block.

Since a plurality of rules for numbering each candidate reference motion vector are stored in the prior art, as a possible implementation manner, each candidate reference motion vector is numbered according to the first candidate reference motion vector, which may be :

First, each candidate reference motion vector is numbered according to a setting rule;

The numbered candidate reference motion vectors are then reordered according to the first candidate reference motion vector.

The reordering of the numbered candidate reference motion vectors according to the first candidate reference motion vector may be:

The first candidate reference motion vector may be interchanged with the number of the candidate motion vector having the smallest number. Taking a set of candidate reference motion vectors provided in the first embodiment of the reference motion vector acquisition method as an example, the group of candidate reference motion vectors is numbered according to an existing rule, and the following is obtained:

No. 0: The motion vector obtained by the median prediction method; No. 1: The motion vector of the image block or sub-image block on the left side of the current image block/sub-image block, that is, the first available motion vector in AO to AnA-1 shown in FIG. 4;

No. 2: The motion vector of the image block or sub-image block on the current image block/sub-image block, that is, the first motion vector available in B0 to BnA-1 shown in Fig. 4;

No. 3: Current image block / motion vector of the image block or sub-image block in the upper left of the sub-image block, that is, the motion vector of C shown in Fig. 4;

No. 4: The motion vector of the image block or sub-image block in the upper right side of the current image block/sub-image block, that is, the motion vector of D shown in Fig. 4;

No. 5: Current image block / The motion vector of the image block or sub-image block at the lower right of the sub-image block, that is, the motion vector of E shown in Fig. 4;

No. 6: The motion vector of the previous coded image or the next coded image with the same position of the current image block/sub-image block.

Since the smallest number in the set of candidate reference motion vectors is number 0, and the number 0 corresponds to the motion vector obtained by the median prediction method, the motion vector obtained by the first candidate reference motion vector and the median prediction method may be used. The number is reversed. For the partitioning method of the left and right division, for the sub-image block obtained on the right side, the motion vector corresponding to the image block or the sub-image block of the upper right side, that is, the image block or the sub-image block of the upper right side is the highest. The first candidate reference motion vector, therefore, is reordered to get:

No. 0: The motion vector of the image block or sub-image block in the upper right side of the current image block/sub-image block, that is, the motion vector of D shown in Fig. 4;

No. 1 : The motion vector of the image block or sub-image block on the left side of the current image block/sub-image block, that is, the first motion vector available in AO to AnA-1 as shown in Fig. 4;

No. 3: the motion vector of the image block or sub-image block in the upper left side of the current image block/sub-image block, that is, the motion vector of C shown in FIG. 4; No. 4: The motion vector obtained by the median prediction method;

For the partitioning method of the upper and lower divisions, for the upper sub-image block obtained after the division, the motion vector corresponding to the upper image block or the sub-image block is the first candidate because the upper image block or the sub-image block is the most relevant. The reference motion vector, therefore, the reordered candidate reference motion vector table can be:

No. 0: The motion vector of the image block or sub-image block on the current image block/sub-image block, that is, the first motion vector available in B0 to BnA-1 shown in Fig. 4;

No. 2: The motion vector obtained by the median prediction method;

No. 3: Current image block / The motion vector of the image block or sub-image block on the upper left side of the sub-image block, that is, the motion vector of C shown in Fig. 4;

No. 4: The motion vector of the image block or sub-image block on the upper right side of the current image block/sub-image block, that is, the motion vector of D shown in Fig. 4;

No. 5: The motion vector of the image block or sub-image block at the lower right of the current image block/sub-image block, that is, the motion vector of E shown in Fig. 4;

For the left sub-image block obtained after the left and right division mode, and the lower sub-image block obtained by dividing the upper and lower divisions, since the most relevant one is the left sub-image block, the reordered candidate reference motion The vector table can be:

No. 0: The motion vector of the image block or sub-image block on the left side of the current image block/sub-image block, ie Figure 4 The first available motion vector from AO to AnA-1;

No. 1 : The motion vector obtained by the median prediction method;

Another reordering mode is also provided in this embodiment: the number corresponding to the first candidate reference motion vector is adjusted to the smallest number, and the numbers corresponding to the remaining candidate reference motion vectors are sequentially incremented by one. The number of the first candidate reference motion vector that will need to be given a smaller numbered value is assigned a smaller value, such as 0, and the number after the value is shifted back.

For the partitioning method of the left and right partitions, for the right sub-image block obtained after the partitioning, since the highest correlation is the upper right image block or the sub-image block, the reordered candidate reference motion vector table may be:

No. 0: Current image block / The motion vector of the image block or sub-image block on the upper right side of the sub-image block, that is, the motion vector of D shown in Fig. 4;

No. 1 : The motion vector obtained by the median prediction method;

No. 2: The motion vector of the image block or sub-image block on the left side of the current image block/sub-image block, that is, the first motion vector available in AO to AnA-1 as shown in Fig. 4;

No. 3: The motion vector of the image block or sub-image block on the current image block/sub-image block, that is, the first motion vector available in B0 to BnA-1 shown in FIG. 4; No. 4: the motion vector of the image block or sub-image block on the upper left side of the current image block/sub-image block, that is, the motion vector of C shown in FIG. 4;

For the partitioning method of the upper and lower divisions, for the upper sub-image block obtained after the division, since the uppermost image block or sub-image block is the most correlated, the reordered candidate reference motion vector table may be:

No. 1 : The motion vector obtained by the median prediction method;

For the left sub-image block obtained after the left and right division mode, and the lower sub-image block obtained by dividing the upper and lower divisions, since the image element or the sub-image block having the highest correlation is the left, the reordering is performed. The candidate reference motion vector table can be:

No. 1 : The motion vector obtained by the median prediction method;

A person skilled in the art can understand that all or part of the process of implementing the above embodiment method 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 method as described above 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).

FIG. 7 is a schematic structural diagram of a first embodiment of a reference motion vector obtaining module according to the present invention. As shown in FIG. 7, the module includes: an identifying unit 11, an analyzing unit 12, a numbering unit 13, and a selecting unit 14;

The identifying unit 11 is configured to obtain a dividing manner of the sub-image block in the current image block, and the analyzing unit 12 is configured to determine, in the dividing manner, the first candidate reference motion in each candidate reference motion vector of the sub-image block. Vector, the candidate reference motion vector is a motion vector of an image block or a sub-image block adjacent to the current image block, or a motion vector obtained after the motion vector of the adjacent image block or sub-image block is operated. ;

a numbering unit 13, configured to select each candidate reference motion vector according to the first candidate reference motion vector Numbering;

The selecting unit 14 is configured to select a reference motion vector from each of the candidate reference motion vectors after the numbering.

The reference motion vector acquisition module provided by the embodiment of the present invention determines a first candidate reference motion vector of the sub-image block according to a division manner of the sub-image block in the current image block, and performs, for each candidate reference motion vector, according to the first reference motion vector. The numbering is such that the first candidate reference motion vector corresponds to a smaller number. Since the encoding end uses variable length coding for the numbering, the coding efficiency of the number can be improved, thereby improving the video coding efficiency.

Based on the first embodiment of the reference motion vector acquisition module, referring to FIG. 8, further, the numbering unit 13 includes:

a number sub-unit 131, configured to number each candidate reference motion vector according to a setting rule; a reordering sub-unit 132, configured to perform weighting on each of the numbered candidate reference motion vectors according to the first candidate reference motion vector Sort.

The reordering subunit 132 may be specifically configured to:

The first candidate reference motion vector is interchanged with the number of the candidate motion vector with the smallest number; or, the number corresponding to the first candidate reference motion vector is adjusted to the smallest number, and the numbers corresponding to the remaining candidate reference motion vectors are sequentially incremented by one.

Further, the identification unit 11 can be specifically configured to:

Identifying a division manner of the sub-image block in the current image block according to the coding mode identifier;

Or identifying a division mode of the sub-image block in the current image block according to a ratio of a dividing line in the current image block to a distance between a boundary of the current image block or a focus of the boundary line extension line to an end point of the image block;

Or, the manner of dividing the sub-image block in the current image block is identified according to the angle and distance of the dividing line in the current image block.

Further, the analyzing unit 12 can be specifically used for:

If the current image block is divided into two sub-image blocks, the image adjacent to the left side of the current image block The motion vector of the block or sub-image block is the first candidate reference motion vector of the left sub-image block in the current image block, and the motion vector of the adjacent upper right image block or sub-image block of the current image block is the current image. a first candidate reference motion vector of the right sub-image block in the block;

The reference motion vector acquisition module provided by the embodiment of the present invention is a function device for performing the reference motion vector acquisition method provided by the embodiment of the present invention. For the specific implementation process, refer to the embodiment of the reference motion vector acquisition method, and details are not described herein again.

FIG. 9 is a schematic structural diagram of a first embodiment of an encoding apparatus according to an embodiment of the present invention. As shown in FIG. 9, the apparatus includes: a prediction module 1, a transform module 2, a quantization module 3, an entropy encoding module 4, and a reference motion vector acquisition. Module 5;

The reference motion vector acquisition block 5 is configured to obtain a division manner of the sub image block in the current image block, and determine a first candidate reference motion vector in each candidate reference motion vector of the sub image block according to the division manner. And the candidate reference motion vector is a motion vector of an image block or a sub-image block adjacent to the current image block, or a motion vector obtained after the motion vector of the adjacent image block or sub-image block is operated; Each candidate reference motion vector is numbered according to the first candidate reference motion vector; a reference motion vector is selected from each of the numbered candidate reference motion vectors;

a prediction module 1 , configured to acquire a pixel residual between the sub image block and the predicted image block, and obtain a difference between the motion vector of the predicted image block and the reference motion vector;

a transform module 2, configured to separately transform the residual obtained by the prediction module 1;

a quantization module 3, configured to quantize the residuals obtained by the prediction module transformed by the transform module 2, respectively;

The entropy coding module 4 is configured to entropy encode the residual quantized by the quantization module 3, and entropy encode the difference between the motion vector of the predicted image block and the reference motion vector. The encoding apparatus provided by the embodiment of the present invention firstly subtracts the pixel value corresponding to the current image block or the sub-image block and the matched sub-image block by the prediction module 1 to obtain a residual, and then sequentially passes the transformation module 2 and the quantization module 3 to the residual. The difference is transformed and quantized. Finally, the transformed and quantized residual is entropy encoded by the entropy encoding module 4, and the entropy encoded bit stream is written into the encoded code stream.

At the same time, the encoding end of the prediction module 1 also obtains the difference between the motion vector of the predicted image block and the reference motion vector according to the reference motion vector acquired by the reference motion vector obtaining module 5, and then sends the difference between the motion vector and the reference motion vector, and then sends the entropy coding module 4, the entropy coding module. 4 encoding the difference between the motion vector and the reference motion vector, and performing variable length entropy encoding on the corresponding number of the reference motion vector in the candidate reference motion vector table, and then writing the entropy encoded bit stream into the encoded code stream.

For the process of obtaining the reference motion vector, refer to the reference motion vector acquisition method embodiment, and details are not described herein again.

The encoding apparatus provided in this embodiment determines a first candidate reference motion vector of the sub-image block according to a division manner of the sub-image block in the current image block, and numbers each candidate reference motion vector according to the first reference motion vector, so that The first candidate reference motion vector corresponds to a smaller number as much as possible. Since the encoding end uses variable length coding for the number, the coding efficiency of the number can be improved, thereby improving the video coding efficiency.

FIG. 10 is a schematic structural diagram of a first embodiment of a decoding apparatus according to an embodiment of the present invention. As shown in FIG. 10, the apparatus includes: an entropy decoding module 1 ', an inverse quantization module 2, an inverse transform module 3', and a prediction compensation module. 4, and a reference motion vector acquisition module 5';

An entropy decoding module 1 ′, configured to perform entropy decoding on the received encoded code stream, obtain a pixel residual between each sub image block and the predicted image block in the current image block, and obtain a motion vector of the predicted image block and the reference The difference between the motion vectors and the number corresponding to the reference motion vector;

An inverse quantization module 2', configured to inverse quantize the residual;

The inverse transform module 3' is configured to perform inverse transform on the residual inversely quantized by the inverse quantization module; the reference motion vector obtaining module 5' is configured to obtain a partition mode of the sub image block in the current image block; Means determining a first candidate reference among each candidate reference motion vector of the sub-image block a motion vector, where the candidate reference motion vector is a motion vector of an image block or a sub-image block adjacent to the current image block, or a motion obtained by the motion vector of the adjacent image block or sub-image block Vector; numbering each candidate reference motion vector according to the first candidate reference motion vector; selecting a reference motion vector from each of the numbered candidate reference motion vectors.

a prediction compensation module 4', configured to acquire a corresponding reference motion vector according to the number, obtain the motion vector according to a difference between the motion vector and the reference motion vector; and determine the predicted image block according to the motion vector, And acquiring the current image block according to the predicted image block and the residual.

The decoding apparatus provided by the embodiment of the present invention determines a first candidate reference motion vector of the sub-image block according to a division manner of the sub-image block in the current image block, and numbers each candidate reference motion vector according to the first reference motion vector, The smaller number of the first candidate reference motion vector is matched as much as possible. Since the encoding end uses variable length coding for the number, the decoding efficiency of the number can be improved, and the video decoding efficiency is improved.

The encoding and decoding apparatus provided by the embodiment of the present invention can be integrated into digital signal processing

(Digital Signal Processing; abbreviated as: DSP) chip or Field-Programmable Gate Array (FPGA) chip for software development, or through Application Specific Integrated Circuit (ASIC) The way to cure is achieved.

It should be noted that the above embodiments are only for explaining the technical solutions of the present invention, and are not intended to be limiting; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those skilled in the art that: The technical solutions described in the foregoing embodiments are modified, or some of the technical features are equivalently replaced. The modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

Claim

A reference motion vector acquisition method, comprising:

Obtaining a division manner of the sub image block in the current image block;

Determining, in the dividing manner, a first candidate reference motion vector in each candidate reference motion vector of the sub-image block, where the candidate reference motion vector is an motion of an image block or a sub-image block adjacent to the current image block. a vector, or a motion vector obtained by the motion vector of the adjacent image block or the sub-image block;

The method according to claim 1, wherein the numbering each candidate reference motion vector according to the first candidate reference motion vector comprises:

The candidate reference motion vectors are numbered according to a setting rule;

The numbered candidate reference motion vectors are reordered according to the first candidate reference motion vector.

The method according to claim 2, wherein the reordering the numbered candidate reference motion vectors according to the first candidate reference motion vector comprises:

The method according to any one of claims 1 to 3, wherein the obtaining the division manner of the sub-image block in the current image block comprises:

Or identifying a division manner of the sub-image block in the current image block according to a ratio of a dividing line in the current image block to a distance between a boundary of the current image block or a focus of the boundary line extension line to an end point of the image block; Or, the division manner of the sub-image block in the current image block is identified according to the angle and the distance of the dividing line in the current image block.

The method according to any one of claims 1-3, wherein the determining the first candidate reference motion vector of the sub-image block in each candidate reference motion vector according to the dividing manner comprises:

6. A reference motion vector acquisition module, comprising:

And a selecting unit, configured to select a reference motion vector according to each of the candidate reference motion vectors after the numbering.

The reference motion vector acquisition module according to claim 6, wherein the numbering unit comprises:

a numbering subunit, configured to number each candidate reference motion vector according to a setting rule; a reordering subunit, configured to use each of the candidate parameters after the first candidate reference motion vector pair The motion vector is reordered.

The reference motion vector acquisition module according to claim 7, wherein the reordering subunit is specifically configured to:

9. The reference motion vector acquisition module according to claim 6, wherein the identification unit is configured to:

10. The reference motion vector acquisition module according to claim 6, wherein the analysis unit is configured to:

An encoding device, comprising: a prediction module, a transformation module, a quantization module, an entropy coding module, and a reference motion vector acquisition module according to any one of claims 6-10, wherein The reference motion vector acquisition module is configured to obtain a division manner of the sub image block in the current image block, and determine a first candidate reference motion vector in each candidate reference motion vector of the sub image block according to the division manner, a candidate reference motion vector is a motion vector of an image block or a sub-image block adjacent to the current image block, or a motion vector obtained by calculating a motion vector of the adjacent image block or sub-image block; The first candidate reference motion vector numbers each candidate reference motion vector; and selects a reference motion vector from each of the numbered candidate reference motion vectors;

12. A decoding apparatus, comprising: an entropy decoding module, an inverse quantization module, an inverse transform module, a prediction compensation module, and a reference motion vector acquisition module according to any one of claims 6-10; a decoding module, configured to perform entropy decoding on the received encoded code stream, obtain a pixel residual between each sub-image block and the predicted image block in the current image block, and obtain a motion vector of the predicted image block and the reference motion vector The difference and the number corresponding to the reference motion vector;

The inverse transform module is configured to perform inverse transform on a residual that is inversely quantized by the inverse quantization module, where the reference motion vector acquisition module is configured to obtain a partition mode of the sub image block in the current image block; And determining, in each candidate reference motion vector of the sub-image block, a first candidate reference motion vector, where the candidate reference motion vector is a motion vector of an image block or a sub-image block adjacent to the current image block, or a motion vector obtained by calculating a motion vector of an adjacent image block or a sub-image block; and encoding each candidate reference motion vector according to the first candidate reference motion vector; Selecting a reference motion vector from each of the candidate reference motion vectors after the numbering;

The prediction compensation module is configured to acquire a corresponding reference motion vector according to the number, obtain the motion vector according to a difference between the motion vector and the reference motion vector, and determine the predicted image block according to the motion vector, And acquiring the current image block according to the predicted image block and the residual.