CN113473154B - Video encoding method, video decoding method, video encoding device, video decoding device and storage medium - Google Patents

Abstract

The embodiment of the application discloses a video coding method, a video decoding method, a video coding device, a video decoding device and a storage medium, and belongs to the technical field of multimedia. In the embodiment of the application, a vector quantization method may be adopted to quantize coefficients to be encoded in a plurality of transform coefficient matrices to obtain corresponding first indication information, and then a video code stream of a target video is generated according to a plurality of code word indication information including the first indication information. Subsequently, the decoding end can decode the vector quantized coefficient according to the first indication information. Therefore, the embodiment of the application provides a video coding and decoding method based on vector quantization, the application of the vector quantization in video coding and decoding is realized, and compared with scalar quantization and network coding quantization, the quantization performance is improved, and further the performance of the video coding and decoding is improved.

Description

Video encoding, video decoding method, device and storage medium

technical field

The present application relates to the field of multimedia technologies, and in particular to a video encoding and decoding method, device and storage medium.

Background technique

With the development of multimedia technology, video coding technology has been widely valued. When encoding a video, it is necessary to perform a discrete cosine transform on the residual signal of the video, so as to obtain a transform coefficient matrix. Afterwards, the transform coefficients included in the transform coefficient matrix are quantized, and then the video code stream is determined according to the quantization result. At present, quantization techniques mainly include scalar quantization, vector quantization and trellis coding quantization. Among them, scalar quantization and trellis coding quantization have been widely used in video coding and decoding, but vector quantization has not yet been applied in video coding and decoding. Based on this, it is urgent to provide a video encoding and decoding method based on vector quantization technology.

Contents of the invention

Embodiments of the present application provide a video encoding, video decoding method, device, and storage medium, which can obtain multiple transform coefficient matrices corresponding to the target video, quantize the transform coefficients included in the multiple transform coefficient matrices, and obtain Video stream. Described technical scheme is as follows:

In one aspect, a video encoding method is provided, wherein the method comprises:

Acquiring a plurality of transformation coefficient matrices corresponding to the target video, each of the plurality of transformation coefficient matrices includes a plurality of coefficients to be encoded;

Quantizing the coefficients to be encoded in the plurality of transformation coefficient matrices to obtain a plurality of codeword indication information, the plurality of codeword indication information is used to indicate the quantization results obtained when quantizing the coefficients to be encoded, the plurality of The codeword indication information includes first indication information, and the first indication information is used to indicate a codeword obtained when the coefficient to be coded is vector quantized;

Generate a video code stream of the target video according to the multiple codeword indication information.

Optionally, said quantizing the coefficients to be encoded in the plurality of transformation coefficient matrices to obtain a plurality of codeword indication information includes:

Perform vector quantization or mixed quantization on multiple coefficients to be encoded in each of the multiple transform coefficient matrices to obtain at least one codeword information corresponding to each transform coefficient matrix, wherein the mixed quantization includes For vector quantization and quantization other than vector quantization, at least one codeword information corresponding to each transform coefficient matrix includes the first indication information, and the first indication information is used to indicate that all or A codeword obtained during vector quantization of some of the coefficients to be encoded.

Optionally, the quantizing the coefficients to be encoded in the plurality of transformation coefficient matrices includes:

performing vector quantization or mixed quantization on a plurality of coefficients to be encoded in each target transform coefficient matrix that satisfies the vector quantization condition among the plurality of transform coefficient matrices, to obtain at least one codeword indication information corresponding to each target transform coefficient matrix, Wherein, the mixed quantization includes vector quantization and quantization other than vector quantization, and at least one codeword information corresponding to each target transform coefficient matrix includes the first indication information, and the first indication information is used to indicate the Codewords obtained during vector quantization of all or part of the coefficients to be encoded in the corresponding transformation coefficient matrix.

Optionally, the plurality of coefficients to be encoded included in each of the plurality of transform coefficient matrices are all the transform coefficients included in the corresponding transform coefficient matrix.

Optionally, the vector quantization condition includes that the number of rows of the transformation coefficient matrix reaches the first row number threshold, or the number of columns of the transformation coefficient matrix reaches the first column number threshold, or the area of the transformation coefficient matrix reaches the first area threshold at least one of .

Optionally, the plurality of coefficients to be encoded included in each of the plurality of transform coefficient matrices are transform coefficients in a first area of the corresponding transform coefficient matrix, and the first area is obtained by scanning an area based on The coefficients are coded by the SRCC technique to scan the resulting area.

Optionally, the vector quantization condition includes that the number of rows in the first region in the transform coefficient matrix reaches the second row number threshold, or that the column number in the first region in the transform coefficient matrix reaches the second column number threshold, or the transformation The area of the first region in the coefficient matrix reaches at least one of the second area thresholds.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix in the video code stream of the target video corresponds to a vector quantization switch identifier, and the vector quantization switch identifier is used to indicate that in the corresponding at least one codeword information Whether it contains the first indication information obtained through vector quantization.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix in the video code stream of the target video corresponds to the number of vectors or the number of vector quantization times, and the number of vectors or the number of times of vector quantization is used to indicate The corresponding at least one codeword indicates the number of first indication information included in the information.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix in the video code stream of the target video also corresponds to a vector quantization start position, and the vector quantization start position is used to indicate the corresponding at least one codeword The position of the first first indication message in the word indication message.

Optionally, the quantizing the coefficients to be encoded in the plurality of transformation coefficient matrices includes:

If the number of coefficients to be encoded in the first transformation coefficient matrix is N times the codeword dimension, then N input signal vectors are generated according to all coefficients to be encoded in the first transformation coefficient matrix, and N is positive Integer, the dimension of the input signal vector is equal to the dimension of the codeword, the first transformation coefficient matrix is any one of the plurality of transformation coefficient matrices, or the first transformation coefficient matrix is the Transform coefficient matrices satisfying vector quantization conditions among the plurality of transform coefficient matrices;

determining first indication information corresponding to each of the N input signal vectors;

The determined N pieces of first indication information are used as at least one codeword indication information corresponding to the first transformation coefficient matrix.

Optionally, the quantizing the coefficients to be encoded in the plurality of transformation coefficient matrices includes:

If the number of coefficients to be coded in the first transform coefficient matrix is not an integer multiple of the codeword dimension, then M coefficients are generated according to part of the coefficients to be coded in the first transform coefficient matrix and the codeword dimension An input signal vector, the dimension of the input signal vector is equal to the dimension of the code word, the first transformation coefficient matrix is any one of the plurality of transformation coefficient matrices, or the first transformation coefficient matrix is Transform coefficient matrices satisfying vector quantization conditions among the plurality of transform coefficient matrices;

Determine the first indication information corresponding to each input signal vector in the M input signal vectors, and perform non-vector quantization on the remaining coefficients to be encoded in the first transform coefficient matrix to obtain the remaining coefficients to be encoded Corresponding second instruction information;

The determined first indication information and second indication information are used as indication information of multiple codewords corresponding to the first transformation coefficient matrix.

Optionally, said quantizing the coefficients to be coded in multiple transformation coefficient matrices includes:

If the number of coefficients to be coded in the first transformation coefficient matrix is not an integer multiple of the codeword dimension, then according to all the coefficients to be coded in the first transformation coefficient matrix and the codeword dimension, generate R An input signal vector, the dimension of the input signal vector is equal to the dimension of the codeword, and one of the R input signal vectors includes one or more coefficient filling values, and the first transform coefficient The matrix is any one of the plurality of transform coefficient matrices, or the first transform coefficient matrix is a transform coefficient matrix satisfying vector quantization conditions among the plurality of transform coefficient matrices;

determining first indication information corresponding to each input signal vector in the R input signal vectors;

The determined R pieces of first indication information are used as at least one codeword indication information corresponding to the first transform coefficient matrix.

Optionally, the first indication information is a codeword index, codeword index indication information, or indication information of a codeword component included in the codeword.

Optionally, the codeword index is a one-dimensional index value, or the codeword index is a multidimensional index coordinate.

Optionally, the video code stream of the target video includes a vector quantization enable flag, and the vector quantization enable flag is used to indicate whether the video code stream of the target video supports vector quantization.

Optionally, the video code stream of the target video contains at least one vector dimension, and the at least one vector dimension is used to indicate the code indicated by each first indication information in the video code stream of the target video. Dimensions of words.

Optionally, the video code stream of the target video further includes at least one codebook index, and the at least one codebook index is used to indicate the corresponding first indication information in the video code stream of the target video. The codebook used when codewording.

On the other hand, a video decoding method is provided, wherein the method comprises:

Acquire the video code stream of the target video, the video code stream of the target video includes a plurality of codeword indication information, and the multiple codeword indication information is used to indicate the coefficients to be encoded in the plurality of transformation coefficient matrices corresponding to the target video Quantization results obtained when quantization is performed, the plurality of codeword indication information includes first indication information, and the first indication information is used to indicate codewords obtained when vector quantization is performed on coefficients to be encoded, and the plurality of transform coefficients Each transformation coefficient matrix in the matrix includes a plurality of coefficients to be encoded;

Acquire decoded data of the target video according to the plurality of codeword indication information.

Optionally, the instruction information for the plurality of codewords is a plurality of first instruction information, and the obtaining the decoded data of the target video according to the plurality of codeword instruction information includes:

Acquiring a codeword corresponding to each first indication information in the plurality of first indication information;

Generate decoded data of the target video according to the acquired multiple codewords.

Optionally, the generating the decoded data of the target video according to the obtained multiple codewords includes:

Using the multiple codewords as dequantized reconstruction values, or processing the multiple codewords to obtain dequantized reconstruction values;

Generate decoded data of the target video according to the dequantized reconstruction value.

Optionally, the plurality of codeword indication information includes a plurality of first indication information and a plurality of second indication information, and the second indication information is used to indicate quantization results obtained through other quantizations other than vector quantization;

The obtaining the decoded data of the target video according to the plurality of codeword indication information includes:

Acquire a codeword corresponding to each first indication information in the plurality of first indication information, and acquire a quantization result corresponding to each second indication information in the plurality of second indication information;

Generate the decoded data of the target video according to the acquired multiple codewords and the quantization result corresponding to each second indication information.

Optionally, the first indication information is a codeword index, codeword index indication information, or indication information of a codeword component included in the codeword.

Optionally, the video code stream of the target video further includes at least one codebook index, and the at least one codebook index is used to indicate the corresponding first indication information in determining the plurality of codeword indication information The codebook used when codewording.

Optionally, the video code stream of the target video further includes a vector quantization enable flag, and the vector quantization enable flag is used to indicate whether the video code stream of the target video supports vector quantization.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix among the plurality of codeword indication information included in the video code stream of the target video corresponds to a vector quantization switch identifier, and the vector quantization switch identifier is used to indicate Whether the corresponding at least one piece of codeword information includes the first indication information obtained through vector quantization.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix in the plurality of codeword indication information included in the video code stream of the target video corresponds to the number of vectors or the number of times of vector quantization, and the number of vectors or The number of times of vector quantization is used to indicate the number of first indication information included in the corresponding at least one codeword indication information.

Optionally, the at least one codeword indication information corresponding to each transformation coefficient matrix also corresponds to a vector quantization start position, and the vector quantization start position is used to indicate the first and the first number in the corresponding at least one codeword indication information. - indicates the location of the information.

Optionally, the video code stream of the target video contains at least one vector dimension, and the at least one vector dimension is used to indicate the code indicated by each first indication information in the video code stream of the target video. Dimensions of words.

In another aspect, a video encoding device is provided, the device comprising:

An acquisition module, configured to acquire a plurality of transform coefficient matrices corresponding to the target video, where each transform coefficient matrix in the plurality of transform coefficient matrices includes a plurality of coefficients to be encoded;

A quantization module, configured to quantize the coefficients to be encoded in the plurality of transformation coefficient matrices to obtain a plurality of codeword indication information, and the plurality of codeword indication information is used to indicate the quantization results obtained when the coefficients to be encoded are quantized , the plurality of codeword indication information includes first indication information, and the first indication information is used to indicate a codeword obtained when vector quantization is performed on a coefficient to be encoded;

A generating module, configured to generate a video code stream of the target video according to the multiple codeword indication information.

Optionally, the quantization module is mainly used for:

Perform vector quantization or mixed quantization on multiple coefficients to be encoded in each of the multiple transform coefficient matrices to obtain at least one codeword information corresponding to each transform coefficient matrix, wherein the mixed quantization includes For vector quantization and quantization other than vector quantization, at least one codeword information corresponding to each transform coefficient matrix includes the first indication information, and the first indication information is used to indicate that all or A codeword obtained during vector quantization of some of the coefficients to be encoded.

Optionally, the quantization module is also used for:

performing vector quantization or mixed quantization on a plurality of coefficients to be encoded in each target transform coefficient matrix that satisfies the vector quantization condition among the plurality of transform coefficient matrices, to obtain at least one codeword indication information corresponding to each target transform coefficient matrix, Wherein, the mixed quantization includes vector quantization and quantization other than vector quantization, and at least one codeword information corresponding to each target transform coefficient matrix includes the first indication information, and the first indication information is used to indicate the Codewords obtained during vector quantization of all or part of the coefficients to be encoded in the corresponding transformation coefficient matrix.

Optionally, the plurality of coefficients to be encoded included in each of the plurality of transform coefficient matrices are all the transform coefficients included in the corresponding transform coefficient matrix.

Optionally, the vector quantization condition includes that the number of rows of the transformation coefficient matrix reaches the first row number threshold, or the number of columns of the transformation coefficient matrix reaches the first column number threshold, or the area of the transformation coefficient matrix reaches the first area threshold at least one of .

Optionally, the plurality of coefficients to be encoded included in each of the plurality of transform coefficient matrices are transform coefficients in a first area of the corresponding transform coefficient matrix, and the first area is obtained by scanning an area based on The coefficients are coded by the SRCC technique to scan the resulting area.

Optionally, the vector quantization condition includes that the number of rows in the first region in the transform coefficient matrix reaches the second row number threshold, or that the column number in the first region in the transform coefficient matrix reaches the second column number threshold, or the transformation The area of the first region in the coefficient matrix reaches at least one of the second area thresholds.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix in the video code stream of the target video corresponds to a vector quantization switch identifier, and the vector quantization switch identifier is used to indicate that in the corresponding at least one codeword information Whether it contains the first indication information obtained through vector quantization.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix in the video code stream of the target video corresponds to the number of vectors or the number of vector quantization times, and the number of vectors or the number of times of vector quantization is used to indicate The corresponding at least one codeword indicates the number of first indication information included in the information.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix in the video code stream of the target video also corresponds to a vector quantization start position, and the vector quantization start position is used to indicate the corresponding at least one codeword The position of the first first indication message in the word indication message.

Optionally, the quantization module is also used for:

If the number of coefficients to be encoded in the first transformation coefficient matrix is N times the codeword dimension, then N input signal vectors are generated according to all coefficients to be encoded in the first transformation coefficient matrix, and N is positive Integer, the dimension of the input signal vector is equal to the dimension of the codeword, the first transformation coefficient matrix is any one of the plurality of transformation coefficient matrices, or the first transformation coefficient matrix is the Transform coefficient matrices satisfying vector quantization conditions among the plurality of transform coefficient matrices;

determining first indication information corresponding to each of the N input signal vectors;

The determined N pieces of first indication information are used as at least one codeword indication information corresponding to the first transformation coefficient matrix.

Optionally, the quantization module is also used for:

If the number of coefficients to be coded in the first transform coefficient matrix is not an integer multiple of the codeword dimension, then M coefficients are generated according to part of the coefficients to be coded in the first transform coefficient matrix and the codeword dimension An input signal vector, the dimension of the input signal vector is equal to the dimension of the code word, the first transformation coefficient matrix is any one of the plurality of transformation coefficient matrices, or the first transformation coefficient matrix is Transform coefficient matrices satisfying vector quantization conditions among the plurality of transform coefficient matrices;

Determine the first indication information corresponding to each input signal vector in the M input signal vectors, and perform non-vector quantization on the remaining coefficients to be encoded in the first transform coefficient matrix to obtain the remaining coefficients to be encoded Corresponding second instruction information;

The determined first indication information and second indication information are used as indication information of multiple codewords corresponding to the first transformation coefficient matrix.

Optionally, the quantization module is also used for:

If the number of coefficients to be coded in the first transformation coefficient matrix is not an integer multiple of the codeword dimension, then according to all the coefficients to be coded in the first transformation coefficient matrix and the codeword dimension, generate R An input signal vector, the dimension of the input signal vector is equal to the dimension of the codeword, and one of the R input signal vectors includes one or more coefficient filling values, and the first transform coefficient The matrix is any one of the plurality of transform coefficient matrices, or the first transform coefficient matrix is a transform coefficient matrix satisfying vector quantization conditions among the plurality of transform coefficient matrices;

determining first indication information corresponding to each input signal vector in the R input signal vectors;

The determined R pieces of first indication information are used as at least one codeword indication information corresponding to the first transform coefficient matrix.

Optionally, the first indication information is a codeword index, codeword index indication information, or indication information of a codeword component included in the codeword.

Optionally, the codeword index is a one-dimensional index value, or the codeword index is a multidimensional index coordinate.

Optionally, the video code stream of the target video includes a vector quantization enable flag, and the vector quantization enable flag is used to indicate whether the video code stream of the target video supports vector quantization.

Optionally, the video code stream of the target video contains at least one vector dimension, and the at least one vector dimension is used to indicate the code indicated by each first indication information in the video code stream of the target video. Dimensions of words.

Optionally, the video code stream of the target video further includes at least one codebook index, and the at least one codebook index is used to indicate the corresponding first indication information in the video code stream of the target video. The codebook used when codewording.

In another aspect, a video decoding device is provided, the device comprising:

The first acquisition module is used to acquire the video code stream of the target video, the video code stream of the target video includes a plurality of codeword indication information, and the multiple codeword indication information is used to indicate multiple transformations corresponding to the target video Quantization results obtained when the coefficients to be encoded in the coefficient matrix are quantized, the plurality of codeword indication information includes first indication information, and the first indication information is used to indicate the codewords obtained when the coefficients to be encoded are vector quantized , each of the plurality of transform coefficient matrices includes a plurality of coefficients to be encoded;

The second acquiring module is configured to acquire decoded data of the target video according to the multiple codeword indication information.

Optionally, the plurality of codeword indication information is a plurality of first indication information, and the second acquisition module is mainly used for:

Acquiring a codeword corresponding to each first indication information in the plurality of first indication information;

Generate decoded data of the target video according to the acquired multiple codewords.

Optionally, the second acquisition module is mainly used for:

Using the multiple codewords as dequantized reconstruction values, or processing the multiple codewords to obtain dequantized reconstruction values;

Generate decoded data of the target video according to the dequantized reconstruction value.

Optionally, the plurality of codeword indication information includes a plurality of first indication information and a plurality of second indication information, and the second indication information is used to indicate quantization results obtained through other quantizations other than vector quantization; the The second acquisition module is also mainly used for:

Acquire a codeword corresponding to each first indication information in the plurality of first indication information, and acquire a quantization result corresponding to each second indication information in the plurality of second indication information;

Generate the decoded data of the target video according to the acquired multiple codewords and the quantization result corresponding to each second indication information.

Optionally, the first indication information is a codeword index, codeword index indication information, or indication information of a codeword component included in the codeword.

Optionally, the video code stream of the target video further includes at least one codebook index, and the at least one codebook index is used to indicate the corresponding first indication information in determining the plurality of codeword indication information The codebook used when codewording.

Optionally, the video code stream of the target video further includes a vector quantization enable flag, and the vector quantization enable flag is used to indicate whether the video code stream of the target video supports vector quantization.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix among the plurality of codeword indication information included in the video code stream of the target video corresponds to a vector quantization switch identifier, and the vector quantization switch identifier is used to indicate Whether the corresponding at least one piece of codeword information includes the first indication information obtained through vector quantization.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix in the plurality of codeword indication information included in the video code stream of the target video corresponds to the number of vectors or the number of times of vector quantization, and the number of vectors or The number of times of vector quantization is used to indicate the number of first indication information included in the corresponding at least one codeword indication information.

Optionally, the at least one codeword indication information corresponding to each transformation coefficient matrix also corresponds to a vector quantization start position, and the vector quantization start position is used to indicate the first and the first number in the corresponding at least one codeword indication information. - indicates the location of the information.

Optionally, the video code stream of the target video contains at least one vector dimension, and the at least one vector dimension is used to indicate the code indicated by each first indication information in the video code stream of the target video. Dimensions of words.

In another aspect, a video encoding device is provided, the device comprising:

processor;

memory for storing processor-executable instructions;

Wherein, the processor executes the executable instructions in the memory to implement the above video encoding method.

In another aspect, a video decoding device is provided, the device comprising:

processor;

memory for storing processor-executable instructions;

Wherein, the processor executes the executable instructions in the memory to execute the above video decoding method.

In another aspect, a computer-readable storage medium is provided, wherein a computer program is stored in the storage medium, and when the computer program is executed by a computer, the steps of the above-mentioned video encoding method or video decoding method are realized.

In another aspect, there is provided a computer program product including instructions, which, when run on a computer, causes the computer to execute the steps of the above video encoding or video decoding method.

The beneficial effects brought by the technical solutions provided by the embodiments of the present application at least include:

In this embodiment of the present application, the vector quantization method may be used to quantize the coefficients to be encoded in multiple transformation coefficient matrices to obtain the corresponding first indication information, and then indicate according to the multiple codewords containing the first indication information The information generates the video code stream of the target video. Subsequently, the decoding end can decode the vector quantized coefficients according to the first indication information. It can be seen that the embodiment of the present application provides a video encoding and decoding method based on vector quantization, which realizes the application of vector quantization in video encoding and decoding. Compared with scalar quantization and network coding quantization, the quantization performance is improved, thereby improving The performance of video codec.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings that need to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those skilled in the art, other drawings can also be obtained based on these drawings without creative effort.

FIG. 1 is a system architecture diagram related to a video codec provided by an embodiment of the present application;

FIG. 2 is a flow chart of a video encoding method provided by an embodiment of the present application;

Fig. 3 is a schematic diagram of a scanning area provided by an embodiment of the present application;

FIG. 4 is a flow chart of a video decoding method provided by an embodiment of the present application;

FIG. 5 is a structural diagram of a video encoding device provided by an embodiment of the present application;

FIG. 6 is a structural diagram of a video decoding device provided by an embodiment of the present application;

Fig. 7 is a schematic structural diagram of a server for video encoding or video decoding provided by an embodiment of the present application.

Detailed ways

In order to make the purpose, technical solution and advantages of the present application clearer, the implementation manners of the present application will be further described in detail below in conjunction with the accompanying drawings.

Before explaining the embodiment of the present application in detail, the system architecture involved in the embodiment of the present application is firstly introduced.

FIG. 1 is a system architecture diagram involved in a video encoding and video decoding method provided by an embodiment of the present application. As shown in FIG. 1 , the system video encoding device 101 and video decoding device 102 . Wherein, the video encoding device 101 and the video decoding device 102 can communicate.

The video coding device 101 is used to obtain a video signal, and obtain a plurality of transformation coefficient matrices according to the video signal, and then quantize and encode the coefficients to be encoded in the plurality of transformation coefficient matrices through the method provided in the embodiment of the present application, thereby obtaining a video code flow. Send the generated video code stream to the video decoding device 102 .

Wherein, the video coding device 101 may include a prediction module 1011 , a transformation module 1012 , a quantization module 1013 and an entropy coding module 1014 .

Wherein, the prediction module 1011 includes an intra-frame prediction unit and an inter-frame prediction unit, wherein the intra-frame prediction unit can use the reconstructed pixels of the coded blocks around the current coded block to predict the pixels of the current coded block, so as to obtain the predicted block; frame The inter-prediction unit can use reconstructed pixels of adjacent encoded image frames in time domain to predict the pixels of the current unencoded image frame, so as to obtain the prediction block. The prediction module 1011 may obtain a prediction block through any one of the above prediction units, and then calculate a residual signal according to the current coding block and the prediction block.

The transform module 1012 maps the residual signal of the image to the transform domain to obtain a transform coefficient matrix, so that the energy of the image in the transform domain is more concentrated, and the frequency domain correlation of the signal is removed.

The quantization module 1013 maps the coefficients to be encoded in the transformation coefficient matrix obtained by the transformation module 1012 into codewords. This mapping is a "many-to-one" process, which is irreversible and will cause signal loss. But the advantage is that it can greatly reduce the value range of the signal, so that a small number of symbols can be used to give a better approximation of the original signal and improve the compression rate.

The entropy coding module 1014 is a lossless coding method based on the principle of information entropy, which converts the codeword index corresponding to the codeword mapped by the quantization module 1013 or other information used to indicate the codeword into a binary code stream, thereby generating a video code flow.

The video decoding device 102 receives the video code stream sent by the video coding device 101, and decodes the video code stream according to the codeword indication information included in the video code stream, so as to obtain decoded data.

Wherein, the video decoding device 102 may include an entropy decoding module 1021 , an inverse quantization module 1022 , an inverse transformation module 1023 , a reconstruction module 1024 and a filtering module 1025 .

Wherein, the entropy decoding module 1021 is used to receive the video code stream sent by the video coding device 101, since the video code stream is a binary code stream obtained by the entropy coding module 1014 by converting the codeword indication information, the entropy decoding module 1021 can Entropy decoding is performed on the video code stream to obtain codeword indication information included in the video code stream.

The inverse quantization module 1022 is configured to inverse quantize the codeword indication information obtained by the entropy decoding module 1021 to obtain an inverse quantization reconstruction value.

The inverse transformation module 1023 is configured to perform DCT inverse transformation on the inverse quantization reconstruction value obtained by the inverse quantization module 1022, so as to obtain a plurality of transformation coefficient matrices.

The reconstruction module 1024 is configured to generate a reconstructed image according to the plurality of transformation coefficient matrices obtained by the inverse transformation module 1023 .

The filtering module 1025 is used to perform image enhancement on the reconstructed image obtained by the reconstruction module 1024, so that the reconstructed image is closer to the original image, and at the same time reduce the influence of block effect and ringing effect, and improve the quality of the reconstructed image.

It should be noted that, in this embodiment of the application, the video encoding device 101 may be a terminal device, for example, a smart phone, a tablet computer, a notebook computer, or a desktop computer, and the video encoding device 101 may also be a server or a server cluster. The video decoding device 102 may be a terminal device, for example, a terminal device such as a tablet computer or a desktop computer, and the video decoding device 102 may also be a server or a server cluster, which is not limited in this embodiment of the present application.

Next, the video coding method provided by the embodiment of the present application will be introduced.

FIG. 2 is a flow chart of a video coding method provided by an embodiment of the present application. The method can be applied to the video encoding device in the aforementioned system architecture, as shown in Figure 2, the method includes the following steps:

Step 201: Obtain a plurality of transformation coefficient matrices corresponding to the target video, and each transformation coefficient matrix in the plurality of transformation coefficient matrices includes a plurality of coefficients to be encoded.

In the embodiment of the present application, DCT (Discrete Cosine Transform, Discrete Cosine Transform) transformation is performed on the residual signal of the target video. During the transformation process, DCT transformation is performed on each TB (Transform Block, transformation block) of each frame of video image in the target video to obtain a transformation coefficient matrix corresponding to each TB. In this way, one or more transformation coefficient matrices can be obtained for one frame of video image.

In the embodiment of the present application, each of the plurality of transform coefficient matrices includes a plurality of coefficients to be encoded.

Wherein, in a possible implementation manner, the multiple coefficients to be encoded included in each of the multiple transform coefficient matrices may be all transform coefficients included in the corresponding transform coefficient matrix.

Optionally, in another possible implementation manner, the plurality of coefficients to be encoded included in each transform coefficient matrix may be the transform coefficients in the first area of the corresponding transform coefficient matrix, where the first area is obtained by The coefficients of the scanned area encode the area scanned by the SRCC technique.

That is, in the embodiment of the present application, the video encoding device can determine the abscissa SRx of the rightmost non-zero coefficient and the ordinate SRy of the bottom non-zero coefficient in the transformation coefficient matrix of the TB through the SRCC technology, using (SRx , SRy) to determine the scanning area to be scanned in the transform coefficient matrix, that is, the first area, as shown in FIG. 3 , and then encode the transform coefficients in the first area determined by (SRx, SRy).

Step 202: Quantize the coefficients to be encoded in multiple transformation coefficient matrices to obtain multiple codeword indication information, the multiple codeword indication information is used to indicate the quantization results obtained when the coefficients to be encoded are quantized, and the multiple codeword indication information The information includes first indication information, and the first indication information is used to indicate a codeword obtained when vector quantization is performed on the coefficient to be encoded.

It can be known from the foregoing step 201 that each transform coefficient matrix includes multiple coefficients to be encoded, based on this, the video encoding device may quantize the multiple coefficients to be encoded in each transform coefficient matrix.

Exemplarily, in a possible implementation manner, the video coding device performs vector quantization or mixed quantization on multiple coefficients to be encoded in each of the multiple transform coefficient matrices to obtain the corresponding At least one piece of codeword information, wherein the hybrid quantization includes vector quantization and quantization other than vector quantization, and the at least one piece of codeword information corresponding to each transform coefficient matrix includes first indication information, which is used to indicate A codeword obtained during vector quantization of all or part of the coefficients to be encoded in the corresponding transformation coefficient matrix.

Wherein, taking the first transform coefficient matrix among the plurality of transform coefficient matrices as an example, the video encoding device may determine whether the number of coefficients to be encoded in the first transform coefficient matrix is N times the dimension of the code word, if the first transform The number of coefficients to be coded in the coefficient matrix is N times the codeword dimension, then generate N input signal vectors according to all coefficients to be coded in the first transformation coefficient matrix, N is a positive integer, and the dimension of the input signal vector Equal to the codeword dimension, determine the first indication information corresponding to each input signal vector in the N input signal vectors; use the determined N first indication information as at least one codeword indication information corresponding to the first transform coefficient matrix.

It should be noted that the dimension of the codeword refers to the dimension of the codeword included in the preset reference codebook to be used when performing vector quantization. Wherein, the reference codebook includes a plurality of codewords, and each codeword is a vector of the dimensionality of the codeword. For example, the codeword dimension can be 8, 16, etc. Moreover, in the embodiment of the present application, each codeword may correspond to one codeword indication information.

If the number of coefficients to be encoded contained in the first transformation coefficient matrix is an integer multiple of the codeword dimension, it means that the coefficients to be encoded contained in the first transformation coefficient matrix can just form N vectors of the codeword dimension, Based on this, the video coding device can generate N input signal vectors according to the codeword dimension.

For example, assuming that the codeword dimension is 8, and the number of coefficients to be encoded contained in the first transform coefficient matrix is 24, then according to the codeword dimension, the 24 coefficients to be encoded in the first transform coefficient matrix can form 3 An 8-dimensional input signal vector.

After obtaining the N input signal vectors, the video encoding device can search the codewords that match each input signal vector from the codewords contained in the reference codebook, and then obtain For the indication information of the corresponding codeword, the obtained indication information is used as the first indication information corresponding to the corresponding input signal vector. Afterwards, the video coding device may use the first indication information corresponding to each input signal vector as the multiple codeword indication information corresponding to the first transformation coefficient matrix.

Wherein, the first indication information may be a codeword index, codeword index indication information, or indication information of a codeword component included in the codeword.

Wherein, when the first indication information is a codeword index, the codeword index may be a one-dimensional index value, or may also be a multi-dimensional index coordinate. For example, for a cluster type codebook, the codeword index of a codeword is a one-dimensional index value, that is, a specific index value can correspond to a codeword; for a lattice type codebook, a codeword corresponds to a Multi-dimensional index coordinates, for example, index coordinates (i, j) correspond to a codeword.

When the first indication information is the codeword index indication information, the codeword index indication information may be information related to the codeword index, that is, the codeword index indication information can indicate a codeword index. For example, a certain codebook contains 31 codewords, divide these 31 codewords into 5 groups, and number each group, the first group is numbered 0, the second group is numbered 1, the third group is numbered 2, The fourth group is numbered 3, and the fifth group is numbered 4; meanwhile, the codewords in the first group are 1, the codewords in the second group are 2 and 3, and the codewords in the third group are 4 to 7, The codewords in the fourth group are 8 to 15, and the codewords in the fifth group are 16 to 31. According to the position of the codeword in each group, the group offset of the codeword in each group can be obtained, and then, according to the group number and the group offset of each codeword, the corresponding codeword can be obtained. Codeword index indication information corresponding to the codeword index, where the codeword index indication information includes the group number of the group where the codeword is located and the offset within the group.

When the first indication information is indication information of codeword components included in the codeword, the indication information of codeword components may be information obtained by directly encoding the codeword components included in the codeword. For example, in some specially constructed codebooks, for some artificially designed codewords, each codeword component in a codeword may be an integer. In this case, each codeword component in the codeword can be directly Encoding to obtain the indication information of the codeword component.

Optionally, if the number of coefficients to be coded in the first transformation coefficient matrix is not an integer multiple of the dimension of the codeword, then vector quantization can be performed on all the coefficients to be coded in the first transformation coefficient matrix, or the second Coefficients to be coded in a transformation coefficient matrix are mixed and quantized.

Wherein, if vector quantization is performed on all the coefficients to be coded in the first transform coefficient matrix, R input signal vectors are generated according to all coefficients to be coded and codeword dimensions in the first transform coefficient matrix, and the dimension of the input signal vector The number is equal to the codeword dimension. Wherein, one of the R input signal vectors includes one or more coefficient filling values. Afterwards, the first indication information corresponding to each input signal vector in the R input signal vectors is determined. The determined R pieces of first indication information are used as at least one codeword indication information corresponding to the first transform coefficient matrix.

Wherein, if the number of coefficients to be encoded contained in the first transformation coefficient matrix is not an integer multiple of the dimension of the codeword, it means that the coefficients to be encoded in the first transformation coefficient matrix cannot form an integer number of vectors of the dimension of the codeword . In this case, if all vector quantization is to be performed on the plurality of coefficients to be encoded, the video encoding device may first calculate the integer r and the remainder t obtained by dividing the number of coefficients to be encoded by the dimension of the codeword, and then , the video encoding device can generate R input signal vectors according to a plurality of coefficients to be encoded contained in the first transform coefficient matrix, at this time, R=r+1, and the last input signal vector can include (codeword dimension- t) coefficient filling values.

For example, the first transformation coefficient matrix contains 20 coefficients to be coded, and the dimension of the code word is 8, calculate 20 divided by 8 to get the integer 2 and the remainder 4, then three 8-dimensional The input signal vector of , wherein, the last input signal vector among the 3 input signal vectors will contain 4 coefficient padding values.

After obtaining the input signal vector, for each input signal vector, the video encoding device can refer to the above introduction to determine the first indication information corresponding to each input signal vector, and use the determined first indication information corresponding to each input signal vector The indication information is used as at least one codeword indication information corresponding to the first transformation coefficient matrix.

Optionally, when the number of coefficients to be encoded in the first transformation coefficient matrix is not an integer multiple of the codeword dimension, if mixed quantization is performed on the coefficients to be encoded in the first variable coefficient matrix, the video encoding The device may generate M input signal vectors according to some of the coefficients to be encoded in the first transformation coefficient matrix and the dimension of the codeword, where the dimension of the input signal vector is equal to the dimension of the codeword. Afterwards, the first indication information corresponding to each input signal vector among the M input signal vectors is determined, and other quantization other than vector quantization is performed on the remaining coefficients to be encoded in the first transformation coefficient matrix, so as to obtain the remaining coefficients to be encoded corresponding to The second instruction message for .

Wherein, the video coding device may firstly calculate the integer r and the remainder t obtained by dividing the number of coefficients to be coded by the codeword dimension, and then the video coding device may generate M input signal vectors of the codeword dimension, at this time, M=r. For the remaining t coefficients to be encoded, other quantization other than vector quantization may be used to quantize them, so as to obtain the second indication information corresponding to each coefficient to be encoded. In this way, vector quantization is performed on some coefficients to be encoded among the plurality of coefficients to be encoded contained in the first transformation coefficient matrix, corresponding first indication information is obtained, and other quantization other than vector quantization is performed on some coefficients to be encoded, The corresponding second indication information is obtained. On this basis, the video coding device may use the obtained first indication information and second indication information as the multiple codeword indication information corresponding to the first transform coefficient matrix.

For example, the coefficients to be encoded in the first transformation coefficient matrix are 20 coefficients, and the dimension of the vector is 8, and the number of vectors is 3. In this case, the 20 coefficients to be encoded in the first transformation coefficient matrix The first 16 coefficients to be encoded generate two 8-dimensional input signal vectors, obtain the first indication information corresponding to each input signal vector, and perform non-vector quantization on the remaining 4 coefficients to be encoded to obtain the corresponding Second instruction message. Optionally, other quantization other than vector quantization may be performed on the first 4 coefficients to be coded of the 20 coefficients to be coded in the first transformation coefficient matrix to obtain the corresponding second indication information, and the remaining 16 coefficients to be coded to form 2 An 8-dimensional input signal vector. Wherein, each signal vector in the two input signal vectors corresponds to the first indication information. Afterwards, the determined two pieces of first indication information and second indication information are used as multiple codeword indication information corresponding to the first transformation coefficient matrix.

It should be noted that, in the process of mixed quantization described above, for the remaining coefficients to be encoded after vector quantization, the non-vector quantization performed by the video encoding device may be index quantity quantization, trellis coding quantization or other quantization. One, which is not limited in this embodiment of the present application.

Optionally, in another possible implementation manner, the video encoding device may perform vector quantization or mixed quantization on multiple coefficients to be encoded in each target transform coefficient matrix that satisfies the vector quantization condition among the multiple transform coefficient matrices, At least one codeword indication information corresponding to each target transformation coefficient matrix is obtained. Wherein, mixed quantization includes vector quantization and quantization other than vector quantization, and at least one codeword information corresponding to each target transform coefficient matrix includes first indication information, and the first indication information is used to indicate the A codeword obtained during vector quantization of all or part of the coefficients to be coded.

It should be noted that, as can be seen from the introduction in the aforementioned step 201, the plurality of coefficients to be encoded included in each of the plurality of transform coefficient matrices may be all the transform coefficients included in the corresponding transform coefficient matrix, or may be Transform coefficients in the first region of the corresponding transform coefficient matrix. Wherein, when the plurality of coefficients to be encoded included in each of the plurality of transform coefficient matrices is all the transform coefficients included in the corresponding transform coefficient matrix, the vector quantization condition may include that the number of rows of the transform coefficient matrix reaches the first At least one of the row number threshold, or the column number of the transformation coefficient matrix reaching the first column number threshold, or the area of the transformation coefficient matrix reaching the first area threshold. That is to say, for any transform coefficient matrix, if the transform coefficient matrix satisfies one or more of the above conditions, vector quantization or mixed quantization can be performed on all the transform coefficients in the corresponding transform coefficient matrix. Wherein, the area of the transformation coefficient matrix refers to the product of the number of rows and the number of columns of the transformation coefficient matrix.

Optionally, when the multiple coefficients to be encoded included in each of the multiple transform coefficient matrices are transform coefficients in the first region of the corresponding transform coefficient matrix, the vector quantization condition may include the first region The number of rows in the transform coefficient matrix reaches the second row number threshold, or the column number of the first region in the transform coefficient matrix reaches the second column number threshold, or the area of the first region in the transform coefficient matrix reaches at least one of the second area thresholds kind. That is to say, for any transformation coefficient matrix, if the first region in the transformation coefficient matrix satisfies one or more of the above conditions, the transformation coefficients contained in the first region in the corresponding transformation coefficient matrix can be Vector quantization or hybrid quantization. Wherein, the area of the first region refers to the product of the number of rows and the number of columns of the first region.

In the embodiment of the present application, the video encoding device may acquire a target transform coefficient matrix satisfying the vector quantization condition from multiple transform coefficient matrices by using the vector quantization condition introduced above. For each target transform coefficient matrix, the video encoding device may perform vector quantization or mixed quantization on the coefficients to be encoded contained in each target transform coefficient matrix by referring to the aforementioned quantization method for the coefficients to be encoded in the first transform coefficient matrix. For variable coefficient matrices that do not meet the vector quantization conditions, the video coding device can quantize the coefficients to be encoded in other quantization methods except vector quantization and mixed quantization, for example, scalar quantization or trellis coding quantization. Quantification, which is not limited in this embodiment of the present application. In this way, the at least one codeword indication information corresponding to the target transformation coefficient matrix that satisfies the vector quantization condition may be the first indication information, or may include the first indication information and the second indication obtained by using other quantization methods other than vector quantization information. The at least one codeword indication information corresponding to the transformation coefficient matrix that does not meet the vector quantization condition will be the second indication information obtained by other quantization methods other than vector quantization, that is, the first indication information will not be included.

The above are only two possible implementations of the vector quantization conditions given in the embodiment of the present application. Optionally, the vector quantization conditions may also be conditions determined according to the block division information of the TB corresponding to the transform coefficient matrix, or according to specific The conditions for setting the quantization parameters corresponding to the quantization mode, or other conditions, are not limited in this embodiment of the present application.

Step 203: Generate a video code stream of the target video according to multiple codeword indication information.

In the embodiment of the present application, after quantizing the coefficients to be encoded in the multiple transformation coefficient matrices corresponding to the target video, and obtaining multiple codeword indication information, the video encoding device can generate the target video according to the multiple codeword indication information. Video stream.

Exemplarily, the video coding device may perform binary coding on at least one codeword indication information corresponding to each transform coefficient matrix according to the sequence of TBs corresponding to each transform coefficient matrix, so as to obtain the video code stream of the target video.

Optionally, it can be seen from the introduction in the foregoing step 202 that the coefficients to be encoded in each transformation coefficient matrix may all be vector quantized, or may be partially vector quantized, or not vector quantized at all. In this way, at least one codeword indication information corresponding to each transformation coefficient matrix may be the first indication information, and may also include the first indication information and indication information corresponding to other quantization methods such as the second indication information, or, It is also possible that the first indication information is not included. Based on this, in the embodiment of the present application, after the video encoding device performs binary coding on at least one codeword indication information corresponding to each transformation coefficient matrix, it may also Add the corresponding vector quantization switch identifier to obtain the video code stream of the target video. Wherein, the vector quantization switch identifier may be used to identify whether the corresponding at least one codeword indication information includes the first indication information obtained through vector quantization.

Optionally, after the video encoding device performs binary coding on at least one codeword indication information corresponding to each transformation coefficient matrix, it may also add the corresponding vector number before the at least one codeword indication information corresponding to each transformation coefficient matrix or the number of vector quantizations to generate a video code stream of the target video, wherein the number of vectors or the number of vector quantizations is used to indicate the number of first indication information contained in the corresponding at least one codeword indication information.

For example, when the coefficients to be encoded in a transformation coefficient matrix form 4 input signal vectors, 4 codewords are obtained according to the 4 input signal vectors, and 4 first indication information are obtained according to the 4 codewords, At this time, the number of vectors or the number of times of vector quantization is 4.

Optionally, a corresponding vector quantization start position may also be added before at least one codeword indication information corresponding to each transform coefficient matrix, and the vector quantization start position is used to indicate the first of the corresponding at least one codeword indication information The location of the first indication message. In this way, when at least one codeword indication information corresponding to a certain transformation coefficient matrix contains both the first indication information and other indication information, the vector quantization start position can be used to know from which position the vector quantization is obtained. The first indication of the . On this basis, if there is also a corresponding number of vectors, it can be determined according to the vector quantization start position and the number of vector quantizations, which of the at least one codeword indication information corresponding to the transformation coefficient matrix is The first indication information is used to facilitate subsequent decoding.

Optionally, the video code stream of the target video may further include a vector quantization enable flag, and the vector quantization enable flag is used to indicate whether the video code stream of the target video supports vector quantization. That is, if vector quantization or mixed quantization is supported in the process of quantizing the coefficients to be encoded in multiple transformation coefficient matrices corresponding to the target video, multiple codewords included in the video code stream of the target video Add the enable flag of vector quantization before the instruction information. In this way, the subsequent decoding end can know whether the video code stream of the target video supports vector quantization through the enabling flag.

Optionally, the video code stream of the target video may also contain at least one vector dimension, at least one vector dimension is used to indicate the dimension of the codeword indicated by each first indication information in the video code stream of the target video .

Wherein, if the codeword dimensions used when performing vector quantization on multiple coefficients to be encoded in each transformation coefficient matrix are the same, then the video code stream of the target video will include a vector dimension, where In this case, the vector dimension can be added before the multiple codeword indication information contained in the video code stream of the target video.

If the dimensions of codewords used in vector quantization of multiple coefficients to be coded in different transform coefficient matrices are the same or different, and the code word dimension used in vector quantization of coefficients to be coded in the same transform coefficient matrix If the word dimensions are the same, the video code stream of the target video may include multiple vector dimensions. At this time, a corresponding vector dimension may be added before at least one codeword indication information corresponding to each transformation coefficient matrix.

For example, the target video corresponds to 10 transformation coefficient matrices, among which, the coefficients to be encoded in 5 transformation coefficient matrices are vector quantized using 8-dimensional codewords, and the coefficients to be encoded in the other 5 transformation coefficient matrices are vector quantized using Vector quantization of 16-dimensional codewords. In this case, the video code stream of the target video will include two kinds of vector dimensions, which are 8 and 16 respectively, and the corresponding vector can be added before at least one codeword indicating information corresponding to each transform coefficient matrix dimension.

If the dimensionality of codewords used when performing vector quantization on multiple coefficients to be encoded in the transformation coefficient matrix may also be different, the video code stream of the target video may include multiple vector dimensions. In this case, the first indication information included in the at least one codeword indication information corresponding to each transformation coefficient matrix may be bound with the corresponding vector dimension.

For example, some of the coefficients to be encoded in a transformation coefficient matrix are vector quantized by using 8-dimensional codewords, and some of the coefficients to be encoded are vector quantized by using 16-dimensional codewords. The first indication information obtained through codeword quantization may be bound to a vector dimension of 8, and the first indication information obtained through 16-dimensional codeword quantization may be bound to a vector dimension of 16.

Optionally, the video code stream of the target video may also include at least one codebook index, at least one codebook index is used to indicate the codeword used when determining the codeword corresponding to each first indication information in the video code stream of the target video codebook.

Wherein, when the same codebook is used to perform vector quantization on multiple coefficients to be encoded in each of the multiple transform coefficient matrices, the video code stream of the target video will include a codebook index. At this time, the codebook index can be added before the multiple codeword indication information contained in the video code stream of the target video, so that the subsequent decoder can obtain the corresponding codebook according to the code book index to decode the video code stream of the target video .

When different codebooks are used to vector quantize the coefficients to be encoded in multiple transform coefficient matrices, and the codebooks used by the coefficients to be encoded in the same transform coefficient matrix are the same, the video code stream of the target video will include Multiple codebook indexes. In this case, a corresponding codebook index may be added before at least one codeword indication information corresponding to each transformation coefficient matrix to indicate which codebook quantization is used to obtain each first indication information in the corresponding transformation coefficient matrix .

When vector quantization is performed on multiple coefficients to be encoded included in the transformation coefficient matrix using different codebooks, the video code stream of the target video will also include multiple codebook indexes. In this case, the first indication information included in at least one codeword indication information corresponding to each transformation coefficient matrix can be bound to the corresponding codebook index to indicate which codebook the corresponding first indication information is obtained by quantization .

It should be noted that the video code stream of the target video may include the vector quantization switch identifier, the vector quantization enable identifier, the number of vectors or the number of times of vector quantization, the starting position of vector quantization, the number of vector dimensions, and the codebook index. One or more, which is not limited in this embodiment of the present application.

In the embodiment of the present application, the vector quantization method is used to quantize the coefficients to be encoded in multiple transformation coefficient matrices to obtain the corresponding first indication information, and then indicate the information according to the multiple codewords containing the first indication information Generate the video code stream of the target video. Subsequently, the decoding end can decode the vector quantized coefficients according to the first indication information. It can be seen that the embodiment of the present application provides a video encoding and decoding method based on vector quantization, which realizes the application of vector quantization in video encoding and decoding. Compared with scalar quantization and network coding quantization, the quantization performance is improved, thereby improving The performance of video codec.

After the target video is encoded by the above method to obtain the video code stream of the target video, the video encoding device may send the video code stream of the target video to the video decoding device. next. The video decoding device may decode the video code stream of the target video by using the video decoding method provided in the embodiment of the present application.

FIG. 4 is a flowchart of a video decoding method provided by an embodiment of the present application. The method can be applied to the video decoding device in the system architecture shown in Figure 1, as shown in Figure 4, the method includes the following steps:

Step 401: Obtain the video code stream of the target video. The video code stream of the target video includes a plurality of codeword indication information, and the multiple codeword indication information includes first indication information, and the first indication information is used to indicate the encoding coefficient to be coded. The codeword obtained during vector quantization.

The video decoding device receives the video code stream of the target video sent by the video coding device, or the video decoding device may also acquire the video code stream of the target video sent to the storage device by the video coding device from the storage device.

Wherein, the video code stream of the target video includes multiple codeword indication information, and the multiple codeword indication information includes first indication information. That is, the video code stream of the target video is obtained by vector quantizing part or all of the coefficients to be encoded included in the plurality of transform coefficient matrices corresponding to the target video. Wherein, for possible implementation manners of the multiple codeword indication information, reference may be made to the description in step 202 in the foregoing embodiments, and the embodiments of the present application are not repeated here.

Optionally, the video code stream of the target video may also include vector quantization switch identifier, vector quantization enable identifier, vector number or vector quantization times, vector quantization start position, vector dimension and codebook index For one or more details, refer to the description in step 203 in the foregoing embodiments for details, and details are not repeated in this embodiment of the present application.

Step 402: Obtain decoded data of the target video according to multiple codeword indication information.

After the video decoding device acquires the video code stream of the target video, it can sequentially dequantize each code word indication information starting from the first code word indication information of the video code stream, so as to obtain the quantization corresponding to each code word indication information As a result, decoded data of the target video is further generated based on the obtained quantization result. The quantization result corresponding to the codeword indication information may be the codeword corresponding to the first indication information, or the quantization result corresponding to the second indication information.

It can be seen from the introduction in the foregoing step 202 that the plurality of codeword indication information of the video code stream of the target video may all be the first indication information, or may include the first indication information and other indication information. Based on this, in the embodiment of the present application, the decoded data of the target video may be obtained through the following two implementation manners.

In the first implementation manner, when the plurality of codeword indication information are all the first indication information, the video decoding device may sequentially acquire the codeword corresponding to each first indication information in the plurality of first indication information; multiple codewords to generate the decoded data of the target video.

Exemplarily, the video decoding device first acquires a codebook for dequantizing multiple first indication information, and then dequantizes each first indication information according to the acquired codebook, so as to obtain the corresponding codeword.

In the first possible situation, the plurality of first indication information are quantized by using the same codebook. In this way, the video encoding device and the video decoding device can pre-agreed on the target codebook to be used. At this time, the video decoding device obtains target codebook. Afterwards, the video decoding device may acquire a codeword corresponding to each piece of first indication information from the target codebook.

In the second possible situation, the multiple first indications are quantized using different codebooks, and at least one of the multiple codeword indications corresponding to the same transformation coefficient matrix will correspond to the same codebook index. At this time, when the video decoding device inversely quantizes each first indication information in turn, when the first codebook index is detected, it obtains a target codebook according to the codebook index, and uses the target codebook to The first indication information after the codebook index is dequantized until the next codebook index is detected, and then the target codebook corresponding to the next codebook index is obtained, and then the next codebook is processed according to the latest acquired target codebook The first indication information after the index is dequantized, and so on.

In a third possible situation, the plurality of first indication information is quantized by using different codebooks, where each first indication information is bound with a codebook index used to indicate its corresponding codebook. In this case, the video decoding device obtains the codebook corresponding to the corresponding first indication information according to the codebook index bound with each first indication information and used to indicate the codebook corresponding to itself. Afterwards, the video decoding device may acquire the codeword corresponding to the corresponding first indication information from the codebook corresponding to each first indication information.

It should be noted that, when inverse quantization is performed according to the obtained first indication information corresponding to the target codebook, the video decoding device may acquire the code corresponding to the first indication information from the corresponding target codebook according to the first indication information. Character.

Wherein, when the first indication information is a codeword index, and when the codeword index is a one-dimensional index value, since one index value can correspond to one codeword, the video decoding device can store the information between the index value and the codeword In this way, the video decoding device can obtain the codeword corresponding to each codeword index by looking up a table, so as to obtain the codeword corresponding to each first indication information.

When the first indication information is a codeword index, and the codeword index is a multidimensional index coordinate, the video decoding device may determine a codeword corresponding to the codeword index through a correlation formula according to the multidimensional index coordinate.

When the first indication information is codeword index indication information, the video decoding device may first acquire a corresponding codeword index according to the codeword index indication information, and then acquire a corresponding codeword according to the codeword index.

Exemplarily, still taking the aforementioned codeword index indication information including the codeword group number and offset as an example, the video decoding device may use the following codeword index indication information according to the group number and group offset included in the The above formula is used to calculate the codeword index.

Codeword index=(1<<group number)+offset within the group-1

After obtaining the codeword index, the video decoding device may obtain the corresponding codeword from the codeword index-to-codeword mapping relationship table according to the codeword index.

When the first indication information is indication information of codeword components included in the codeword, the video decoding device may acquire the codeword components through the indication information of the codeword components, and then form the acquired codeword components into a codeword.

For example, as described in the foregoing embodiments, the first indication information is the indication information of codeword components obtained by encoding each codeword component of the codeword. In this case, the video decoding device can directly indicate each codeword component The information is decoded to obtain corresponding codeword components, and the obtained codeword components are composed into codewords.

After obtaining the codewords corresponding to each first indication information, the video decoding device may generate decoded data of the target video according to the multiple codewords.

In an implementation manner, the multiple codewords are directly used as dequantized reconstruction values, and the video decoding device generates decoded data of the target video according to the dequantized reconstruction values of the multiple codewords.

Optionally, in a possible implementation manner, multiple codewords are respectively calculated through correlation formulas to generate inverse quantization reconstruction values corresponding to the multiple codewords. For example, the codeword can be multiplied by a number, or a number can be added, or a shift operation can be performed to generate an inverse quantization reconstruction value corresponding to the codeword. Afterwards, the video decoding device generates decoded data of the target video according to the dequantized reconstruction value obtained after calculating a plurality of codewords.

Optionally, in the second implementation manner, when the multiple codeword indication information includes the first indication information and the second indication information, the video decoding device may obtain the corresponding first indication information among the multiple first indication information. codeword, and obtain a quantization result corresponding to each second indication information in a plurality of second indication information; generate the decoded data of the target video according to the obtained multiple codewords and the quantization result corresponding to each second indication information .

Wherein, in this implementation manner, at least one codeword indication information corresponding to each transformation coefficient matrix among the plurality of codeword indication information included in the video code stream of the target video may correspond to a vector quantization switch identifier, and the vector quantization switch The identifier is used to indicate whether the corresponding at least one codeword information contains the first indication information obtained through vector quantization.

Based on this, in the first possible situation, at least one codeword indication information corresponding to each transformation coefficient matrix is the first indication information, so that the vector quantization switch identification corresponding to the at least one codeword indication information will be able to identify Whether the at least one codeword indication information is all first indication information or all is second indication information. In this case, when the video decoding device dequantizes a plurality of codeword indication information, when it detects that the vector quantization switch flag is turned on, it can know that the codeword indication information after the vector quantization switch flag is the first 1. Instructions. Afterwards, the video decoding device may refer to the method introduced in the first implementation manner to obtain a codebook for inverse quantization of each first indication information marked by the vector quantization switch, and obtain each first indication information from the obtained codebook A codeword corresponding to the indication information, until the next vector quantization switch flag is detected, if the next vector quantization switch flag is still on, you can continue to refer to the above operations for processing, if the next vector quantization switch flag is off , then the video decoding device can learn that the codeword indication information after the next vector quantization switch identification is the second indication information, so that the video decoding device can sequentially perform the first code word after the next vector quantization switch identification based on other quantization methods other than vector quantization The second instruction information is dequantized until the vector quantization switch flag is detected again, and so on.

Optionally, in the first possible situation above, at least one codeword indication information corresponding to each transformation coefficient matrix may also correspond to the number of vectors or the number of vectorization times, and the number of vectors or the number of vectorization times is used to indicate The corresponding at least one codeword indicates the number of first indication information included in the information. In this way, after the video decoding device detects that the vector quantization switch flag is turned on, it can also obtain the number of vectors or the number of times of vector quantization, and according to the number of vectors or the number of times of vector quantization, know the number of times after the logo of the vector quantization switch. The codeword indication information is the first indication information corresponding to a transform coefficient matrix.

In the second possible situation, at least one codeword indication information corresponding to each transformation coefficient matrix may contain both the first indication information and the second indication information. In this case, each transformation coefficient matrix corresponds to The at least one codeword indication information of may also correspond to the number of vectors or the number of times of vectorization. At this point, in a possible implementation manner, the video decoding device and the video encoding device may agree in advance that the first a after each vector quantization switch identifier is the first indication information, where a is the vector number or Vector quantization times. On this basis, after the video decoding device detects that the vector quantization switch flag is an on flag, it acquires the indication information of the previous a codewords after the flag is turned on, and at this time, the indication information of the previous a codewords are all first indications information. Afterwards, the video decoding device can use the method of acquiring codebooks described above to acquire the codebooks corresponding to the previous a first indication information, and then obtain the codewords corresponding to each first indication information from the acquired codebooks. After obtaining the codeword corresponding to each first indication information, if the next vector quantization switch identifier has not been detected, it means that the remaining codeword indication information is the second indication information. At this time, the video decoding device can be based on non-vector quantization Other quantization methods of quantization perform inverse quantization on the remaining second indication information to obtain corresponding quantization results, until the next vector quantization switch flag is detected, refer to the above operations for processing.

Optionally, in the second possible situation above, in another possible implementation, at least one codeword indication information corresponding to each transformation coefficient matrix not only corresponds to the number of vectors or the number of times of vectorization, but also corresponds to There is a vector quantization start position, and the vector quantization start position is used to indicate the position of the first first indication information in the corresponding at least one codeword indication information. Based on this, after the video decoding device detects that the vector quantization switch flag is an on flag, it can determine, according to the vector quantization start position, which codeword indication information to start from the first codeword indication information after the on flag is The first indication information, that is, determine where is the first first indication information after the activation flag. In this way, for the codeword indication information before the first indication information, the video decoding device performs inverse quantization based on other quantization methods other than vector quantization to obtain corresponding quantization results. Afterwards, according to the number of vectors or the number of times of vector quantization, a corresponding number of codeword indication information is obtained from the first first indication information, and these acquired codeword indication information is the first indication information, and these first indications are obtained by referring to the aforementioned method The codebook corresponding to the information, and then obtain the codeword corresponding to each first indication information from the obtained codebook. After obtaining the codeword corresponding to each first indication information, if the next vector quantization switch identifier has not been detected, it means that the remaining codeword indication information is the second indication information. At this time, the video decoding device can be based on non-vector quantization Other quantization methods of quantization perform inverse quantization on the remaining second indication information to obtain corresponding quantization results, until the next vector quantization switch flag is detected, refer to the above operations for processing.

After obtaining the quantization result corresponding to each codeword indication information in the plurality of codeword indication information, the video decoding device can refer to the method introduced in the first implementation mode above to obtain an inverse quantization reconstruction value according to each quantization result, and then according to Dequantize the reconstructed value to generate the decoded data of the target video, which will not be repeated in this embodiment of the present application.

It should be noted that, in the first implementation manner and the second implementation manner above, the video code stream of the target video may further include at least one vector dimension.

Wherein, when the multiple codeword indication information of the video code stream of the target video contains a vector dimension before, the video decoding device can learn the multiple codeword indication information contained in the video code stream after detecting the vector dimension The codeword dimension corresponding to the first indication information in . In the case that the video code stream of the target video does not carry a codebook index, the video decoding device may also determine the corresponding codebook through the vector dimension, which is not limited in this embodiment of the present application.

When at least one codeword indication information corresponding to each transformation coefficient matrix in the multiple codeword indication information of the video code stream of the target video is preceded by a vector dimension, after the video decoding device detects a vector dimension, that is The codeword dimension corresponding to the first indication information in the codeword indication information between the vector dimension and the next vector dimension may be known.

When each first indication information among the multiple codeword indication information of the video code stream of the target video is bound with a vector dimension, when the video decoding device dequantizes each first indication information, it can The codeword dimension of the codeword corresponding to the first indication information is obtained through the vector dimension corresponding to the first indication information. At this time, the vector dimension can also be used to indicate that the bundled codeword indication information is the first indication information obtained through vector quantization.

Optionally, in this embodiment of the application, the multiple codeword indication information contained in the video code stream of the target video may also contain a vector quantization enable flag, so that when the video decoding device starts to code the video code of the target video When the stream is being decoded, whether the video code stream of the target video supports vector quantization can be known through the enable flag.

In this embodiment of the present application, the video code stream of the target video contains a plurality of codeword indication information, and the plurality of codeword indication information includes the first indication information obtained through vector quantization, and the video decoding device can be based on the first The instruction information generates the decoded data corresponding to the video code stream. It can be seen that the embodiment of the present application provides a video decoding method based on vector quantization, which realizes the application of vector quantization in video decoding. Since the video decoding device can perform vector quantization The obtained video code stream is decoded, therefore, the video coding device can also generate the video code stream by means of vector quantization. In this way, compared with scalar quantization and network coding quantization, the quantization performance is improved, and the video codec is further improved. performance.

Next, a video encoding device provided by an embodiment of the present application is introduced.

Referring to FIG. 5, an embodiment of the present application provides a video encoding device 500, which includes:

An acquisition module 501, configured to acquire a plurality of transform coefficient matrices corresponding to the target video, each of the plurality of transform coefficient matrices includes a plurality of coefficients to be encoded;

The quantization module 502 is configured to quantize the coefficients to be encoded in multiple transform coefficient matrices to obtain multiple codeword indication information, the multiple codeword indication information is used to indicate the quantization results obtained when the coefficients to be encoded are quantized, and the multiple The codeword indication information includes first indication information, and the first indication information is used to indicate the codeword obtained when the coefficient to be coded is vector quantized;

The generation module 503 is configured to generate a video code stream of the target video according to multiple codeword indication information.

Optionally, the quantization module 502 is mainly used for:

Perform vector quantization or mixed quantization on a plurality of coefficients to be encoded in each of the plurality of transform coefficient matrices to obtain at least one codeword information corresponding to each transform coefficient matrix, wherein the mixed quantization includes vector quantization and division For quantization other than vector quantization, at least one codeword information corresponding to each transformation coefficient matrix includes first indication information, and the first indication information is used to indicate the vector quantization of all or part of the coefficients to be encoded in the corresponding transformation coefficient matrix The code word obtained in the process.

Optionally, the quantization module 502 is also used for:

Perform vector quantization or mixed quantization on a plurality of coefficients to be encoded in each target transform coefficient matrix satisfying the vector quantization condition among the plurality of transform coefficient matrices, and obtain at least one codeword indication information corresponding to each target transform coefficient matrix, wherein, Hybrid quantization includes vector quantization and quantization other than vector quantization, and at least one codeword information corresponding to each target transform coefficient matrix includes first indication information, and the first indication information is used to indicate that all or A codeword obtained during vector quantization of some of the coefficients to be encoded.

Optionally, the plurality of coefficients to be encoded included in each of the plurality of transform coefficient matrices are all the transform coefficients included in the corresponding transform coefficient matrix.

Optionally, the vector quantization condition includes that the number of rows of the transform coefficient matrix reaches the first row number threshold, or the number of columns of the transform coefficient matrix reaches the first column number threshold, or the area of the transform coefficient matrix reaches at least one of the first area thresholds A sort of.

Optionally, the plurality of coefficients to be encoded included in each of the plurality of transform coefficient matrices are transform coefficients in the first area of the corresponding transform coefficient matrix, and the first area is coded by scanning area-based coefficient SRCC The area scanned by technology.

Optionally, the vector quantization condition includes that the number of rows in the first region in the transform coefficient matrix reaches the second row number threshold, or that the column number in the first region in the transform coefficient matrix reaches the second column number threshold, or that the transform coefficient matrix The area of the first region in reaches at least one of the second area thresholds.

Optionally, at least one codeword indication information corresponding to each transform coefficient matrix in the video code stream of the target video corresponds to a vector quantization switch identifier, and the vector quantization switch identifier is used to indicate whether the corresponding at least one codeword information contains The first indication information obtained by vector quantization.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix in the video code stream of the target video corresponds to the number of vectors or the number of times of vector quantization, and the number of vectors or the number of times of vector quantization is used to indicate the corresponding at least one The number of first indication information included in the codeword indication information.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix in the video code stream of the target video also corresponds to a vector quantization start position, and the vector quantization start position is used to indicate the corresponding at least one codeword indication information The position of the first first indication message of .

Optionally, the quantization module 502 is also used for:

If the number of coefficients to be encoded in the first transformation coefficient matrix is N times the codeword dimension, then N input signal vectors are generated according to all coefficients to be encoded in the first transformation coefficient matrix, N is a positive integer, and the input signal The dimension of the vector is equal to the dimension of the code word, the first transformation coefficient matrix is any one of a plurality of transformation coefficient matrices, or the first transformation coefficient matrix is a transformation coefficient matrix satisfying the vector quantization condition among the plurality of transformation coefficient matrices;

Determine the first indication information corresponding to each input signal vector in the N input signal vectors;

The determined N pieces of first indication information are used as at least one codeword indication information corresponding to the first transformation coefficient matrix.

Optionally, the quantization module 502 is also used for:

If the number of coefficients to be coded in the first transform coefficient matrix is not an integer multiple of the codeword dimension, M input signal vectors are generated according to the partial codeword coefficients and the codeword dimension in the first transform coefficient matrix, The dimension of the input signal vector is equal to the dimension of the code word, the first transform coefficient matrix is any one of multiple transform coefficient matrices, or the first transform coefficient matrix is a transform coefficient matrix satisfying the vector quantization condition among the multiple transform coefficient matrices ;

Determine the first indication information corresponding to each input signal vector in the M input signal vectors, and perform non-vector quantization on the remaining coefficients to be encoded in the first transformation coefficient matrix, so as to obtain the second instructions;

The determined first indication information and second indication information are used as indication information of a plurality of codewords corresponding to the first transformation coefficient matrix.

Optionally, the quantization module 502 is also used for:

If the number of coefficients to be encoded in the first transformation coefficient matrix is not an integer multiple of the codeword dimension, then R input signal vectors are generated according to all coefficients to be encoded and the codeword dimension in the first transformation coefficient matrix, The dimensionality of the input signal vector is equal to the codeword dimensionality, and one of the R input signal vectors includes one or more coefficient filling values, and the first transform coefficient matrix is any one of a plurality of transform coefficient matrices, Alternatively, the first transformation coefficient matrix is a transformation coefficient matrix that satisfies the vector quantization condition among the plurality of transformation coefficient matrices;

Determine the first indication information corresponding to each input signal vector in the R input signal vectors;

The determined R pieces of first indication information are used as at least one codeword indication information corresponding to the first transformation coefficient matrix.

Optionally, the first indication information is a codeword index, codeword index indication information, or indication information of a codeword component included in the codeword.

Optionally, the codeword index is a one-dimensional index value, or the codeword index is a multidimensional index coordinate.

Optionally, the video code stream of the target video includes a vector quantization enabling flag, and the vector quantization enabling flag is used to indicate whether the video code stream of the target video supports vector quantization.

Optionally, the video code stream of the target video contains at least one vector dimension, and the at least one vector dimension is used to indicate the dimension of the codewords indicated by each first indication information in the video code stream of the target video.

Optionally, the video code stream of the target video further includes at least one codebook index, and the at least one codebook index is used to indicate the codeword used when determining the codeword corresponding to each first indication information in the video code stream of the target video code book.

Next, a video decoding device provided by an embodiment of the present application is introduced.

Referring to FIG. 6, an embodiment of the present application provides a video encoding device 600, which includes:

The first acquisition module 601 is configured to acquire the video code stream of the target video, the video code stream of the target video includes a plurality of codeword indication information, and the plurality of codeword indication information is used to indicate a plurality of transformation coefficient matrices corresponding to the target video The quantization result obtained when the coefficient to be encoded is quantized, the plurality of codeword indication information includes the first indication information, the first indication information is used to indicate the codeword obtained when the coefficient to be encoded is vector quantized, and the plurality of transform coefficient matrices Each transformation coefficient matrix of includes a plurality of coefficients to be coded;

The second acquiring module 602 is configured to acquire decoded data of the target video according to multiple codeword indication information.

Optionally, the multiple codeword indication information is multiple first indication information, and the second obtaining module 602 is mainly used for:

Acquiring a codeword corresponding to each first indication information in the plurality of first indication information;

Generate decoded data of the target video according to the acquired multiple codewords.

Optionally, the second acquiring module 602 is mainly used for:

Using multiple codewords as dequantized reconstruction values, or processing multiple codewords to obtain dequantized reconstruction values;

The decoded data of the target video is generated according to the dequantized reconstruction value.

Optionally, the plurality of codeword indication information includes a plurality of first indication information and a plurality of second indication information, and the second indication information is used to indicate quantization results obtained through other quantizations other than vector quantization; the second obtaining module 602 also Mainly used:

Acquiring a codeword corresponding to each first indication information in a plurality of first indication information, and acquiring a quantization result corresponding to each second indication information in a plurality of second indication information;

The decoded data of the target video is generated according to the obtained multiple codewords and the quantization result corresponding to each second indication information.

Optionally, the first indication information is a codeword index, codeword index indication information, or indication information of a codeword component included in the codeword.

Optionally, the video code stream of the target video further includes at least one codebook index, at least one codebook index is used to indicate the codeword used when determining the codeword corresponding to each first indication information among the multiple codeword indication information code book.

Optionally, the video code stream of the target video further includes a vector quantization enable flag, and the vector quantization enable flag is used to indicate whether the video code stream of the target video supports vector quantization.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix among the plurality of codeword indication information included in the video code stream of the target video corresponds to a vector quantization switch identifier, and the vector quantization switch identifier is used to indicate at least one corresponding Whether the codeword information includes the first indication information obtained through vector quantization.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix in the multiple codeword indication information included in the video code stream of the target video corresponds to the number of vectors or the number of times of vector quantization, and the number of vectors or the number of times of vector quantization is represented by Indicates the number of first indication information included in the corresponding at least one codeword indication information.

Optionally, at least one codeword indication information corresponding to each transformation coefficient matrix also corresponds to a vector quantization start position, and the vector quantization start position is used to indicate the first first indication in the corresponding at least one codeword indication information The location of the information.

Optionally, the video code stream of the target video contains at least one vector dimension, and the at least one vector dimension is used to indicate the dimension of the codewords indicated by each first indication information in the video code stream of the target video.

To sum up, in the embodiment of the present application, the vector quantization method can be used to quantize the coefficients to be encoded in multiple transformation coefficient matrices to obtain the corresponding first indication information, and then according to the Multiple codewords indicate information to generate a video code stream of the target video. Subsequently, the decoding end can decode the vector quantized coefficients according to the first indication information. It can be seen that the embodiment of the present application provides a video encoding and decoding method based on vector quantization, which realizes the application of vector quantization in video encoding and decoding. Compared with scalar quantization and network coding quantization, the quantization performance is improved, thereby improving The performance of video codec.

It should be noted that, the video encoding and video decoding devices provided in the above-mentioned embodiments only use the division of the above-mentioned functional modules for illustration during video encoding and video decoding. In practical applications, the above-mentioned functions can be assigned by different The functional modules are completed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the video encoding and video decoding apparatuses provided in the above embodiments belong to the same concept as the video encoding and video decoding method embodiments, and the specific implementation process thereof is detailed in the method embodiments, and will not be repeated here.

Fig. 7 is a schematic structural diagram of a server according to an exemplary embodiment. The functions of the video encoding or video decoding server in the above embodiments can be implemented by the server shown in FIG. 7 . The server may be a server in a background server cluster. Specifically:

The server 700 includes a central processing unit (Central Processing Unit, CPU) 701, a system memory 704 including a random access memory (Random Access Memory, RAM) 702 and a read-only memory (Read-Only Memory, ROM) 703, and a connection system memory 704 and the system bus 705 of the central processing unit 701. The server 700 also includes a basic input/output system (Input/Output, I/O system) 706 that helps to transmit information between various devices in the computer, and a large computer for storing operating systems 713, application programs 714, and other program modules 715. capacity storage device 707 .

The basic input/output system 706 includes a display 708 for displaying information and an input device 709 such as a mouse and a keyboard for user input of information. Both the display 708 and the input device 709 are connected to the central processing unit 701 through the input and output controller 710 connected to the system bus 705 . The basic input/output system 706 may also include an input output controller 710 for receiving and processing input from a number of other devices such as a keyboard, mouse, or electronic stylus. Similarly, input output controller 710 also provides output to a display screen, printer, or other type of output device.

Mass storage device 707 is connected to central processing unit 701 through a mass storage controller (not shown) connected to system bus 705 . Mass storage device 707 and its associated computer-readable media provide non-volatile storage for server 700 . That is to say, the mass storage device 707 may include a computer-readable medium (not shown) such as a hard disk or a CD-ROM (Compact Disc Read-Only Memory, Compact Disc Read-Only Memory) drive.

Without loss of generality, computer-readable media may comprise computer storage media and communication media. Computer storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media include RAM, ROM, EPROM (Erasable Programmable Read Only Memory, erasable programmable read-only memory), EEPROM (Electrically Erasable Programmable Read Only Memory, electrically erasable programmable read-only memory), flash memory or other solid-state storage Device, CD-ROM, DVD (Digital Versatile Disc, Digital Versatile Disc) or other optical storage, cassette, tape, disk storage or other magnetic storage device. Certainly, those skilled in the art know that the computer storage medium is not limited to the above-mentioned ones. The above-mentioned system memory 704 and mass storage device 707 may be collectively referred to as memory.

According to various embodiments of the present application, the server 700 can also run on a remote computer connected to the network through a network such as the Internet. That is to say, the server 700 can be connected to the network 712 through the network interface unit 711 connected to the system bus 705, or in other words, the network interface unit 711 can also be used to connect to other types of networks or remote computer systems (not shown).

The above-mentioned memory also includes one or more programs, one or more programs are stored in the memory and configured to be executed by the CPU. The one or more programs include instructions for performing the video encoding and video decoding methods provided by the embodiments of the present application.

The embodiment of the present application also provides a computer-readable storage medium. When the instructions in the storage medium are executed by the processor of the server, the server can execute the video encoding and video decoding methods provided in the foregoing embodiments. For example, the computer readable storage medium may be ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, and the like. It should be noted that the computer-readable storage medium mentioned in the embodiment of the present application may be a non-volatile storage medium, in other words, may be a non-transitory storage medium.

It should be understood that all or part of the steps for implementing the above embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented using software, it may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. The computer instructions may be stored in the above-mentioned computer-readable storage medium.

That is, in some embodiments, there is also provided a computer program product including instructions, which, when run on a computer, causes the computer to execute the video encoding or video decoding methods provided in the above embodiments.

The above description is not intended to limit the embodiments of the present application, and any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the embodiments of the present application shall be included within the scope of protection of the embodiments of the present application.

Claims (30)

1. A method of video encoding, the method comprising:

obtaining a plurality of transformation coefficient matrixes corresponding to a target video, wherein each transformation coefficient matrix in the plurality of transformation coefficient matrixes comprises a plurality of coefficients to be coded;

quantizing a coefficient to be coded in the multiple transformation coefficient matrixes to obtain multiple code word indication information, wherein the multiple code word indication information is used for indicating a quantization result obtained when the coefficient to be coded is quantized, the multiple code word indication information comprises first indication information, and the first indication information is used for indicating a code word obtained when the coefficient to be coded is subjected to vector quantization;

generating a video code stream of the target video according to the plurality of code word indication information;

the quantizing the coefficients to be encoded in the plurality of transform coefficient matrices includes:

if the number of the coefficients to be coded in a first transformation coefficient matrix is N times of the dimension of the code word, generating N input signal vectors according to all the coefficients to be coded in the first transformation coefficient matrix, wherein N is a positive integer, the dimension of the input signal vector is equal to the dimension of the code word, and the first transformation coefficient matrix is any one of the plurality of transformation coefficient matrices, or the first transformation coefficient matrix is the transformation coefficient matrix which meets the vector quantization condition in the plurality of transformation coefficient matrices; determining first indication information corresponding to each input signal vector in the N input signal vectors; and taking the determined N pieces of first indication information as at least one piece of code word indication information corresponding to the first transformation coefficient matrix.

2. The method according to claim 1, wherein the quantizing the coefficients to be encoded in the plurality of transform coefficient matrices to obtain a plurality of codeword indication information includes:

and performing vector quantization or mixed quantization on a plurality of coefficients to be coded in each of the plurality of transform coefficient matrices to obtain at least one piece of codeword information corresponding to each transform coefficient matrix, wherein the mixed quantization includes vector quantization and other quantization except for vector quantization, the at least one piece of codeword information corresponding to each transform coefficient matrix includes the first indication information, and the first indication information is used for indicating codewords obtained in the process of performing vector quantization on all or part of coefficients to be coded in the corresponding transform coefficient matrix.

3. The method according to claim 1, wherein the quantizing the coefficients to be encoded in the plurality of transform coefficient matrices to obtain a plurality of codeword indication information includes:

and performing vector quantization or mixed quantization on a plurality of coefficients to be coded in each target transform coefficient matrix which meets the vector quantization condition in the plurality of transform coefficient matrices to obtain at least one piece of codeword indication information corresponding to each target transform coefficient matrix, wherein the mixed quantization comprises vector quantization and other quantization except for the vector quantization, the at least one piece of codeword information corresponding to each target transform coefficient matrix comprises the first indication information, and the first indication information is used for indicating codewords obtained in the process of performing the vector quantization on all or part of the coefficients to be coded in the corresponding transform coefficient matrix.

4. The method of claim 3, wherein the plurality of coefficients to be encoded included in each of the plurality of matrices of transform coefficients is all transform coefficients included in the respective matrix of transform coefficients.

5. The method of claim 4, wherein the vector quantization condition comprises at least one of a number of rows of the matrix of transform coefficients reaching a first row number threshold, a number of columns of the matrix of transform coefficients reaching a first column number threshold, or an area of the matrix of transform coefficients reaching a first area threshold.

6. The method of claim 3, wherein the plurality of coefficients to be encoded included in each of the plurality of transform coefficient matrices are transform coefficients in a first region of the corresponding transform coefficient matrix, the first region being a region scanned by a scanning region-based coefficient coding (SRCC) technique.

7. The method of claim 6, wherein the vector quantization condition comprises at least one of a number of rows of a first region in the transform coefficient matrix reaching a second row number threshold, or a number of columns of the first region in the transform coefficient matrix reaching a second column number threshold, or an area of the first region in the transform coefficient matrix reaching a second area threshold.

8. The method according to claim 3, wherein the at least one codeword indication information corresponding to each transform coefficient matrix in the video stream of the target video corresponds to a vector quantization switch identifier, and the vector quantization switch identifier is used to indicate whether the corresponding at least one codeword information includes the first indication information obtained by vector quantization.

9. The method according to claim 2 or 3, wherein the at least one codeword indication information corresponding to each transform coefficient matrix in the video code stream of the target video corresponds to a vector number or a vector quantization number, and the vector number or the vector quantization number is used to indicate the number of the first indication information included in the corresponding at least one codeword indication information.

10. The method according to claim 9, wherein at least one codeword indication information corresponding to each transform coefficient matrix in the video stream of the target video further corresponds to a vector quantization start position, and the vector quantization start position is used to indicate a position of a first indication information in the corresponding at least one codeword indication information.

11. The method of claim 1, wherein quantizing the coefficients to be encoded in the plurality of transform coefficient matrices, further comprises:

if the number of the coefficients to be coded in the first transformation coefficient matrix is not integer times of the number of dimensions of the code word, generating M input signal vectors according to part of the coefficients to be coded in the first transformation coefficient matrix and the number of dimensions of the code word, wherein the number of dimensions of the input signal vectors is equal to the number of dimensions of the code word;

determining first indication information corresponding to each input signal vector in the M input signal vectors, and performing other non-vector quantization on the residual coefficients to be coded in the first transform coefficient matrix to obtain second indication information corresponding to the residual coefficients to be coded;

and taking the determined first indication information and the second indication information as a plurality of code word indication information corresponding to the first transformation coefficient matrix.

12. The method of claim 1, wherein quantizing the coefficients to be encoded in the plurality of transform coefficient matrices, further comprises:

if the number of the coefficients to be coded in the first transformation coefficient matrix is not an integer multiple of the code word dimension, generating R input signal vectors according to all the coefficients to be coded in the first transformation coefficient matrix and the code word dimension, wherein the dimension of the input signal vector is equal to the code word dimension, and one input signal vector in the R input signal vectors comprises one or more coefficient filling values;

determining first indication information corresponding to each input signal vector in the R input signal vectors;

and taking the determined R pieces of first indication information as at least one piece of code word indication information corresponding to the first transformation coefficient matrix.

13. The method according to claim 1, wherein the first indication information is a codeword index, codeword index indication information, or indication information of codeword components included in a codeword, and the codeword index is a one-dimensional index value or the codeword index is a multi-dimensional index coordinate.

14. The method of claim 1, wherein the video bitstream of the target video comprises an enable flag of vector quantization, and wherein the enable flag of vector quantization is used to indicate whether the video bitstream of the target video supports vector quantization.

15. The method of claim 1, wherein at least one vector dimension is included in the video code stream of the target video, and the at least one vector dimension is used to indicate a dimension of the codeword indicated by each first indication information in the video code stream of the target video.

16. The method according to claim 1, wherein the video stream of the target video further includes at least one codebook index, and the at least one codebook index is used for indicating a codebook adopted when determining a codeword corresponding to each first indication information in the video stream of the target video.

17. A method of video decoding, the method comprising:

the method comprises the steps of obtaining a video code stream of a target video, wherein the video code stream of the target video comprises a plurality of code word indication information, the code word indication information is used for indicating a quantization result obtained when a coefficient to be coded in a plurality of transformation coefficient matrixes corresponding to the target video is quantized, the code word indication information comprises a plurality of first indication information and a plurality of second indication information, the first indication information is used for indicating a code word obtained when the coefficient to be coded is subjected to vector quantization, the second indication information is used for indicating a quantization result obtained through other quantization which is not subjected to vector quantization, and each transformation coefficient matrix in the transformation coefficient matrixes comprises a plurality of coefficients to be coded;

acquiring decoding data of the target video according to the plurality of code word indication information;

the obtaining of the decoded data of the target video according to the multiple codeword indication information includes:

acquiring a code word corresponding to each first indication information in the plurality of first indication information, and acquiring a quantization result corresponding to each second indication information in the plurality of second indication information;

and generating the decoded data of the target video according to the acquired multiple code words and the quantization result corresponding to each piece of second indication information.

18. The method of claim 17, wherein the plurality of codeword indication information is a plurality of first indication information, and the obtaining the decoded data of the target video according to the plurality of codeword indication information comprises:

acquiring a code word corresponding to each piece of first indication information in the plurality of pieces of first indication information;

and generating the decoding data of the target video according to the acquired multiple code words.

19. The method according to claim 18, wherein the generating decoded data of the target video according to the obtained multiple code words comprises:

taking the multiple code words as inverse quantization reconstruction values, or processing the multiple code words to obtain inverse quantization reconstruction values;

and generating the decoding data of the target video according to the inverse quantization reconstruction value.

20. The method according to any of claims 17-19, wherein the first indication information is a codeword index, codeword index indication information, or indication information of codeword components included in a codeword, and the codeword index is a one-dimensional index value, or the codeword index is a multi-dimensional index coordinate.

21. The method of claim 17, further comprising at least one codebook index in the video bitstream of the target video, wherein the at least one codebook index is used to indicate a codebook used in determining the codeword corresponding to each first indication information in the plurality of codeword indication information.

22. The method of claim 17, wherein the video bitstream of the target video further comprises an enable flag for vector quantization, and wherein the enable flag for vector quantization is used to indicate whether the video bitstream of the target video supports vector quantization.

23. The method according to claim 17, 19 or 22, wherein at least one codeword indication information corresponding to each transform coefficient matrix in a plurality of codeword indication information included in the video stream of the target video corresponds to a vector quantization switch identifier, and the vector quantization switch identifier is used to indicate whether the corresponding at least one codeword information includes the first indication information obtained by vector quantization.

24. The method of claim 17, wherein at least one of the codeword indication information corresponding to each transform coefficient matrix in the multiple codeword indication information included in the video stream of the target video corresponds to a vector number or a vector quantization time, and the vector number or the vector quantization time is used to indicate a number of the first indication information included in the corresponding at least one codeword indication information.

25. The method of claim 24, wherein the at least one codeword indication corresponding to each transform coefficient matrix further corresponds to a vector quantization start position, and the vector quantization start position is used to indicate a position of a first indication information in the corresponding at least one codeword indication information.

26. The method of claim 17, wherein at least one vector dimension is included in the video stream of the target video, and the at least one vector dimension is used to indicate a dimension of a codeword indicated by each first indication information in the video stream of the target video.

27. A video encoding apparatus, characterized in that the apparatus comprises:

the video coding device comprises an acquisition module, a coding module and a decoding module, wherein the acquisition module is used for acquiring a plurality of transformation coefficient matrixes corresponding to a target video, and each transformation coefficient matrix in the plurality of transformation coefficient matrixes comprises a plurality of coefficients to be coded;

the quantization module is configured to quantize coefficients to be encoded in the multiple transform coefficient matrices to obtain multiple codeword indication information, where the multiple codeword indication information is used to indicate a quantization result obtained when the coefficients to be encoded are quantized, the multiple codeword indication information includes first indication information, and the first indication information is used to indicate a codeword obtained when the coefficients to be encoded are vector quantized;

the generating module is used for generating a video code stream of the target video according to the multiple code word indicating information;

the quantization module is configured to generate N input signal vectors according to all the coefficients to be encoded in a first transform coefficient matrix if the number of the coefficients to be encoded in the first transform coefficient matrix is N times of a codeword dimension, where N is a positive integer, the dimension of the input signal vector is equal to the codeword dimension, the first transform coefficient matrix is any one of the transform coefficient matrices, or the first transform coefficient matrix is a transform coefficient matrix satisfying a vector quantization condition among the transform coefficient matrices; determining first indication information corresponding to each input signal vector in the N input signal vectors; and taking the determined N pieces of first indication information as at least one piece of code word indication information corresponding to the first transformation coefficient matrix.

28. A video decoding apparatus, characterized in that the apparatus comprises:

the video code stream of the target video comprises a plurality of code word indication information, the plurality of code word indication information are used for indicating quantization results obtained when coefficients to be coded in a plurality of transformation coefficient matrixes corresponding to the target video are quantized, the plurality of code word indication information comprise a plurality of first indication information and a plurality of second indication information, the first indication information is used for indicating code words obtained when the coefficients to be coded are subjected to vector quantization, the second indication information is used for indicating quantization results obtained through other quantization which is not subjected to vector quantization, and each transformation coefficient matrix in the plurality of transformation coefficient matrixes comprises a plurality of coefficients to be coded;

a second obtaining module, configured to obtain, according to the multiple codeword indication information, decoded data of the target video;

the second obtaining module is configured to obtain a codeword corresponding to each of the plurality of first indication information, and obtain a quantization result corresponding to each of the plurality of second indication information; and generating the decoded data of the target video according to the acquired multiple code words and the quantization result corresponding to each piece of second indication information.

29. A computer-readable storage medium, in which a computer program is stored, which, when executed by a computer, implements the video encoding method of any one of claims 1 to 16.

30. A computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, and when executed by a computer, the computer program implements the video decoding method according to any one of claims 17 to 26.

Images