WO2022077490A1

WO2022077490A1 - Intra prediction method, encoder, decoder, and storage medium

Info

Publication number: WO2022077490A1
Application number: PCT/CN2020/121668
Authority: WO
Inventors: 谢志煌
Original assignee: Oppo广东移动通信有限公司
Priority date: 2020-10-16
Filing date: 2020-10-16
Publication date: 2022-04-21
Also published as: CN116325727A; MX2023000279A; CN116980596A

Abstract

Embodiments of the present application provide an intra prediction method, an encoder, a decoder, and a storage medium, comprising: traversing intra prediction modes to determine an initial prediction value of an initial prediction block corresponding to a current block; performing intra prediction filtering and intra prediction smooth filtering respectively on the initial prediction block to obtain a first-type prediction value and a second-type prediction value, wherein the intra prediction smooth filtering is a filtering process performed on the current block using multiple adjacent reference pixels in each of at least two adjacent reference pixel sets; using the initial prediction value, the first-type prediction value and the second-type prediction value to calculate a rate distortion cost of an original pixel value of the current block, and determining a current prediction mode corresponding to an optimal rate distortion cost; performing intra prediction on the current block using the current prediction mode; and writing index information of the current prediction mode and a filtering identifier into a code stream, wherein the filtering identifier represents an identifier corresponding to intra prediction filtering and/or intra prediction smooth filtering.

Description

Intra-frame prediction method, encoder, decoder and storage medium

technical field

The embodiments of the present application relate to video coding technologies, and relate to, but are not limited to, an intra-frame prediction method, an encoder, a decoder, and a storage medium.

Background technique

In video encoding and decoding, in the process of encoding and decoding the current block, in addition to the inter-frame prediction mode, intra-frame prediction can also be used. Intra-frame prediction may perform an intra-frame prediction process separately for each coding unit to calculate a prediction block.

In the prior art, intra-frame prediction usually uses the respective angle mode and non-angle mode to predict the current block to obtain the predicted block, and then selects the optimal prediction of the current block according to the rate-distortion cost information calculated from the predicted block and the original block. mode, and then perform intra prediction based on the prediction mode. The current intra-frame angle prediction does not use all reference pixels, and it is easy to ignore the correlation between some pixels and the current block. The intra-frame prediction filtering technology improves the pixel prediction accuracy through point-to-point filtering, which can effectively enhance the spatial correlation. Therefore, the intra-frame prediction filtering technology is adopted to improve the pixel prediction accuracy through point-to-point filtering, which can effectively enhance the spatial correlation, thereby improving the intra-frame prediction accuracy.

However, in the intra-frame prediction filtering technique for image blocks with relatively blurred textures, over-sharp prediction may lead to an increase and larger residual error, and the prediction effect is not good, resulting in a decrease in encoding and decoding efficiency.

SUMMARY OF THE INVENTION

The embodiments of the present application provide an intra-frame prediction method, an encoder, a decoder, and a storage medium, which can smooth the prediction process of processed image blocks and improve encoding and decoding efficiency.

In a first aspect, an embodiment of the present application provides an intra-frame prediction method, which is applied to an encoder, and the method includes:

Traverse the intra-frame prediction mode to determine the initial prediction value of the initial prediction block corresponding to the current block;

Perform intra-frame prediction filtering and intra-frame prediction smoothing filtering on the initial prediction block respectively, to obtain a first-type prediction value and a second-type prediction value; wherein, the intra-frame prediction smoothing filtering adopts at least two adjacent reference values A process of filtering the current block for multiple adjacent reference pixels in each adjacent reference pixel set in the pixel set;

Using the initial predicted value, the first-type predicted value and the second-type predicted value, respectively, perform rate-distortion cost calculation with the original pixel value of the current block, and determine the current prediction mode corresponding to the optimal rate-distortion cost;

using the current prediction mode to perform intra prediction on the current block;

The index information of the current prediction mode and the filter identifier are written into the code stream, where the filter identifier represents an identifier corresponding to intra-frame prediction filtering and/or intra-frame prediction smoothing filtering.

In a second aspect, an embodiment of the present application provides an intra-frame prediction method, which is applied to a decoder, and the method includes:

Obtain the code stream, parse the header information of the code stream, and obtain the filter identification;

When the filtering flag indicates that the intra-frame prediction filtering permission flag is valid and the intra-frame prediction smoothing filtering permission flag is valid, at least obtain the current prediction mode and the intra-frame prediction filtering use flag from the code stream;

When the intra-frame prediction filtering use flag is invalid and a valid intra-frame prediction smoothing filter use flag is obtained from the code stream, based on the current prediction mode, perform intra-frame prediction smoothing filtering processing on the current block to obtain the first a predicted value.

In a third aspect, an embodiment of the present application provides an encoder, including:

The first determining part is configured to traverse the intra prediction mode, and determine the initial prediction value of the initial prediction block corresponding to the current block;

The first filtering part is configured to perform intra-frame prediction filtering and intra-frame prediction smoothing filtering respectively on the initial prediction block to obtain a first-type prediction value and a second-type prediction value; wherein, the intra-frame prediction smoothing filtering process A process of filtering the current block for using a plurality of adjacent reference pixels in each adjacent reference pixel set in at least two adjacent reference pixel sets;

The first determining part is further configured to use the initial prediction value, the first type prediction value and the second type prediction value to calculate the rate-distortion cost with the original pixel value of the current block, respectively, and determine the corresponding optimal rate-distortion cost. The current forecast mode of ;

a first prediction part configured to use the current prediction mode to perform intra prediction on the current block;

The writing part is configured to write the index information of the current prediction mode and the filtering identification into the code stream, wherein the filtering identification represents the identification corresponding to intra-frame prediction filtering and/or intra-frame prediction smoothing filtering.

In a fourth aspect, an embodiment of the present application provides a decoder, including:

The parsing part is configured to obtain the code stream, parse the header information of the code stream, and obtain the filtering identifier; and when the filtering identifier is an indication that the intra-frame prediction filtering permission flag is valid, and the intra-frame prediction smoothing filtering permission flag is valid, Obtain at least the current prediction mode and intra-frame prediction filtering use identifier from the code stream;

The second prediction part is configured to, based on the current prediction mode, frame the current block when the intra-frame prediction filtering use flag is invalid and a valid intra-frame prediction smoothing filter use flag is obtained from the code stream The intra-prediction smoothing filtering process is performed to obtain the first predicted value.

In a fifth aspect, the embodiments of the present application also provide an encoder, including:

a first memory and a first processor;

The first memory stores a computer program executable on a first processor, which when executed by the first processor implements the intra prediction method of the encoder.

In a sixth aspect, an embodiment of the present application also provides a decoder, including:

a second memory and a second processor;

The second memory stores a computer program executable on a second processor, which when executed by the second processor implements the intra prediction method of the decoder.

In a seventh aspect, an embodiment of the present application provides a storage medium, including:

A computer program is stored thereon, and when the computer program is executed by the first processor, the intra-frame prediction method of the encoder is realized; or, when the computer program is executed by the second processor, the intra-frame prediction method of the decoder is realized. method of prediction.

Embodiments of the present application provide an intra-frame prediction method, an encoder, a decoder, and a storage medium. By using intra-frame prediction filtering in the encoder, it is possible to use at least two adjacent reference pixel sets in each adjacent reference pixel set. The process of filtering the current block of a neighboring reference pixel. Intra-frame prediction smoothing filtering can refer to multiple adjacent reference pixels in one adjacent reference pixel set, so such prediction will be smoother, and the prediction for blurred images will be more accurate, thereby improving coding efficiency. .

Description of drawings

1A-1C are exemplary distribution diagrams of components in different color formats provided by the embodiments of the present application;

2 is a schematic diagram of the division of an exemplary coding unit provided by an embodiment of the present application;

3 is a schematic diagram of an exemplary prediction mode provided by an embodiment of the present application;

4A-4C are schematic diagrams of three filtering situations of exemplary intra-frame prediction filtering provided by an embodiment of the present application;

FIG. 5 is a schematic diagram of the composition and structure of an exemplary network architecture for video encoding and decoding provided by an embodiment of the present application;

FIG. 6 is a structural diagram of an exemplary video coding system provided by an embodiment of the present application;

FIG. 7 is a structural diagram of an exemplary video decoding system provided by an embodiment of the present application;

FIG. 8 is a flowchart 1 of an intra-frame prediction method provided by an embodiment of the present application;

9A-9C are schematic diagrams of exemplary target reference pixels provided by embodiments of the present application;

10 is a second flowchart of an intra-frame prediction method provided by an embodiment of the present application;

11 is a flowchart of an intra-frame prediction method further provided by an embodiment of the present application;

FIG. 12 is a schematic structural diagram 1 of an encoder according to an embodiment of the present application;

FIG. 13 is a second schematic structural diagram of an encoder according to an embodiment of the application;

FIG. 14 is a schematic structural diagram 1 of a decoder according to an embodiment of the present application;

FIG. 15 is a second schematic structural diagram of a decoder according to an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. It should be understood that the specific embodiments described herein are only used to explain the related application, but not to limit the application. In addition, it should be noted that, for the convenience of description, only the parts related to the relevant application are shown in the drawings.

In a video image, a first image component, a second image component and a third image component are generally used to represent the current block (Coding Block, CB); wherein, the three image components are a luminance component and a blue chrominance component respectively. and a red chrominance component, specifically, the luminance component is usually represented by the symbol Y, the blue chrominance component is usually represented by the symbol Cb or U, and the red chrominance component is usually represented by the symbol Cr or V; in this way, the video image can use the YCbCr format Representation can also be represented in YUV format.

Usually, digital video compression technology works on image data whose color coding method is YCbCr (YUV) format, YUV ratio is generally measured as 4:2:0, 4:2:2 or 4:4:4, Y represents brightness ( Luma), Cb(U) represents blue chromaticity, Cr(V) represents red chromaticity, U and V represent chromaticity (Chroma) to describe color and saturation. 1A to 1C show the distribution diagrams of each component in different color formats, wherein white is the Y component, and black and gray are the UV components. As shown in Figure 1A, in the color format, 4:2:0 means that every 4 pixels has 4 luminance components and 2 chrominance components (YYYYCbCr), as shown in Figure 1B, 4:2:2 means that every 4 A pixel has 4 luminance components, 4 chrominance components (YYYYCbCrCbCr), and as shown in Figure 1C, 4:4:4 represents a full pixel display (YYYYCbCrCbCrCbCrCbCr).

Currently, common video coding standards are based on the block-based hybrid coding framework. Each frame in the video image is divided into square largest coding units (Largest Coding Units, LCUs) of the same size (such as 128×128, 64×64, etc.), and each LCU can also be divided into rectangular ones according to the rules. Coding unit (Coding Unit, CU); and the coding unit may be divided into smaller prediction units (Prediction Unit, PU). Specifically, the hybrid coding framework may include modules such as prediction, transform (Transform), quantization (Quantization), entropy coding (EntropyCoding), and loop filtering (In Loop Filter); wherein, the prediction module may include intraPrediction (intraPrediction) And inter prediction (interPrediction), inter prediction can include motion estimation (motion estimation) and motion compensation (motion compensation). Since there is a strong correlation between adjacent pixels in a frame of a video image, the use of intra-frame prediction in the video coding and decoding technology can eliminate the spatial redundancy between adjacent pixels. Inter-frame prediction can refer to the image information of different frames, and use motion estimation to search for the motion vector information that best matches the current divided block to eliminate temporal redundancy; transformation converts the predicted image block to the frequency domain, redistributes the energy, and combines quantization The information that is not sensitive to the human eye can be removed to eliminate visual redundancy; entropy coding can eliminate character redundancy according to the current context model and the probability information of the binary code stream.

It should be noted that in the video coding process, the encoder first reads the image information, and divides the image into several coding tree units (Coding Tree Units, CTUs), and a coding tree unit can be further divided into several coding units ( Coding Unit, CU), these coding units can be rectangular blocks or square blocks, and the specific relationship can be referred to as shown in FIG. 2 .

In the process of intra-frame prediction, the current coding unit cannot refer to the information of different frame images, and can only use the adjacent coding units of the same frame image as reference information for prediction, that is, according to most of the current left-to-right, top-to-bottom The current coding unit can refer to the upper left coding unit, the upper coding unit and the left coding unit are used as reference information to predict the current coding unit, and the current coding unit is used as the reference information of the next coding unit. image for prediction. If the input digital video is in color format, that is, the input source of the current mainstream digital video encoder is YUV 4:2:0 format, that is, every 4 pixels of the image consists of 4 Y components and 2 UV components. The Y component and UV component will be encoded separately, and the encoding tools and technologies used are slightly different. At the same time, the decoding end will also decode according to different formats.

For the intra-frame prediction part in digital video coding and decoding, the current block is mainly predicted with reference to the image information of adjacent blocks of the current frame, the residual information is obtained by calculating the residual between the predicted block and the original image block, and then the process of transformation and quantization is performed. The residual information is transmitted to the decoding end. After the decoding end receives and parses the code stream, the residual information is obtained through inverse transformation and inverse quantization, and the reconstructed image block is obtained by superimposing the residual information on the predicted image block predicted by the decoding end.

In this process, intra prediction usually uses the respective angle mode and non-angle mode to predict the current coding block to obtain the prediction block, and selects the optimal prediction mode for the current coding unit according to the rate-distortion information calculated from the prediction block and the original block. , and then transmit the prediction mode to the decoding end via the code stream. The decoding end parses the prediction mode, predicts the prediction image of the current decoding block, and superimposes the residual pixels transmitted by the code stream to obtain the reconstructed image. After the development of digital video coding and decoding standards over the past generations, the non-angle mode remains relatively stable, including the average mode and the plane mode; the angle mode continues to increase with the evolution of digital video coding standards. Take the international digital video coding standard H series as an example , the H.264/AVC standard has only 8 angle prediction modes and 1 non-angle prediction mode; H.265/HEVC is extended to 33 angle prediction modes and 2 non-angle prediction modes; and the latest general video coding standard H .266/VVC adopts 67 prediction modes, among which 2 non-angle prediction modes are reserved, and the angle mode is expanded from 33 kinds of H.265 to 65 kinds. Undoubtedly, with the increase of the angle mode, the intra-frame prediction will be more accurate, and it will be more in line with the needs of the current society for the development of high-definition and ultra-high-definition video. Not only the international standard, but the domestic digital audio and video coding standard 3 (Audio Video coding Standard, AVS3) also continues to expand the angle mode and non-angle mode. The development of ultra-high-definition digital video puts forward higher requirements for intra-frame prediction, not only It only relies on simply increasing the angle prediction mode and expanding the wide angle to submit coding efficiency. Therefore, the domestic digital audio and video coding standard AVS3 adopts the Intra Prediction Filter (IPF). The intra prediction filtering technology points out that not all reference pixels are used in the current intra-frame angle prediction, and it is easy to ignore some pixels and The correlation between the current coding units, the intra-frame prediction filtering technology improves the pixel prediction accuracy through point-to-point filtering, which can effectively enhance the spatial correlation, thereby improving the intra-frame prediction accuracy. The IPF technology takes the prediction mode from upper right to lower left in AVS3 as an example, as shown in FIG. 3 .

The direction of the prediction mode is from upper right to lower left, and the generated prediction value of the current coding unit mainly uses the reference pixels of the adjacent blocks in the upper MRB row, that is, the prediction pixels of the current coding unit do not refer to the reconstructed pixels of the adjacent blocks on the left. However, the current coding unit and the left reconstruction block are in a spatially adjacent relationship. If only the upper MRB pixels are referenced instead of the left URB pixels, the spatial correlation is likely to be lost, resulting in poor prediction effect.

In order to solve this problem, the intra-frame prediction filter technology (Intra Prediction Filter, IPF) is proposed. The intra-frame prediction filter technology is a filtering method applied to all prediction modes of intra-frame prediction to improve the accuracy of intra-frame prediction.

IPF is mainly realized through the following processes:

1. The intra-frame prediction filtering technology judges the current prediction mode of the coding unit and divides it into a horizontal angle prediction mode, a vertical angle prediction mode and a non-angle prediction mode;

2. According to different types of prediction modes, IPF technology uses different filters to filter input pixels;

3. According to the different distances from the current pixel to the reference pixel, the IPF technology uses different filter coefficients to filter the input pixels;

4. The input pixels of the IPF technology are the predicted pixels obtained in each prediction mode, and the output pixels are the final predicted pixels after IPF filtering.

The IPF technology has an allowable flag bit ipf_enable_flag (intra-frame prediction filtering permission flag), which is an exemplary, binary variable. A value of '1' indicates that intra-frame prediction filtering can be used; a value of '0' indicates that intra-frame prediction filtering should not be used. . At the same time, the IPF technology also uses the flag ipf_flag (intra-frame prediction filtering use flag), which is an exemplary, binary variable. A value of '1' indicates that intra-frame prediction filtering should be used; a value of '0' indicates that intra-frame prediction filtering should not be used. Predictive filtering, if the flag bit ipf_flag does not exist in the code stream, the default value is 0.

Among them, the syntax element representation of ipf_flag is as follows:

It should be noted that the IPF technology classifies

prediction modes

0, 1 and 2 as non-angle prediction modes, and uses the first three-tap filter to filter the prediction pixels; Mode 50 is classified as a vertical angle-like prediction mode, and the predicted pixels are filtered using the first two-tap filter; The second two-tap filter filters the predicted pixels. in,

The above-mentioned first three-tap filter suitable for IPF technology, the filtering formula is shown in formula (1):

P'(x,y)=f(x)·P(-1,y)+f(y)·P(x,-1)+(1-f(x)-f(y))·P( x,y) (1)

The above-mentioned first three-tap filter suitable for IPF technology, the filtering formula is shown in formula (2):

P′(x,y)＝f(x)·P(-1,y)+(1-f(x))·P(x,y) (2)

The above-mentioned second two-tap filter suitable for IPF technology, the filtering formula is shown in formula (3):

P′(x,y)＝f(y)·P(x,-1)+(1-f(y))·P(x,y) (3)

Among them, P'(x, y) is the final prediction value of the pixel located at the position (x, y) of the current chrominance prediction block, and f(x) and f(y) are the horizontal filtering of the reconstructed pixels with reference to the adjacent blocks on the left, respectively. The coefficient and the vertical filter coefficient of the reconstructed pixel of the adjacent block on the reference upper side, P(-1,y) and P(x,-1) are the reconstructed pixels on the left side of row y and the reconstructed pixels on the upper side of column x, respectively, P(x,y) is the original predicted pixel value in the current chrominance component prediction block. The values of x and y do not exceed the range of the width and height of the current coding unit.

It should be noted that the values of the above-mentioned horizontal filter coefficients and vertical filter coefficients are related to the size of the current coding unit and the distances from the predicted pixels in the current prediction block to the left reconstructed pixels and the top reconstructed pixels. The above-mentioned horizontal filter coefficients may be divided into different filter coefficient groups according to the size of the current coding unit. Table 1 presents the filter coefficients of the intra-frame prediction filtering technique.

Table 1

Exemplarily, FIGS. 4A to 4C show schematic diagrams of three filtering situations of intra-frame prediction filtering, respectively only referring to the upper reference pixels to filter the prediction value in the current coding unit (that is, the upper two-tap filtering in FIG. 4A ); Only the left reference pixel is referenced to filter the measurements in the current coding unit pre-filter (ie, the left two-tap filtering of FIG. 4B ); and both the upper and left reference pixels are referenced to the prediction in the current coding unit block values are filtered (ie, three-tap filtering on the upper and left sides of Figure 4C).

The IPF technology acts on the intra-frame prediction part of the video coding hybrid framework, and is specifically applied to all prediction modes of intra-frame prediction, acting on both the encoding end and the decoding end.

Based on the above concepts, an embodiment of the present application provides a network architecture of a video encoding and decoding system including an intra-frame prediction method. FIG. 5 is a schematic diagram of the composition and structure of the network architecture for video encoding and decoding according to an embodiment of the present application, as shown in FIG. 5 . As shown, the network architecture includes one or more electronic devices 12 to 1N and a communication network 01 , wherein the electronic devices 12 to 1N can perform video interaction through the communication network 01 . In the implementation process, the electronic device can be various types of devices with video codec functions, for example, the electronic device can include mobile phones, tablet computers, personal computers, personal digital assistants, navigators, digital phones, video phones, Televisions, sensing devices, servers, etc., are not limited in the embodiments of the present application. The intra-frame prediction apparatus in the embodiment of the present application may be the above-mentioned electronic device.

The electronic device in the embodiments of the present application has a video encoding and decoding function, and generally includes a video encoder (ie, encoder) and a video decoder (ie, decoder).

The present application provides a video coding system. As shown in FIG. 6 , the video coding system 11 includes:

Transform unit 111, quantization unit 112, mode selection and coding control logic unit 113, intra prediction unit 114, inter prediction unit 115 (including: motion compensation and motion estimation), inverse quantization unit 116, inverse transform unit 117, loop Filtering unit 118, coding unit 119 and decoding image buffering unit 110; for the input original video signal, a video reconstruction block can be obtained by the division of coding tree block (Coding Tree Unit, CTU), by mode selection and coding control logic unit 113 Determine the encoding mode, and then transform the video reconstruction block through the transform unit 111 and the quantization unit 112 for the residual pixel information obtained after intra-frame or inter-frame prediction, including transforming the residual information from the pixel domain to the transform domain , and quantize the obtained transform coefficients to further reduce the bit rate; the intra-frame prediction unit 114 is used to perform intra-frame prediction on the video reconstruction block; wherein, the intra-frame prediction unit 114 is used to determine the maximum value of the video reconstruction block. optimal intra prediction mode (ie, target prediction mode); inter prediction unit 115 is used to perform inter prediction encoding of the received video reconstruction block relative to one or more blocks in one or more reference frames to provide temporal Prediction information; where, motion estimation is the process of generating a motion vector that can estimate the motion of the video reconstruction block, and then motion compensation is performed based on the motion vector determined by motion estimation; After the mode, the inter prediction unit 115 is also used to provide the selected inter prediction data to the encoding unit 119, and also send the calculated motion vector data to the encoding unit 119; in addition, the inverse quantization unit 116 and the inverse The transform unit 117 is used for the reconstruction of the video reconstruction block, reconstructing the residual block in the pixel domain, the reconstructed residual block is passed through the loop filtering unit 118 to remove the blocking artifacts, and then, the reconstructed residual block is A predictive block added to the frame of the decoded image buffer unit 110 to generate a reconstructed video reconstruction block; the encoding unit 119 is used to encode various encoding parameters and quantized transform coefficients. The decoded image buffer unit 110 is used for storing reconstructed video reconstruction blocks for prediction reference. As the video image encoding proceeds, new reconstructed video reconstruction blocks are continuously generated, and these reconstructed video reconstruction blocks are all stored in the decoded image buffer unit 110 .

An embodiment of the present application provides a video decoding system, and FIG. 7 is a schematic structural diagram of a video encoding system according to an embodiment of the present application. As shown in FIG. 7 , the video encoding system 12 includes:

Decoding unit 121, inverse transform unit 127, inverse quantization unit 122, intra-frame prediction unit 123, motion compensation unit 124, loop filtering unit 125 and decoded image buffer unit 126; the input video signal is encoded by the video encoding system 11 After processing, the code stream of the video signal is output; the code stream is input into the video decoding system 12, and firstly passes through the decoding unit 121 to obtain the decoded transform coefficient; for the transform coefficient, the inverse transform unit 127 and the inverse quantization unit 122 processed to generate a residual block in the pixel domain; intra-prediction unit 123 may be used to generate prediction data for the current video decoding block based on the determined intra-prediction direction and data from previously decoded blocks of the current frame or picture Motion compensation unit 124 determines prediction information for the video decoding block by parsing the motion vector and other associated syntax elements, and uses the prediction information to generate the predictive block of the video decoding block being decoded; The residual blocks of unit 127 and inverse quantization unit 122 are summed with the corresponding predictive blocks produced by intra prediction unit 123 or motion compensation unit 124 to form a decoded video block; the decoded video signal is passed through the loop filtering unit 125 to remove blocking artifacts, which can improve video quality; the decoded video blocks are then stored in a decoded picture buffer unit 126, which stores reference pictures for subsequent intra prediction or motion compensation, while also storing It is used for the output of the video signal to obtain the restored original video signal.

The implementation of the encoder using IPT for intra-frame prediction coding mode is as follows:

In the current coding unit, if the allowable identification bit of IPF is '1' then all the following steps are carried out, if the allowable identification bit of IPF is '0' then only a), b) and f) steps are carried out:

a), the intra-frame prediction first traverses all prediction modes, calculates the predicted pixels under each intra-frame prediction mode, and calculates the rate-distortion cost according to the original pixels;

b), select the optimal prediction mode of the current coding unit according to the principle of the minimum rate-distortion cost of all the above-mentioned prediction modes, and record the optimal prediction mode information and the corresponding rate-distortion cost information;

c), traverse all intra-frame prediction modes again, this process opens the IPF technology, first calculates the predicted pixels under each intra-frame prediction mode, and obtains the prediction block of the current coding unit;

d), carry out IPF filtering to the prediction block of the current coding unit, select the corresponding filter according to the current prediction mode, select the corresponding filter coefficient group according to the current coding unit size, and the specific correspondence can look up Table 1;

e), calculate the rate-distortion cost information of each prediction mode according to the final prediction pixel and original pixel obtained through the above-mentioned IPF filtering technology, and record the prediction mode and the corresponding cost value of the minimum rate-distortion cost information;

f), if the IPF allows the identification bit to be '0', then transmit the prediction mode index recorded in b) to the decoding end through the code stream; if the IPF allows the identification bit to be '1', then the minimum code recorded in b) The value is compared with the minimum cost value recorded in e), if the rate-distortion cost in b) is smaller, the prediction mode index encoding recorded in b) is used as the optimal prediction mode of the current coding unit and transmitted to decoding through the code stream end, the IPF current coding unit identification position is no, indicating that the IPF technology is not used, and is also transmitted to the decoding end through the code stream; if the rate-distortion in e) is smaller, then the prediction mode index code recorded in e) is used as the current code. The optimal prediction mode of the unit is transmitted to the decoding end through the code stream, and the current IPF coding unit identification position is true, indicating that the IPF technology is used, and is also transmitted to the decoding end through the code stream.

After that, the predicted value is superimposed with the residual information after operations such as transformation and quantization to obtain the reconstructed block of the current coding unit as the reference information of the subsequent coding unit.

The implementation of the decoder for intra-frame prediction coding mode using IPT is as follows:

The decoding end obtains the code stream and parses to obtain the digital video sequence information, and parses to obtain the IPF permission flag of the current video sequence, the coding mode of the current decoding unit is the intra-frame prediction coding mode, and the IPF usage flag of the current decoding unit.

In the current decoding unit, if the allowable identification bit of IPF is '1' then all the following steps are carried out; if the allowable identification bit of IPF is '0' then only a), b) and e) steps are carried out:

a), obtain the code stream information, analyze the residual information of the current decoding unit, and obtain the time-domain residual information through the inverse transformation and inverse quantization process;

b), parse the code stream and obtain the prediction mode index of the current decoding unit, calculate the prediction block of the current decoding unit according to the adjacent reconstruction block and the prediction mode index;

c), parse and obtain the use identification bit of IPF, if the use identification bit of IPF is '0', then do not perform additional operations on the current prediction block; if the use identification bit of IPF is '1', then execute d);

d), select the corresponding filter according to the prediction mode classification information of the current decoding unit, select the corresponding filter coefficient group according to the size of the current decoding unit, then filter each pixel in the prediction block to obtain the final prediction block;

The reconstructed block of the current decoding unit is obtained by superimposing the restored residual information of the prediction block, and output after post-processing.

However, the intra-frame prediction filtering technology can effectively improve the coding efficiency of intra-frame prediction, greatly enhance the spatial correlation of intra-frame prediction, and solve the problem of ignoring unused reference pixels in the process of intra-frame prediction. Impact. However, when the intra-frame prediction process needs to be smoothed, the intra-frame prediction filtering technology and the current intra-frame prediction mode cannot solve similar problems well, and pixel-by-pixel filtering based on reference pixels can improve the correlation between the prediction block and the reference block. However, it cannot solve the smoothing problem inside the prediction block. That is to say, a prediction block calculated according to a single prediction mode usually shows a better prediction effect in an image with a clearer texture, and therefore the residual error will be smaller and smaller, and the coding efficiency will be improved. However, in image blocks with blurred texture, over-sharpened prediction may lead to increased and larger residual errors, poor prediction effect and reduced coding efficiency.

Based on this, the present technology is implemented on the prediction block when performing intra-frame prediction, and for some image blocks that need smoothing processing, the present application proposes an intra-frame prediction method based on the smoothing-processing intra-frame prediction filtering enhancement technology. An intra-frame prediction method provided in this embodiment of the present application mainly acts on the intra-frame prediction unit 114 of the video coding system 11 and the intra-frame prediction unit 123 of the video decoding system 12; The encoding system 11 can obtain a better prediction effect through the intra-frame prediction method provided in the embodiment of the present application, and correspondingly, at the decoding end, the video decoding and restoration quality can also be improved.

Based on this, the technical solutions of the present application are further elaborated below with reference to the accompanying drawings and embodiments. Before going into detail, it should be noted that the "first", "second", "third", etc. mentioned throughout the specification are only for distinguishing different features, and do not have a limited priority or sequence. , size relationship and other functions.

An embodiment of the present application provides an intra-frame prediction method, which is applied to a video encoding device, that is, an encoder. The functions implemented by the method can be implemented by the processor in the video encoding device calling program codes, and of course the program codes can be stored in a computer storage medium. It can be seen that the video encoding device includes at least a processor and a storage medium. Hereinafter, the current coding unit and the current decoding unit are both represented by the current block.

FIG. 8 is a schematic flowchart of an implementation of an intra-frame prediction method according to an embodiment of the present application. As shown in FIG. 8 , the method includes:

S101. Traverse the intra prediction mode to determine the initial prediction value of the initial prediction block corresponding to the current block.

S102, performing intra-frame prediction filtering and intra-frame prediction smoothing filtering on the initial prediction block respectively, to obtain a first-type prediction value and a second-type prediction value; wherein, the intra-frame prediction smoothing filtering adopts at least two adjacent reference pixel sets The process of filtering the current block for multiple adjacent reference pixels in each adjacent reference pixel set.

S103 , using the initial prediction value, the first type prediction value and the second type prediction value, respectively perform rate-distortion cost calculation with the original pixel value of the current block, and determine the current prediction mode corresponding to the optimal rate-distortion cost.

S104. Using the current prediction mode, perform intra-frame prediction on the current block.

S105. Write the index information of the current prediction mode and the filter identifier into the code stream, where the filter identifier represents an identifier corresponding to intra-frame prediction filtering and/or intra-frame prediction smoothing filtering.

In the embodiments of the present application, an intra-frame prediction smoothing filtering technology (Intra Prediction Smooth, IPS) is proposed, and on the basis of IPF, the IPS technology can be superimposed to realize intra-frame prediction. That is, when both IPF and IPS are allowed, the intra-frame prediction process on the encoder side is performed.

In this embodiment of the present application, a video image may be divided into multiple image blocks, and each image block to be encoded currently may be called an encoding block, wherein each encoding block may include a first image component, a second image component, and a first image component. Three image components; and the current block is the coding block in the video image currently to be predicted for the first image component, the second image component or the third image component.

Wherein, it is assumed that the current block performs the first image component prediction, and the first image component is a luminance component, that is, the image component to be predicted is a luminance component, then the current block may also be called a luminance block; or, it is assumed that the current block performs the second image component prediction prediction, and the second image component is a chrominance component, that is, the image component to be predicted is a chrominance component, then the current block may also be called a chrominance block.

It should be noted that the encoder determines the to-be-predicted image component of the current block; based on the parameters of the current block, various prediction modes are used to predict and encode the to-be-predicted image component to obtain an initial predicted value, and then calculate various predictions based on the initial predicted value. The rate-distortion cost corresponding to each prediction mode in the mode; select the minimum rate-distortion cost (ie, the first minimum rate-distortion cost) from the multiple calculated rate-distortion costs, and determine the minimum rate-distortion cost as the optimal rate-distortion cost cost, the prediction mode corresponding to the optimal rate-distortion cost is determined as the prediction mode of the current block.

In the embodiment of the present application, after determining that the encoder traverses all intra-frame prediction modes, determines the initial prediction value of the initial prediction block corresponding to the current block, and then uses the IPF and IPS technologies to filter the current block, the first block is obtained. The first-class predicted value and the second-class predicted value, the initial predicted value, the first-class predicted value and the second-class predicted value are used to calculate the rate-distortion cost with the original pixel value of the current block respectively, and the three corresponding minimum values are determined. rate-distortion cost, and then select the smallest rate-distortion cost among the three minimum rate-distortion costs as the optimal rate-distortion cost, then the prediction mode corresponding to changing the optimal rate-distortion cost is the current prediction mode. Using the current prediction mode, intra-frame prediction is performed on the current block to complete the prediction of the image component to be predicted.

It should be noted that the intra prediction smoothing filtering is a process of filtering the current block by using multiple adjacent reference pixels in each adjacent reference pixel set in at least two adjacent reference pixel sets. Intra-frame prediction filtering is a process of filtering the current block by using one adjacent reference pixel in each adjacent reference pixel set in at least two adjacent reference pixel sets. Intra-frame prediction smoothing filtering can refer to multiple adjacent reference pixels in one adjacent reference pixel set, so such prediction will be smoother, and the prediction for blurred images will be more accurate, thereby improving coding efficiency. .

In the embodiment of the present application, the encoder adopts IPF and IPS technology, and after filtering the current block, the process of obtaining the first type of prediction value and the second type of prediction value: one is to obtain the initial prediction block corresponding to the current block. On the basis of IPF, the initial prediction block is processed by IPF and IPS respectively; the other is that the encoder traverses all intra prediction modes again, and only calculates the initial prediction block in each intra prediction mode when IPF is turned on. , and then perform IPF filtering on the initial prediction block to determine the first type of prediction value. Then, the encoder traverses all intra-frame prediction modes again, and only calculates the initial prediction block in each intra-frame prediction mode when IPS is turned on, and then performs IPS filtering on the initial prediction block to determine the second type of prediction value. The application examples are not limited.

In some embodiments of the present application, the encoder performs intra-frame prediction filtering and intra-frame prediction smoothing filtering on the initial prediction block respectively, and the process of obtaining the first type prediction value and the second type prediction value may be: The left adjacent reference pixel point and the upper adjacent reference pixel point of each pixel in the prediction block, perform intra-frame prediction filtering on each pixel of the initial prediction block to obtain the first type of prediction value; and the encoder is based on the initial prediction value. According to the preset horizontal distance and the preset vertical distance corresponding to each pixel in the prediction block, the initial prediction block is subjected to intra-frame prediction smoothing filtering processing to obtain the second type of prediction value.

It should be noted that the encoder performs IPF filtering on the prediction block of the current block, selects the corresponding filter according to the currently traversed prediction mode, and selects the corresponding filter coefficient group according to the size of the current block. In Table 1, after the filter coefficient is determined, the prediction block is filtered based on the filter coefficient to obtain the rate-distortion cost of each prediction mode, and the smallest one is the second minimum rate-distortion cost.

In some embodiments of the present application, in the process of performing IPS filtering, the encoder may determine, based on the preset horizontal distance and the preset vertical distance corresponding to each pixel in the initial prediction block, at least two phases corresponding to each pixel. At least two adjacent reference block ranges of the adjacent reference pixel set; from the at least two adjacent reference pixel sets, correspondingly determine at least two target reference pixel sets belonging to the at least two adjacent reference block ranges; wherein, each The target reference pixel sets include at least one target reference pixel; based on the at least two target reference pixel sets, intra-frame prediction smoothing filtering is performed on each pixel to obtain a second-type prediction value.

In some embodiments of the present application, the process of the encoder correspondingly determining at least two target reference pixel sets belonging to the at least two adjacent reference pixel sets from the at least two adjacent reference pixel sets may be:

Manner 1: The encoder determines, from at least two adjacent reference pixel sets, pixels at two boundary pixel positions of each adjacent reference block range as a target reference pixel set, thereby determining at least two target reference pixel sets.

Mode 2: The encoder determines, from at least two adjacent reference pixel sets, M pixels at any pixel position within the range of each adjacent reference block as a target reference pixel set, thereby determining at least two target reference pixel sets , where M is greater than 0 and does not exceed the maximum number of pixels in the corresponding adjacent reference pixel set.

In some embodiments of the present application, the at least two adjacent reference pixel sets include: a left adjacent reference pixel set and an upper adjacent reference pixel set; the at least two adjacent reference block ranges include: a left adjacent reference pixel set A pixel range and an upper adjacent reference pixel range; the at least two target reference pixel sets include: a left target adjacent reference pixel set and an upper target adjacent reference pixel set.

It should be noted that the at least two adjacent reference pixel sets include: when the left adjacent reference pixel set and the upper adjacent reference pixel set, the encoder determines the preset horizontal distance and the preset vertical distance corresponding to the current pixel to The current pixel of the current block is used as the reference, and the preset horizontal distance is moved left and right to obtain the first horizontal pixel position and the second horizontal pixel position; and the preset vertical distance is moved up and down respectively to obtain the first vertical pixel position and the second vertical pixel position. pixel location. Based on the first horizontal pixel position and the second horizontal pixel position, it is mapped to the upper adjacent reference pixel set, and the first boundary limit and the second boundary limit of the pixel position are determined. The pixel position between is the upper adjacent reference pixel range corresponding to the current pixel corresponding to the upper adjacent reference pixel set. Based on the first vertical pixel position and the second vertical pixel position, it is mapped to the left adjacent reference pixel set. , the third boundary limit and the fourth boundary limit of the pixel position are determined, and the pixel position between the third boundary limit and the fourth boundary limit is the left adjacent reference pixel set corresponding to the current pixel corresponding to the left adjacent pixel set Reference pixel range, so that at least two adjacent reference block ranges are obtained. For each adjacent reference pixel set, the encoder may select at least one adjacent reference pixel (ie, the target reference pixel set) from the reference pixels included in the corresponding adjacent reference pixel range for performing intra prediction, thereby At least two target reference pixel sets are selected.

In this embodiment of the present application, the encoder does this for each pixel of the current block, so as to complete the processing of the current block, and for the current pixel, the encoder selects a target reference pixel set for an adjacent reference pixel set The number of reference pixels can be one or more, which is determined based on the position of the current pixel in the current block.

It should be noted that the encoder can select the reference pixel value corresponding to the boundary position of each adjacent reference pixel range to form the target reference pixel set, or can select any M reference pixels from each adjacent reference pixel range to form the target reference pixel. A set of pixels, where M may be greater than 0, which is not limited in this embodiment of the present application.

It should be noted that, for different adjacent reference pixel sets, the corresponding determined number of M reference pixels may be different or may be the same, which is not limited in this embodiment of the present application. For example, the target reference pixel set corresponding to the upper adjacent reference pixel set may be UPNum, the target reference pixel set corresponding to the left adjacent reference pixel set may be LFNum, and UPNum and LFNum may be the same or different.

In some embodiments of the present application, the preset horizontal distance is less than or equal to the width of the initial prediction block; the preset vertical distance is less than or equal to the height of the initial prediction block.

In some embodiments of the present application, the preset horizontal distance and the preset vertical distance corresponding to the first type of pixels in the position of n rows and m columns of the initial prediction block are different from those of other types except n rows and m columns of the initial prediction block. The preset horizontal distance and the preset vertical distance corresponding to the pixels are different; n and m are natural numbers; the preset horizontal distance and the preset vertical distance corresponding to each pixel in the first type of pixels are at least partially the same, or both are different; other The preset horizontal distance and the preset vertical distance corresponding to each pixel in the class of pixels are at least partially the same, or both are different.

It should be noted that the preset horizontal distance and the preset vertical distance are equal, and the specific values can be set in the encoder and decoder according to the actual situation, or can be preset in the encoder and decoder according to the size of the current prediction block. In the device, the embodiments of the present application are not limited.

In this embodiment of the present application, the n rows and m columns of the initial prediction block may be the first n rows and the first m columns, or may be in other forms, which are not limited in the embodiments of the present application.

Exemplarily, the above-mentioned IPS technology may have different filtering processes for the first n rows and the first m columns of the initial prediction block. Specifically, the first n rows and the first m columns may use the reference pixels corresponding to the first positional relationship, while other non-first n rows and m columns may use the reference pixels corresponding to the first positional relationship. The reference pixels corresponding to the second positional relationship may be used for the filtering of the predicted pixels of the row and the first m columns. For example, the first three rows and the first three columns use the same HD and VD, which is 1, while the first three rows and the first three columns use the same HD and VD, which is 2.

Wherein, n and m do not exceed the size of the current block, and n and m may be the same or different, which are not limited in this embodiment of the present application.

In some embodiments of the present application, the encoder may determine filter coefficients according to the principle mechanism of IPF; based on at least two target reference pixel sets and filter coefficients, perform intra-frame prediction smoothing filtering on each pixel to obtain the second type Predictive value.

Among them, the filter coefficients corresponding to the first type of pixels in the position of n rows and m columns of the initial prediction block are different from the filter coefficients corresponding to other types of pixels except n rows and m columns of the initial prediction block; n and m are natural numbers; The filter coefficients corresponding to each pixel in one type of pixels are at least partially the same, or all different; the filter coefficients corresponding to each pixel in other types of pixels are at least partially the same, or all different.

In this embodiment of the present application, the IPS technology may have multiple different filter coefficients, for example, the coefficients used by the current predicted pixel in the first n rows and the first m columns are different from the filter coefficients used in the non-first n rows and non-first m columns, etc. , and the filter coefficients corresponding to each pixel in the first n rows and the first m columns may also be the same, partly the same, or all different; the filter coefficients corresponding to each pixel in the non-first n rows and non-first m columns It can also be the same, partially the same, or all different.

It should be noted that, in some embodiments of the embodiments of the present application, the encoder may determine a weight coefficient; based on at least two target reference pixel sets and the weight coefficient, weighted summation is performed on each pixel to complete intra-frame prediction After smoothing and filtering, the second type of predicted value is obtained. Among them, the IPS technology can have multiple different weight coefficients, such as the coefficients used by the current predicted pixel in the first n rows and the first m columns are different from the weight coefficients used in the non-first n rows and non-first m columns, etc., and, The weight coefficients corresponding to each pixel in the first n rows and the first m columns can also be the same, partially the same, or different; the weight coefficients corresponding to each pixel in the non-first n rows and non-first m columns are also Can be the same, partially the same, or all different.

Exemplarily, the encoder uses the position of the current predicted pixel as a reference, and uses the horizontal distance HD and vertical distance VD as offsets to obtain UPNum reconstruction reference pixels adjacent to the upper side and LFNum reconstruction reference pixels on the left side, and obtain The reference pixel and the current predicted pixel are filtered, or the weighted summation can be used to obtain the accurate predicted pixel as the final predicted pixel of the current position in the predicted block.

Exemplarily, as shown in the schematic diagram of 4 target reference pixels with a distance of 2 as shown in FIG. 9A , a is the current pixel, b and c are the adjacent reference pixel ranges, and the encoder obtains 4 reference pixels and predicted pixels. A 5-tap filter or weighting is formed to obtain the accurately corrected predicted pixel value. In the specific example given, the coefficients of the first two rows and the first two columns of the four reference pixels can be set to be 10. The weighting coefficient of the predicted pixels is 216, and the sum needs to be divided by 256 or shifted to the right by 8 bits, and the VD and HD of the first two rows and columns can be set to 1; set 4 reference pixels not the first two rows The weighting coefficients of the first two columns and the non-corrected ones are both 13, and the uncorrected predicted pixel coefficient is 204, and the sum needs to be divided by 256 or shifted to the right by 8 bits.

As shown in Figure 9B, a schematic diagram of 4 target reference pixels with a distance of 1, a is the current pixel, b and c are the adjacent reference pixel ranges, and the encoder obtains 4 reference pixels and predicted pixels to form a 5-tap Filter or weight to obtain the accurately corrected predicted pixel value. In the specific example given, the weighting coefficient of the 4 reference pixels can be set to be 7, and the weighting coefficient of the imprecisely corrected predicted pixel is 228. The sum of Divide by 256 or shift right by 8 bits.

As shown in Figure 9C, a schematic diagram of two target reference pixels with a distance of 0, a is the current pixel, b and c are the adjacent reference pixel ranges, the encoder obtains two reference pixels and predicted pixels to form a 3-tap Filter or weight to obtain the accurately corrected predicted pixel value. In the specific example given, the weighting coefficient of the two reference pixels can be set to be 14, and the weighting coefficient of the uncorrected predicted pixel is 228. The sum of Divide by 256 or shift right by 8 bits.

In this embodiment of the present application, the predicted pixels obtained by the encoder performing the IPS technology may be different from other cases, for example, the second type of prediction of the predicted pixels used by the IPS is obtained by using 4-tap different filter coefficients or different filter coefficients of other taps. value etc.

In some embodiments of the present application, when the encoder uses IPS technology to filter the current block, if the reconstructed reference pixels need to be acquired, that is, the coordinates of the reference pixels in the target reference pixel set exceed the upper adjacent reference pixel set or the left side The size range of the adjacent reference pixel set can also be filled with the adjacent reference pixel set of the current block, and then the target reference pixel set is selected.

In some embodiments of the present application, when the preset horizontal distance is greater than the width of the initial prediction block, the encoder uses the leftmost reference pixel in the upper adjacent reference pixel set to pad to the left until the upper phase is determined. until the first boundary of the range of adjacent reference blocks, and/or, using the rightmost reference pixel in the set of adjacent reference pixels on the upper side to fill to the right, until the second boundary of the range of adjacent reference blocks on the upper side is determined When the preset vertical distance is less than or equal to the height of the initial prediction block, the lowermost reference pixel in the set of adjacent reference pixels on the left side is used to fill down until the third boundary limit of the range of the adjacent reference block on the left side is determined. and/or, the uppermost reference pixel in the set of adjacent reference pixels on the left side is used to fill upwards until the fourth boundary limit of the range of the adjacent reference block on the left side is determined.

Exemplarily, the encoder fills the leftmost reconstruction reference pixel with the uppermost adjacent reconstructed pixel to the left until the required range is satisfied, and uses the rightmost reconstruction reference pixel of the upper adjacent reconstruction pixel to fill to the right until the requirement is met. The requested range; use the reconstructed reference pixel on the uppermost side of the adjacent reconstructed pixels on the left to fill up until the required range is satisfied, and use the adjacent reconstructed pixels on the left to fill down until the required range is satisfied. The above filling process can also be replaced by a conditional judgment method, and the pixels beyond the range can directly request the pixels at the most boundary, etc., which are not limited in this embodiment of the present application.

It should be noted that the encoder performs IPS filtering on the prediction block of the current block, selects the corresponding filter according to the currently traversed prediction mode, determines the filter coefficient, and filters the prediction block based on the filter coefficient to obtain each filter coefficient. The rate-distortion cost of a prediction mode, and the smallest one is selected to be the third smallest rate-distortion cost.

In some embodiments of the present application, after the encoder obtains the predicted value of the first type and the predicted value of the second type, the encoder adopts the initial predicted value, the predicted value of the first type and the predicted value of the second type, which are respectively the same as the current predicted value. The original pixel value of the block is subjected to rate-distortion cost calculation, and the first minimum rate-distortion cost corresponding to the initial prediction value, the second minimum rate-distortion cost corresponding to the first type of prediction value, and the first minimum rate-distortion cost corresponding to the second type of prediction value are determined. Three minimum rate-distortion costs; determine the minimum optimal rate-distortion cost from the first minimum rate-distortion cost, the second minimum rate-distortion cost and the third minimum rate-distortion cost; take the prediction mode corresponding to the obtained optimal rate-distortion cost as Current forecast mode. In this way, the encoder can use the current prediction mode to perform intra-frame prediction on the current block, and write the index information of the current prediction mode and the filter flag into the code stream, where the filter flag represents intra-frame prediction filtering and/or intra-frame prediction Predictive smoothing filter corresponding flag.

In this embodiment of the present application, the filtering flags include: an intra-frame prediction filtering permission flag, an intra-frame prediction filtering use flag, an intra-frame prediction smoothing filtering permission flag, and an intra-frame prediction smoothing filtering use flag.

In some embodiments of the present application, since both IPF and IPS are adopted in the present application, the encoder determines that the intra-frame prediction filtering permission flag is valid, and the intra-frame prediction smoothing filtering permission flag is valid, and sets the intra-frame prediction filtering permission flag to be valid. Allows markers and intra-frame prediction smoothing filtering to allow markers to be written into the stream;

The transmission of the intra-frame prediction filtering use flag and the intra-frame prediction smoothing filter use flag follows one of the following transmission modes to the decoder:

When the optimal rate-distortion cost is the second minimum rate-distortion cost, determine that the intra-frame prediction filtering use flag is valid, and write the intra-frame prediction filtering use flag into the code stream;

When the optimal rate-distortion cost is the third minimum rate-distortion cost, it is determined that the intra-frame prediction filtering use flag is invalid, and the intra-frame prediction smoothing filter use flag is valid, and the intra-frame prediction filtering use flag and the intra-frame prediction smoothing filter use flag are determined. Write the code stream using the identifier;

When the optimal rate-distortion cost is the first minimum rate-distortion cost, it is determined that the intra-frame prediction filtering use flag is invalid, and the intra-frame prediction smoothing filter use flag is invalid, and the intra-frame prediction filtering use flag and the intra-frame prediction smoothing filter use flag are determined to be invalid. Write the codestream with the flag.

In the embodiment of the present application, a binary number may be used to indicate whether it is valid, for example, 1 indicates valid, 0 indicates invalid, or 0 indicates valid, 1 indicates invalid, etc. The specific setting method and setting value are not limited in this embodiment of the present application.

It can be understood that the encoder implements the process of filtering the current block by using one adjacent reference pixel in each adjacent reference pixel set in at least two adjacent reference pixel sets by using intra-frame prediction filtering. Intra-frame prediction smoothing filtering can refer to multiple adjacent reference pixels in one adjacent reference pixel set, so such prediction will be smoother, and the prediction for blurred images will be more accurate, thereby improving coding efficiency. .

In some embodiments of the present application, there may be three types of components to be predicted, to implement the encoding of the three encoding channels of Y, U, and V, that is, the encoding type of the initial predicted value corresponds to multiple encoding channels; then, each encoding channel corresponds to The filtering identifiers of the s are the same identifier, or use one-to-one corresponding identifiers.

That is to say, the IPS technology is applicable to each encoding channel, wherein the luminance component and the chrominance component respectively use independent identification bits to indicate whether the IPS technology is used or not.

It can be understood that, when the encoder uses one filter flag for three coding channels, it can save bits and data amount of transmitted flag bits, and reduce the computational complexity.

In some embodiments of the present application, the encoder may first determine the current block to see if it meets the conditions for IPS processing. The intra-frame prediction method proposed in this application is adopted. in,

When the size of the initial prediction block satisfies the prediction area threshold range, it is allowed to use intra prediction smoothing filtering for filtering;

When the size of the initial prediction block is larger than the upper limit of the prediction area threshold range, the initial prediction block is divided to obtain prediction sub-blocks that meet the prediction area threshold range; the intra-frame prediction processing flow is implemented for the prediction sub-block.

The range of the predicted area threshold can be no less than 4×4 or 8×8, and no more than 64×64 or 32×32, and the detailed value can be set according to the actual situation. If the area is too small, IPS processing can not be performed, but if the area is too large, it can be divided to meet the conditions for using IPS before processing.

It can be understood that the encoder limits the use range of the IPS, and does not use the IPS technology for the unit with a small prediction block area, so as to reduce the transmission identification bits and reduce the computational complexity.

In some embodiments of the present application, after determining the initial prediction value of the initial prediction block corresponding to the current block, the encoder adopts the current prediction mode, and before performing intra prediction on the current block, the encoder may Type, the prediction mode set is determined, and the prediction mode set is the prediction mode that enables the intra-frame prediction smoothing filtering technology; for the initial prediction value, the initial sub-prediction value predicted by the prediction mode set is used to perform intra-frame prediction smoothing filtering. The second type of prediction value; the initial prediction value and the second type of prediction value are used to calculate the rate-distortion cost with the original pixel value of the current block respectively, and the current prediction mode corresponding to the optimal rate-distortion cost is determined.

It is understandable that the encoder limits the scope of use of the IPS, filters the prediction modes of the current block, and only uses the IPS technology for part of the prediction modes, so as to reduce transmission flag bits and reduce computational complexity.

In some embodiments of the present application, the manner in which the encoder obtains the predicted value of the first type and the predicted value of the second type may further include:

After traversing the intra-frame prediction mode and determining the initial prediction value of the initial prediction block corresponding to the current block, the encoder uses the initial prediction value, the first-type prediction value and the second-type prediction value to compare with the original pixel value of the current block respectively. Calculate the rate-distortion cost. Before determining the current prediction mode corresponding to the optimal rate-distortion cost, perform intra-frame prediction filtering on the initial prediction block to obtain the first type of prediction value; traverse the intra-frame prediction mode and use intra-frame prediction smoothing filtering. Perform prediction estimation on the current block to obtain the second type of prediction value.

That is to say, the encoder can traverse the intra prediction mode again, and based on the preset horizontal distance and preset vertical distance corresponding to each pixel in the current block, perform intra prediction smoothing filtering processing on the current block to obtain the second type of prediction value; without filtering the initial predicted value. Wherein, the implementation principle of the filter processing performed on the current block by the encoder is the same as the principle of filtering the initial prediction block in the foregoing embodiment, and details are not described herein again.

Exemplarily, the encoder obtains coding information, including intra-frame prediction filtering (IPF) allowable identification bits and the allowable identification bits of this technical solution (IPS), etc., and after obtaining the image information, the image is divided into several CTUs, and further divided into several CTUs. Several CUs, each independent CU performs intra-frame prediction, the current CU can limit the minimum area and the maximum area for IPS technology.

In the intra-frame prediction process at the encoding end, if the IPF allowable flag bit and the IPS allowable flag bit are both '1', perform all the following steps; if the IPF allowable flag bit is '1' and the IPS allowable flag bit is '0', Then only a), b), c), d), e), and h) are executed; if the IPF allowable flag is '0' and the IPS allowable flag is '1', only a), b), f), g), and h); if both the IPF allow flag and the IPS allow flag are '0', only a), b) and h) are performed. If the above-mentioned IPF or IPS allows the flag bit to be '0', then the coding end defaults to the use flag bit of the corresponding technology of all coding blocks to be '0':

a), the current coding unit traverses all intra-frame prediction modes, calculates the prediction block under each prediction mode, and calculates the rate-distortion cost information of the current prediction mode according to the original pixel block;

c), traverse all intra-frame prediction modes, this process uses IPF technology, first calculates the predicted pixels under each intra-frame prediction mode, and obtains the prediction block of the current coding unit;

f), traverse all intra-frame prediction modes, this process uses IPS technology, calculates the predicted pixels under each intra-frame prediction mode, and calculates the predicted pixel process using IPS technology, which can be added for several tap interpolation filters. Several reference pixels are used as new interpolation filters to predict the predicted pixels of the process. The above-mentioned several tap interpolation filters can be 8 tap filters, and the above-mentioned several reference pixels can be 4. The above-mentioned several reference pixels are obtained. The process is as follows: taking the position of each predicted pixel as a reference, and taking the horizontal distance HD and vertical distance VD as offsets, obtain UPNum reconstruction reference pixels adjacent to the upper side and LFNum reconstruction reference pixels on the left side. The above HD and VD can be 0, 1, 2, 3, ..., N, etc., and N does not exceed the width or height of the current block. The above UPNum and LFNum can be 0, 1, 2, 3, ..., M, etc., and M does not exceed the number of adjacent reconstruction reference pixels on the upper side of the current block and the number of adjacent reconstruction reference pixels on the left side;

g), calculate the rate-distortion cost information of each prediction mode according to the final predicted pixel obtained through the IPS technology and the original pixel, and record the prediction mode and the corresponding cost value of the minimum rate-distortion cost information;

h), if the IPF allows the identification bit to be '0' and the IPS allows the identification bit to be '0', then transmit the prediction mode index recorded in b) to the decoding end through the code stream; if the IPF allows the identification bit to be '1' and If the IPS allows the flag bit to be '0', compare the minimum cost value recorded in b) with the minimum cost value recorded in e), and if the rate-distortion cost in b) is smaller, compare the prediction recorded in b) The mode index coding is used as the optimal prediction mode of the current coding unit and is transmitted to the decoding end through the code stream, and the IPF current coding unit identification position is '0', indicating that the IPF technology is not used, and is also transmitted to the decoding end through the code stream; if e) in The rate distortion is smaller, then the prediction mode index code recorded in e) is transmitted to the decoding end as the optimal prediction mode of the current coding unit through the code stream, and the IPF current coding unit identification position '1' indicates that the IPF technology is used, It is also transmitted to the decoding end through the code stream; if the IPF allows the identification bit to be '0' and the IPS allows the identification bit to be '1', then compare the minimum cost value recorded in b) with the minimum cost value recorded in g), If the rate-distortion cost in b) is smaller, then use the prediction mode index encoding recorded in b) as the optimal prediction mode of the current coding unit and transmit it to the decoding end through the code stream, and use the IPS identifier position '0 of the current coding unit ', indicating that this technology is not used, and it is also transmitted to the decoding end through the code stream; if the rate-distortion in g) is smaller, the prediction mode recorded in g) will be encoded as the optimal prediction mode of the current coding unit through the code stream It is transmitted to the decoding end, and the IPS use flag position of the current coding unit is '1', indicating that this technology is used, and it is also transmitted to the decoding end through the code stream; if the IPF allowable flag bit is '1' and the IPS allowable flag bit is '1' , then compare the minimum cost values recorded in b), e) and g), if the rate-distortion cost in b) is smaller, then encode the prediction mode index recorded in b) as the optimal prediction for the current coding unit The mode is transmitted to the decoding end through the code stream, and the IPS use identification bit and the IPF use identification position of the current coding unit are '0', indicating that they are not used, and are also transmitted to the decoding end through the code stream; if the rate distortion in e) is smaller , then the prediction mode index code recorded in e) is transmitted to the decoding end as the optimal prediction mode of the current coding unit through the code stream, and the IPF of the current coding unit uses the identification position '1 ' and does not transmit the IPS to use the identification position, indicating Using IPF technology instead of IPS technology, it is also transmitted to the decoding end through the code stream; if the rate-distortion in a), is smaller, then use the prediction mode recorded in g) as the optimal prediction mode of the current coding unit. The code stream is transmitted to the decoding end, and the IPF usage flag position of the current coding unit is '0' and the IPS usage flag position is '1', indicating that the IPF technology is not used but the IPS technology is used, and is also transmitted to the decoding end through the code stream.

Finally, the encoder superimposes the predicted block and the inversely transformed and inversely quantized residuals to obtain a reconstructed coding unit block, which is used as a prediction reference block for the next coding block.

An embodiment of the present application provides an intra-frame prediction method, as shown in FIG. 10 , including:

S201. Traverse the intra-frame prediction mode to determine the initial prediction value of the initial prediction block corresponding to the current block.

S202: Perform intra-frame prediction smoothing filtering processing on the initial prediction block respectively, to obtain a second type of prediction value; wherein, the intra-frame prediction smoothing filtering is performed by using the multi-dimensional data in each adjacent reference pixel set in at least two adjacent reference pixel sets. The process of filtering the current block with adjacent reference pixels.

S203 , using the initial predicted value and the second-type predicted value, respectively perform rate-distortion cost calculation with the original pixel value of the current block, and determine the current prediction mode corresponding to the optimal rate-distortion cost.

S204. Using the current prediction mode, perform intra-frame prediction on the current block.

S205: Write the index information of the current prediction mode and the filter flag into the code stream, where the filter flag represents the flag corresponding to the intra-frame prediction smoothing filter.

In the embodiment of the present application, the encoder can only determine whether to perform IPS, and it is not dependent on the implementation of the IPF technology, so that the intra-frame prediction smoothing filtering permission flag is valid.

In this embodiment of the present application, the encoder may only perform IPS, so that when the IPS technology is enabled, the encoder performs intra-frame prediction smoothing filtering on the initial prediction blocks respectively to obtain the second type of prediction value, and the encoder only needs to use the initial prediction block. The predicted value and the second type of predicted value are respectively calculated with the original pixel value of the current block to calculate the rate-distortion cost, and the current prediction mode corresponding to the optimal rate-distortion cost can be determined. The detailed filtering processing method is consistent with the implementation of the foregoing embodiment, and is not repeated here.

In this embodiment of the present application, the filtering flag includes: an intra-frame prediction smoothing filtering permission flag and an intra-frame prediction smoothing filtering use flag. The encoder writes the index information of the current prediction mode into the code stream; determines that the intra-frame prediction smoothing filtering permission flag is valid, and writes the intra-frame prediction smoothing filtering permission flag into the code stream.

Intra-frame prediction smoothing filtering uses one of the following transmission modes:

When the optimal rate-distortion cost is the third minimum rate-distortion cost, it is determined that the intra-frame prediction smoothing filtering use flag is valid, and the intra-frame prediction smoothing filtering use flag is written into the code stream.

When the optimal rate-distortion cost is the first minimum rate-distortion cost, it is determined that the intra-frame prediction smoothing filtering use flag is invalid, and the intra-frame prediction smoothing filtering use flag is written into the code stream.

During the intra-frame prediction process at the coding end, if the IPF allowable flag bit and the IPS allowable flag bit are both '1', perform all the following steps; if the IPF allowable flag bit is '1' and the IPS allowable flag bit is '0', Then only execute a), b), c), d), e), and i); if the IPF allow flag bit is '0' and the IPS allow flag bit is '1', then only execute a), b), f), g), h), and i); if both the IPF allow flag and the IPS allow flag are '0', only execute a), b) and i). If the above-mentioned IPF or IPS allows the flag bit to be '0', then the coding end defaults to the use flag bit of the corresponding technology of all coding blocks to be '0':

f) traversing all intra-frame prediction modes, this process uses IPS technology, first calculates the predicted pixels in each intra-frame prediction mode, and obtains the prediction block of the current coding unit. The interpolation filter for calculating intra-frame predicted pixels in this process may be different from steps a) and c), and a stronger filter can be used to obtain smoother predicted pixels. The specific implementation can be an 8-tap interpolation filter for the angle prediction mode, Exemplarily, the filter coefficients may be shown in the following 4 groups, and each group may include 32 types of 8-tap filter coefficients.

A set: {15,99,238,315,237,101,18,1},{14,95,233,315,242,105,19,1},{13,91,229,315,246,109,20,1},{12,89,226,314,248,112,22,1},2,54,24 1},{10,81,217,313,257,120,25,1},{9,77,211,312,262,125,27,1},{8,74,208,311,265,129,28,1},{7,71,203,310,269,133,30,1},{7,68,200 1},{6,65,195,307,275,141,34,1},{5,62,190,304,278,146,37,2},{5,59,185,302,282,150,39,2},{4,56,182,301,285,154,40,2},{4,54,1877,29 3},{3,51,172,296,290,163,46,3},{3,48,167,294,294,167,48,3},{3,46,163,290,296,172,51,3},{3,43,158,287,298,177,54,4},{2,40,254,528 4},{2,39,150,282,302,185,59,5},{2,37,146,278,304,190,62,5},{1,34,141,275,307,195,65,6},{1,32,136,271,309,200,68,7},{1,3,2,3,79,30,133, 7},{1,28,129,265,311,208,74,8},{1,27,125,262,312,211,77,9},{1,25,120,257,313,217,81,10},{1,24,117,253,314,220,84,11},{1,22,11 12}, {1, 20, 109, 246, 315, 229, 91, 13}, {1, 19, 105, 242, 315, 233, 95, 14};

Two groups: {7,92,245,339,245,91,6,-1},{6,87,239,339,250,96,8,-1},{5,83,236,339,255,99,8,-1},{4,80,231,338,259,103,10,- 1},{3,75,226,338,265,108,11,-2},{1,70,220,337,271,114,13,-2},{1,67,217,337,275,117,13,-3},{0,64,211,335,279,122,16,-3},{-1 ,61,206,333,283,127,18,-3},{-1,57,202,332,287,131,19,-3},{-2,53,196,330,292,137,21,-3},{-3,49,190,328,297,142,24,,326,47,1 ,25,-4},{-3,43,181,324,305,151,27,-4},{-3,41,177,321,308,155,29,-4},{-4,37,171,319,312,161,32,-4},{-4,35,166,315,315,166,35 ,-4},{-4,32,161,312,319,171,37,-4},{-4,29,155,308,321,177,41,-3},{-4,27,151,305,324,181,43,-3},{-4,25,146,300,326,187,47,- 3},{-3,24,142,297,328,190,49,-3},{-3,21,137,292,330,196,53,-2},{-3,19,131,287,332,202,57,-1},{-3,18,127,283,333,206,61,-1} ,{-3,16,122,279,335,211,64,0},{-3,13,117,275,337,217,67,1},{-2,13,114,271,337,220,70,1},{-2,11,108,215,338,226,75,3},{-1,10,23,1 ,80,4},{-1,8,99,255,339,236,83,5},{-1,8,96,250,339,239,87,6};

Three groups: {0,42,125,178,125,42,0,0},{-1,40,121,177,128,46,1,0},{-1,37,118,177,130,48,3,0},{-1,36,116,176,133,50,3 ,-1},{-1,34,114,176,135,52,3,-1},{-2,32,110,175,138,56,4,-1},{-3,30,107,174,141,59,5,-1},{-3 ,29,106,173,142,60,6,-1},{-3,28,104,172,143,62,7,-1},{-3,25,101,172,146,64,8,-1},{-4,23,97,171,149,68,9 ,-1},{-4,21,94,170,152,71,10,-2},{-4,19,91,168,156,74,11,-3},{-3,19,91,167,155,74,12,- 3},{-3,20,90,165,155,75,13,-3},{-3,18,87,163,157,79,14,-3},{-3,16,83,160,160,83,16,-3} ,{-3,14,79,157,163,87,18,-3},{-3,13,75,155,165,90,20,-3},{-3,12,74,155,167,91,19,-3},{ -3,11,74,156,168,91,19,-4},{-2,10,71,152,170,94,21,-4},{-1,9,68,149,171,97,23,-4},{-1 ,8,64,146,172,101,25,-3},{-1,7,62,143,172,104,28,-3},{-1,6,60,142,173,106,29,-3},{-1,5,59,141,174,107,30,- 3},{-1,4,56,138,175,110,32,-2},{-1,3,52,135,176,114,34,-1},{-1,3,50,133,176,116,36, -1},{0,3, 48,130,177,118,37,-1},{0,1,46,128,177,121,40,-1};

Four groups: 21,105,235,302,235,105,21,0 {}, {} 21,102,231,300,237,108,24,1, 20,100,227,298,239,113,26,1 {}, {} 19,97,223,296,241,117,29,2, 19,94,218,294,243,121,32,3 {}, {18, 92,214,292,246,125,34,3},{17,89,210,290,248,129,37,4},{17,86,206,288,250,133,39,5},{16,84,202,286,252,137,42,5},{15,8,1,141,98,28 78,194,282,256,145,47,7},{14,76,190,280,258,149,50,7},{13,73,186,277,260,154,53,8},{12,71,182,275,262,158,55,9},{12,6,162,58,273,162,58,27 65,174,271,267,166,60,10},{10,63,170,269,269,170,63,10},{10,60,166,267,271,174,65,11},{9,58,162,264,273,17,8,68,12},{9,52}{2,78,27,2} 53,154,260,277,186,73,13},{7,50,149,258,280,190,76,14},{7,47,145,256,282,194,78,15},{6,45,141,254,284,198,81,15},{5,42,13,87,25 39,133,250,288,206,86,17},{4,37,129,248,290,210,89,17},{3,34,125,246,292,214,92,18},{3,32,121,243,294,218,94,19},{2,2,9,11,7,24 26,113,239,298,227,100,20},{1,24,108,237,300,231,102,21}.

g), the prediction block of the current coding unit is accurately corrected for the predicted pixels, and the accurate correction process can be based on the position where the current predicted pixels are located, and the horizontal distance HD and the vertical distance VD are used as offsets to obtain the adjacent UPNum on the upper side. Reconstruction reference pixels and LFNum reconstruction reference pixels on the left, filter the obtained reference pixels and the current prediction pixel, or perform weighted summation to obtain the accurate prediction pixel as the final prediction pixel at the current position in the prediction block. The above HD and VD can be 0, 1, 2, 3, ..., N, etc., and N does not exceed the width or height of the current block. The above UPNum and LFNum can be 0, 1, 2, 3, ..., M, etc., and M does not exceed the number of adjacent reconstruction reference pixels on the upper side of the current block and the number of adjacent reconstruction reference pixels on the left side;

h), calculate the rate-distortion cost information of each prediction mode according to the final predicted pixel and original pixel obtained by the above-mentioned IPS technology, and record the prediction mode and the corresponding cost value of the minimum rate-distortion cost information;

i), if the IPF allows the identification bit to be '0' and the IPS allows the identification bit to be '0', then transmit the prediction mode index recorded in b) to the decoding end through the code stream; if the IPF allows the identification bit to be '1' and If the IPS allows the flag bit to be '0', compare the minimum cost value recorded in b) with the minimum cost value recorded in e), and if the rate-distortion cost in b) is smaller, compare the prediction recorded in b) The mode index coding is used as the optimal prediction mode of the current coding unit and is transmitted to the decoding end through the code stream, and the IPF current coding unit identification position is '0', indicating that the IPF technology is not used, and is also transmitted to the decoding end through the code stream; if e) in The rate distortion is smaller, then the prediction mode index code recorded in e) is transmitted to the decoding end as the optimal prediction mode of the current coding unit through the code stream, and the IPF current coding unit identification position '1' indicates that the IPF technology is used, It is also transmitted to the decoding end through the code stream; if the IPF allows the identification bit to be '0' and the IPS allows the identification bit to be '1', then compare the minimum cost value recorded in b) with the minimum cost value recorded in h), If the rate-distortion cost in b) is smaller, then use the prediction mode index encoding recorded in b) as the optimal prediction mode of the current coding unit and transmit it to the decoding end through the code stream, and use the IPS identifier position '0 of the current coding unit ', indicating that this technology is not used, and is also transmitted to the decoding end through the code stream; if the rate-distortion in h) is smaller, the prediction mode recorded in h) is used as the optimal prediction mode of the current coding unit through the code stream It is transmitted to the decoding end, and the IPS use flag position of the current coding unit is '1', indicating that this technology is used, and it is also transmitted to the decoding end through the code stream; if the IPF allowable flag bit is '1' and the IPS allowable flag bit is '1' , then compare the minimum cost values recorded in b), e) and h), and if the rate-distortion cost in b) is smaller, encode the prediction mode index recorded in b) as the optimal prediction for the current coding unit The mode is transmitted to the decoding end through the code stream, and the IPS use identification bit and the IPF use identification position of the current coding unit are '0', indicating that they are not used, and are also transmitted to the decoding end through the code stream; if the rate distortion in e) is smaller , then the prediction mode index code recorded in e) is transmitted to the decoding end as the optimal prediction mode of the current coding unit through the code stream, and the IPF of the current coding unit uses the identification position '1 ' and does not transmit the IPS to use the identification position, indicating Using IPF technology instead of IPS technology, it is also transmitted to the decoding end through the code stream; if the rate-distortion in h) is smaller, the prediction mode recorded in h) is used as the optimal prediction mode of the current coding unit. The stream is transmitted to the decoding end, and the IPF use identification position of the current coding unit is '0' and the IPS use identification position is '1', indicating that the IPF technology is not used but the IPS technology is used, and is also transmitted to the decoding end through the code stream.

Finally, the encoder superimposes the prediction block and the inversely transformed and inversely quantized residuals to obtain a reconstructed coding unit block, which is used as a prediction reference block for the next coding unit.

It can be understood that the encoder adopts intra-frame prediction smoothing filtering technology, which strengthens the spatial correlation of intra-frame prediction, and multiple reference pixels in a neighboring reference pixel set are smoothed and filtered, avoiding the possibility of over-sharp prediction. It will cause the residual error to increase and become larger, the prediction effect is not good, and the coding efficiency is improved.

An embodiment of the present application provides an intra-frame prediction method, which is applied to a video decoding device, that is, a decoder. The functions implemented by the method can be implemented by the processor in the video decoding device calling program codes, and of course the program codes can be stored in a computer storage medium. It can be seen that the video encoding device includes at least a processor and a storage medium.

FIG. 11 is a schematic flowchart of an implementation of an intra-frame prediction method according to an embodiment of the present application. As shown in FIG. 11 , the method includes:

S301. Acquire a code stream, parse the header information of the code stream, and obtain a filter identifier.

S302. When the filtering flag indicates that the intra-frame prediction filtering permission flag is valid and the intra-frame prediction smoothing filtering permission flag is valid, obtain at least the current prediction mode and the intra-frame prediction filtering use flag from the code stream.

S303, when the intra-frame prediction filtering use flag is invalid and the valid intra-frame prediction smoothing filtering use flag is obtained from the code stream, based on the current prediction mode, perform intra-frame prediction smoothing filtering processing on the current block to obtain the first prediction value.

In this embodiment of the present application, after acquiring the code stream information, the decoder can parse from the header information of the code stream to determine whether the IPF and IPS processing is performed. That is, the intra-frame prediction filtering permission flag in the filtering flags, and/or the intra-frame prediction smoothing filtering permission flag can be parsed out. When IPF and IPS are employed, the filtering flags are indicative of the Intra prediction filtering permission flag being valid, and the intra prediction smoothing filtering permission flag being valid. At this time, the decoder can obtain at least the current prediction mode and the intra-frame prediction filtering use flag from the code stream. When the intra-frame prediction filtering use flag is invalid, and the valid intra-frame prediction smoothing filter use flag is obtained from the code stream When identifying, that is, when the current block adopts the current prediction mode determined by the IPS technology, based on the current prediction mode, perform intra-frame prediction smoothing filtering on the current block to obtain the first prediction value, and then based on the first prediction value. The predicted value goes through the subsequent decoding process.

In some embodiments of the present application, when the intra prediction filtering use flag is valid, the current prediction mode is obtained from the code stream; based on the current prediction mode, intra prediction filtering is performed on the current block to obtain the second prediction value.

In some embodiments of the present application, the decoder performs intra-frame prediction smoothing filtering processing on the current block based on the preset horizontal distance and the preset vertical distance corresponding to each pixel in the current block to obtain the first predicted value.

In some embodiments of the present application, at least two adjacent sets of at least two adjacent reference pixels corresponding to each pixel are determined based on a preset horizontal distance and a preset vertical distance corresponding to each pixel in the current block reference block range;

From the at least two adjacent reference pixel sets, correspondingly determine at least two target reference pixel sets within the scope of at least two adjacent reference blocks; wherein, each target reference pixel set includes at least one target reference pixel;

Based on at least two target reference pixel sets, intra-frame prediction smoothing filtering is performed on each pixel to obtain a first predicted value.

In some embodiments of the present application, the decoder correspondingly determines, from the at least two adjacent reference pixel sets, at least two target reference pixel sets belonging to the range of the at least two adjacent reference blocks, including the following manners :

Manner 1: From the at least two adjacent reference pixel sets, determine the pixels at the two boundary pixel positions of each adjacent reference block range as a target reference pixel set, so as to determine the at least two target reference pixels gather.

Mode 2: From the at least two adjacent reference pixel sets, determine M pixels at any pixel position within the range of each adjacent reference block as a target reference pixel set, so as to determine the at least two target reference pixel sets. A set of pixels, where M is greater than 0 and does not exceed the maximum number of pixels in the corresponding adjacent reference pixel set.

In some embodiments of the present application, the at least two adjacent reference pixel sets include: a left adjacent reference pixel set and an upper adjacent reference pixel set;

The at least two adjacent reference block ranges include: a left adjacent reference pixel range and an upper adjacent reference pixel range;

The at least two target reference pixel sets include: a left target adjacent reference pixel set and an upper target adjacent reference pixel set.

In some embodiments of the present application, the preset horizontal distance is less than or equal to the width of the initial prediction block; the preset vertical distance is less than or equal to the height of the initial prediction block;

The preset horizontal distance and preset vertical distance corresponding to the first type of pixels in the position of n rows and m columns of the initial prediction block, and the preset horizontal distances and pre-set vertical distances corresponding to other types of pixels except n rows and m columns of the initial prediction block. Let the vertical distances be different; n and m are natural numbers;

The preset horizontal distance and the preset vertical distance corresponding to each pixel in the first type of pixels are at least partially the same, or both are different;

The preset horizontal distance and the preset vertical distance corresponding to each pixel in the other types of pixels are at least partially the same, or both are different.

In some embodiments of the present application, performing an intra-frame prediction smoothing filtering process on each pixel based on at least two target reference pixel sets to obtain a second type of prediction value, including: a decoder determining a filter coefficient; and based on at least two The target refers to the set of pixels and filter coefficients, and performs intra-frame prediction smoothing filtering processing on each pixel to obtain the second type of prediction value.

In some embodiments of the present application, the filter coefficients corresponding to the first type of pixels at the positions of n rows and m columns in the initial prediction block are different from the filter coefficients corresponding to other types of pixels except for n rows and m columns in the initial prediction block; n and m are natural numbers;

The filter coefficients corresponding to each pixel in the first type of pixels are at least partially the same, or all different;

The filter coefficients corresponding to each pixel in the other types of pixels are at least partially the same, or all different.

In some embodiments of the present application, when the preset horizontal distance is greater than the width of the initial prediction block, the leftmost reference pixel in the upper adjacent reference pixel set is used to pad to the left until the upper adjacent reference is determined until the first boundary of the block range, and/or, using the rightmost reference pixel in the upper adjacent reference pixel set to fill to the right, until the second boundary of the upper adjacent reference block range is determined;

When the preset vertical distance is less than or equal to the height of the initial prediction block, the lowermost reference pixel in the set of adjacent reference pixels on the left side is used to fill down until the third boundary limit of the range of the adjacent reference block on the left side is determined, And/or, the uppermost reference pixel in the set of adjacent reference pixels on the left side is used to fill upwards until the fourth boundary limit of the range of the adjacent reference block on the left side is determined.

It should be noted that, in this embodiment of the present application, the principle of the decoder implementing filtering processing is the same as that of the encoder, and details are not repeated here.

Exemplarily, the decoder obtains the code stream, and parses the code stream to obtain the IPF permission flag and the IPS permission flag of the current video sequence.

In the intra-frame prediction decoding process, if the allowable flag bit of the IPF and the allowable flag bit of the IPS are both '1', perform all the following steps; if the allowable flag bit of the IPF is '1' and the allowable flag bit of the IPS is ' 0', only perform steps a), b), c), d) and g); if the allowable flag bit of IPF is '0' and the allowable flag bit of IPS is '1', then only a), b ), e), f) and g) steps; if the allowable identification bit of IPF and the allowable identification bit of IPS are both '0', then only steps a), b) and g) are performed. If the above-mentioned IPF or IPS allows the flag bit to be '0', the decoding end defaults to the use flag bit of the corresponding technology of all decoding blocks to be '0':

a), obtain the code stream and decode to obtain residual information, and obtain time-domain residual information through processes such as inverse transformation and inverse quantization;

b), parse the code stream to obtain the prediction mode of the current decoding unit, and calculate the prediction block according to the prediction mode of the current decoding unit and the adjacent reconstruction block;

c), parse and obtain the use identification bit of IPF, if the use identification bit of IPF is '0', then do not perform additional operations on the current prediction block, skip step d); if the use identification bit of IPF is '1', then execute d);

d), select the corresponding filter according to the prediction mode classification information of the current decoding unit, select the corresponding filter coefficient group according to the size of the current decoding unit, then filter each pixel in the prediction block to obtain the prediction block;

e), obtain the use identification bit of IPF, if the use identification bit of IPF is '1', then skip the remaining process of this step and skip step f); if the use identification bit of IPF is '0', then parse and obtain IPS usage flag. If the use flag bit of IPS is '0', then no additional operation is performed on the current prediction block, and step f) is skipped; if the use flag bit of IPS is '1', then f) is executed;

f), the prediction block of the current coding unit is accurately corrected for the predicted pixels. The accurate correction process can be based on the position where the current predicted pixels are located, and the horizontal distance HD and the vertical distance VD are used as offsets to obtain the adjacent UPNum on the upper side. Reconstruction reference pixels and LFNum reconstruction reference pixels on the left, filter the obtained reference pixels and the current prediction pixel, or perform weighted summation to obtain the accurate prediction pixel as the final prediction pixel at the current position in the prediction block. The above HD and VD can be 0, 1, 2, 3, ..., N, etc., and N does not exceed the width or height of the current block. The above UPNum and LFNum can be 0, 1, 2, 3, ..., M, etc., and M does not exceed the number of adjacent reconstruction reference pixels on the upper side of the current block and the number of adjacent reconstruction reference pixels on the left side;

g), superimposing the residual information restored in step a) of the prediction block to obtain the reconstructed block of the current decoding unit, which is output after post-processing.

Exemplarily, the implementation on the decoder side may also adopt the following implementation manners, which are not limited in this embodiment of the present application. As follows:

The decoder obtains the code stream, and parses the code stream to obtain the IPF allowable flag bit and the IPS allowable flag bit of the current video sequence.

In the intra-frame prediction decoding process, if the allowable flag bit of the IPF and the allowable flag bit of the IPS are both '1', perform all the following steps; if the allowable flag bit of the IPF is '1' and the allowable flag bit of the IPS is ' 0', then only execute steps a), b), and d); if the allowable identification bit of IPF is '0' and the allowable identification bit of IPS is '1', then only steps a), c) and d) are executed ; If the allowable identification bit of IPF and the allowable identification bit of IPS are both '0', then only steps a) and d) are performed. If the above-mentioned IPF or IPS allows the flag bit to be '0', the decoding end defaults to the use flag bit of the corresponding technology of all decoding blocks to be '0':

a), obtain the code stream and decode to obtain the prediction mode information and residual information of the current decoding unit, and obtain the time domain residual information through processes such as inverse transformation and inverse quantization;

b), parse the code stream to obtain the use flag of the IPF of the current decoding unit;

c), if the IPF use identification bit of the current decoding unit is '0', then the use identification bit of the IPS is obtained by parsing the code stream; if the IPF use identification bit of the current decoding unit is '1', then the use identification bit of the IPS is not parsed ;

d), if the use identification bit of IPF is '0' and the use identification bit of IPS is '0', then calculate the predicted block according to the current prediction mode, skip all remaining steps; if the use identification bit of IPF is '1' , then execute e); if the use flag of the IPS is '1', execute f);

e), select the corresponding filter according to the prediction mode classification information of the current decoding unit, select the corresponding filter coefficient group according to the size of the current decoding unit, then filter each pixel in the prediction block to obtain the prediction block, skip through all remaining steps;

f), according to the prediction mode parsed by the current decoding unit, the process can use an interpolation filter that is different from the generation of predicted pixels in steps d) and e), and can use an 8-tap interpolation filter plus use the location of each predicted pixel. Several reconstructed reference pixels obtained based on the position of the pixel are used as interpolation filters for generating predicted pixels, and the specific process of obtaining the above-mentioned several reconstructed reference pixels is: taking the horizontal distance HD and the vertical distance VD as offsets, obtaining the adjacent UPNum on the upper side. Reconstructed reference pixels and left LFNum reconstructed reference pixels. The above HD and VD can be 0, 1, 2, 3, ..., N, etc., and N does not exceed the width or height of the current block. The above UPNum and LFNum can be 0, 1, 2, 3, ..., M, etc., and M does not exceed the number of adjacent reconstruction reference pixels on the upper side of the current block and the number of adjacent reconstruction reference pixels on the left side;

It can be understood that the decoder implements the process of filtering the current block by using one adjacent reference pixel in each adjacent reference pixel set in at least two adjacent reference pixel sets by using intra-frame prediction filtering. Using intra-frame prediction smoothing filtering can refer to multiple adjacent reference pixels in one adjacent reference pixel set, so such prediction will be smoother and the prediction for blurred images will be more accurate, thereby improving decoding efficiency. .

As shown in FIG. 12, an embodiment of the present application provides an encoder 1, including:

The first determination part 10 is configured to traverse the intra prediction mode, and determine the initial prediction value of the initial prediction block corresponding to the current block;

The first filtering part 11 is configured to perform intra-frame prediction filtering and intra-frame prediction smoothing filtering respectively on the initial prediction block to obtain a first-type prediction value and a second-type prediction value; wherein, the intra-frame prediction smoothing Filtering is a process of filtering the current block by using a plurality of adjacent reference pixels in each adjacent reference pixel set in at least two adjacent reference pixel sets;

The first determining part 10 is further configured to use the initial prediction value, the first type prediction value and the second type prediction value to calculate the rate-distortion cost with the original pixel value of the current block, respectively, to determine the optimal rate-distortion cost. the corresponding current prediction mode;

a first prediction part 12, configured to use the current prediction mode to perform intra-frame prediction on the current block;

The writing part 13 is configured to write the index information of the current prediction mode and the filter flag into the code stream, wherein the filter flag represents the flag corresponding to intra-frame prediction filtering and/or intra-frame prediction smoothing filtering.

In some embodiments of the present application, the first filtering part 11 is further configured to, based on the adjacent reference pixels on the left and the adjacent reference pixels on the upper side of each pixel in the initial prediction block, filter Perform intra-frame prediction filtering on each pixel of the initial prediction block to obtain the first type of prediction value; and based on the preset horizontal distance and preset vertical distance corresponding to each pixel in the initial prediction block, determine the The initial prediction block is subjected to intra-frame prediction smoothing filtering to obtain the second type of prediction value.

In some embodiments of the present application, the first filtering part 11 is further configured to, based on the preset horizontal distance and the preset vertical distance corresponding to each pixel in the initial prediction block, determine the at least two adjacent reference block ranges of at least two adjacent reference pixel sets corresponding to each pixel; from the at least two adjacent reference pixel sets, correspondingly determined to belong to the at least two adjacent reference pixel sets At least two target reference pixel sets within the block range; wherein, each target reference pixel set includes at least one target reference pixel; based on the at least two target reference pixel sets, intra prediction smoothing is performed on each pixel filtering to obtain the predicted value of the second type.

In some embodiments of the present application, the first filtering part 11 is further configured to determine, from the at least two adjacent reference pixel sets, pixels at two boundary pixel positions of each adjacent reference block range is one target reference pixel set, thereby determining the at least two target reference pixel sets.

In some embodiments of the present application, the first filtering part 11 is further configured to determine, from the at least two adjacent reference pixel sets, M arbitrary pixel positions within the range of each adjacent reference block The pixel is one target reference pixel set, so that the at least two target reference pixel sets are determined, wherein M is greater than 0 and does not exceed the maximum number of pixels of the corresponding adjacent reference pixel sets.

The preset horizontal distance and the preset vertical distance corresponding to the first type of pixels in the position of n rows and m columns of the initial prediction block are the same as those corresponding to other types of pixels except n rows and m columns of the initial prediction block. The preset horizontal distance is different from the preset vertical distance; n and m are natural numbers;

The preset horizontal distance corresponding to each pixel in the first type of pixels is at least partially the same as the preset vertical distance, or both are different;

The preset horizontal distance and the preset vertical distance corresponding to each of the other types of pixels are at least partially the same, or both are different.

In some embodiments of the present application, the first filtering part 11 is further configured to determine a filter coefficient; based on the at least two target reference pixel sets and the filter coefficient, perform intra-frame filtering on each pixel A prediction smoothing filtering process is performed to obtain the second type of prediction value.

In some embodiments of the present application, the filter coefficients corresponding to the first type of pixels in the positions of n rows and m columns of the initial prediction block are corresponding to other types of pixels except for n rows and m columns of the initial prediction block. Different filter coefficients; n and m are natural numbers;

In some embodiments of the present application, the first determining part 10 is further configured to use the initial prediction value, the first type prediction value and the second type prediction value to perform rate-distortion costing on the original pixel values of the current block, respectively. Calculate to determine a first minimum rate-distortion cost corresponding to the initial predicted value, a second minimum rate-distortion cost corresponding to the first type of prediction value, and a third minimum rate-distortion cost corresponding to the second type of prediction value Distortion cost; from the first minimum rate-distortion cost, the second minimum rate-distortion cost, and the third minimum rate-distortion cost, determine the minimum optimal rate-distortion cost; the optimal rate-distortion cost will be obtained The cost corresponds to the prediction mode as the current prediction mode.

In some embodiments of the present application, the filtering flags include: intra-frame prediction filtering permission flags, intra-frame prediction filtering use flags, intra-frame prediction smoothing filtering permission flags, and intra-frame prediction smoothing filtering use flags;

The writing part 13 is further configured to write the index information of the current prediction mode into the code stream; determine that the intra-frame prediction filtering permission flag is valid, and the intra-frame prediction smoothing filtering permission flag is valid, and writing the intra-frame prediction filtering permission flag and the intra-frame prediction smoothing filtering permission flag into the code stream;

When the optimal rate-distortion cost is the second minimum rate-distortion cost, determining that the intra-frame prediction filtering use flag is valid, and writing the intra-frame prediction filtering use flag into the code stream;

When the optimal rate-distortion cost is the third minimum rate-distortion cost, it is determined that the intra-frame prediction filtering use flag is invalid and the intra-frame prediction smoothing filter use flag is valid, and the intra-frame prediction filtering use flag is valid. The prediction filtering uses the flag and the intra-frame prediction smoothing filtering uses the flag to write the code stream;

When the optimal rate-distortion cost is the first minimum rate-distortion cost, it is determined that the intra-frame prediction filtering use flag is invalid, and the intra-frame prediction smoothing filter use flag is invalid, and the intra-frame prediction filtering use flag is invalid The predictive filtering uses the flag and the intra prediction smoothing filtering uses the flag to write the code stream.

In some embodiments of the present application, the first filtering part 11 is further configured to traverse the intra-frame prediction mode, and after determining the initial prediction value of the initial prediction block corresponding to the current block, use the current block Prediction mode, before performing intra-frame prediction on the current block, perform intra-frame prediction smoothing filtering processing on the initial prediction block respectively, to obtain the second type of prediction value; wherein, the intra-frame prediction smoothing filter adopts at least two A process of filtering the current block for multiple adjacent reference pixels in each adjacent reference pixel set in the adjacent reference pixel set;

The first determining part 10 is further configured to use the initial predicted value and the second type of predicted value to perform rate-distortion cost calculation with the original pixel value of the current block, respectively, to determine the rate-distortion cost corresponding to the optimal rate-distortion cost. the current prediction mode.

In some embodiments of the present application, the first filtering part 11 is further configured to traverse the intra-frame prediction mode, and after determining the initial prediction value of the initial prediction block corresponding to the current block, use the current block prediction mode, before performing intra prediction on the current block, determine a prediction mode set according to the type of intra prediction mode, and the prediction mode set is a prediction mode enabling intra prediction smoothing filtering technology; for the initial prediction In the value, the initial sub-prediction value predicted by the prediction mode set is used to perform intra-frame prediction smoothing filtering processing to obtain the second type of prediction value;

In some embodiments of the present application, the filtering flags include: an intra-frame prediction smoothing filtering permission flag and an intra-frame prediction smoothing filtering use flag;

The writing part 13 is further configured to write the index information of the current prediction mode into the code stream; determine that the intra-frame prediction smoothing filtering permission flag is valid, and write the intra-frame prediction smoothing filtering permission flag incoming code stream;

When the optimal rate-distortion cost is the third minimum rate-distortion cost, it is determined that the intra-frame prediction smoothing filter usage flag is valid, and the intra-frame prediction smoothing filter usage flag is written into the code stream.

When the optimal rate-distortion cost is the first minimum rate-distortion cost, it is determined that the intra-frame prediction smoothing filter usage flag is invalid, and the intra-frame prediction smoothing filter usage flag is written into the code stream.

In some embodiments of the present application, the encoding type of the initial prediction value corresponds to multiple encoding channels;

The filtering identifiers corresponding to each encoding channel use the same identifier, or use one-to-one corresponding identifiers.

In some embodiments of the present application, the first filtering part 11 is further configured to allow the use of intra-frame prediction smoothing to perform filtering when the size of the initial prediction block satisfies the prediction area threshold range; When the size is greater than the upper limit of the prediction area threshold range, the initial prediction block is divided to obtain prediction sub-blocks satisfying the prediction area threshold range; the intra prediction processing flow is implemented for the prediction sub-block.

In some embodiments of the present application, the first filtering part 11 is further configured to traverse the intra-frame prediction mode, and after determining the initial prediction value of the initial prediction block corresponding to the current block, use the initial prediction value , the first type of prediction value and the second type of prediction value, respectively carry out rate-distortion cost calculation with the original pixel value of the current block, and before determining the current prediction mode corresponding to the optimal rate-distortion cost, the initial prediction block is respectively framed The intra-prediction filtering is performed to obtain the first type of prediction value; the intra-frame prediction mode is traversed, and the intra-frame prediction smoothing filtering process is used to predict and estimate the current block to obtain the second type of prediction value.

In some embodiments of the present application, the first filtering part 11 is further configured to traverse the intra prediction mode, and based on the preset horizontal distance and the preset vertical distance corresponding to each pixel in the current block, The current block is subjected to intra-frame prediction smoothing filtering to obtain the second type of prediction value.

In some embodiments of the present application, the first determining part 10 is further configured to, when the preset horizontal distance is greater than the width of the initial prediction block, use the leftmost reference pixel set in the upper adjacent reference pixel set The reference pixels on the upper side are filled to the left until the first boundary limit of the range of the upper adjacent reference block is determined, and/or, the rightmost reference pixel in the upper adjacent reference pixel set is used to fill to the right , until the second boundary limit of the upper adjacent reference block range is determined;

When the preset vertical distance is less than or equal to the height of the initial prediction block, the lowermost reference pixel in the set of adjacent reference pixels on the left side is used to fill down, until the range of the adjacent reference block on the left side is determined. Until the third boundary limit, and/or, use the uppermost reference pixel in the left adjacent reference pixel set to fill upwards until the fourth boundary limit of the left adjacent reference block range is determined.

In practical applications, as shown in FIG. 13 , an embodiment of the present application further provides an encoder, including:

a first memory 15 and a first processor 14;

The first memory 15 stores a computer program that can be run on the first processor 14 , and the first processor 14 executes the program for an intra-frame prediction method on the encoder side.

As shown in FIG. 14, an embodiment of the present application provides a decoder 2, including:

The parsing part 20 is configured to obtain the code stream, parse the header information of the code stream, and obtain the filtering identification; and when the filtering identification indicates that the intra-frame prediction filtering permission flag is valid, and the intra-frame prediction smoothing filtering permission flag is valid , at least obtain the current prediction mode and intra prediction filtering use identifier from the code stream;

The second prediction part 21 is configured to, based on the current prediction mode, perform a prediction on the current block when the intra-frame prediction filtering use flag is invalid and a valid intra-frame prediction smoothing filter use flag is obtained from the code stream. The intra-frame prediction is smoothed and filtered to obtain the first predicted value.

In some embodiments of the present application, the parsing part 20 is further configured to at least acquire the current prediction mode and the intra-frame prediction filtering use flag from the code stream, when the intra-frame prediction filtering use flag is valid When , obtain the current prediction mode from the code stream;

The second prediction part 21 is further configured to perform intra-frame prediction filtering processing on the current block based on the current prediction mode to obtain a second prediction value.

In some embodiments of the present application, the second prediction part 21 is further configured to frame the current block based on a preset horizontal distance and a preset vertical distance corresponding to each pixel in the current block The intra-prediction smoothing filtering process is performed to obtain the first predicted value.

In some embodiments of the present application, the second prediction part 21 is further configured to, based on the preset horizontal distance and the preset vertical distance corresponding to each pixel in the current block, determine a at least two adjacent reference block ranges of at least two adjacent reference pixel sets corresponding to each pixel;

From the at least two adjacent reference pixel sets, at least two target reference pixel sets belonging to the at least two adjacent reference blocks are correspondingly determined; wherein, each target reference pixel set includes at least one target reference pixel;

Based on the at least two target reference pixel sets, intra-frame prediction smoothing filtering is performed on each pixel to obtain the first predicted value.

In some embodiments of the present application, the second prediction part 21 is further configured to determine, from the at least two adjacent reference pixel sets, pixels at two boundary pixel positions of each adjacent reference block range is one target reference pixel set, thereby determining the at least two target reference pixel sets.

In some embodiments of the present application, the second prediction part 21 is further configured to determine, from the at least two adjacent reference pixel sets, M arbitrary pixel positions within the range of each adjacent reference block The pixel is one target reference pixel set, so that the at least two target reference pixel sets are determined, wherein M is greater than 0 and does not exceed the maximum number of pixels of the corresponding adjacent reference pixel sets.

In some embodiments of the present application, the second prediction part 21 is further configured to determine filter coefficients; based on the at least two target reference pixel sets and the filter coefficients, perform intra-frame performance on each pixel A prediction smoothing filtering process is performed to obtain the second type of prediction value.

In some embodiments of the present application, when the preset horizontal distance is greater than the width of the initial prediction block, the leftmost reference pixel in the set of adjacent reference pixels on the upper side is used to pad to the left, until it is determined that the upper until the first boundary of the range of the adjacent reference blocks on the side, and/or, use the rightmost reference pixel in the set of adjacent reference pixels on the upper side to fill to the right, until the range of the adjacent reference blocks on the upper side is determined. until the second boundary;

It can be understood that the decoder implements the process of filtering the current block by using one adjacent reference pixel in each adjacent reference pixel set in at least two adjacent reference pixel sets by using intra-frame prediction filtering. Using intra-frame prediction smoothing filtering can refer to multiple adjacent reference pixels in a set of adjacent reference pixels, so such prediction will be smoother, and the prediction for blurred images will be more accurate, thereby improving decoding efficiency. .

In practical applications, as shown in FIG. 15 , an embodiment of the present application further provides a decoder, including:

a second memory 22 and a second processor 23;

The second memory 22 stores a computer program that can be executed on the second processor 23, and the second processor 23 implements the intra-frame prediction method on the decoder side when executing the program.

Correspondingly, an embodiment of the present application provides a storage medium on which a computer program is stored. When the computer program is executed by the first processor, an intra-frame prediction method of the encoder is implemented; or, the computer program is processed by the second processor. When the decoder executes, implements the decoder's intra prediction method.

The above is only the embodiment of the present application, but the protection scope of the present application is not limited to this. Covered within the scope of protection of this application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Industrial Applicability

In the embodiment of the present application, the process of filtering the current block by using one adjacent reference pixel in each adjacent reference pixel set in at least two adjacent reference pixel sets is implemented by using intra-frame prediction filtering. Intra-frame prediction smoothing filtering can refer to multiple adjacent reference pixels in one adjacent reference pixel set, so such prediction will be smoother, and the prediction for blurred images will be more accurate, thereby improving coding efficiency. .

Claims

An intra-frame prediction method, applied to an encoder, comprising:

Traverse the intra-frame prediction mode to determine the initial prediction value of the initial prediction block corresponding to the current block;

Perform intra-frame prediction filtering and intra-frame prediction smoothing filtering on the initial prediction block respectively, to obtain a first-type prediction value and a second-type prediction value; wherein, the intra-frame prediction smoothing filtering adopts at least two adjacent reference values A process of filtering the current block for multiple adjacent reference pixels in each adjacent reference pixel set in the pixel set;

Using the initial predicted value, the first-type predicted value and the second-type predicted value, respectively, perform rate-distortion cost calculation with the original pixel value of the current block, and determine the current prediction mode corresponding to the optimal rate-distortion cost;

using the current prediction mode to perform intra prediction on the current block;

The index information of the current prediction mode and the filter identifier are written into the code stream, where the filter identifier represents an identifier corresponding to intra-frame prediction filtering and/or intra-frame prediction smoothing filtering.
The method according to claim 1, wherein the performing intra-frame prediction filtering and intra-frame prediction smoothing filtering processing on the initial prediction block respectively to obtain the first-type prediction value and the second-type prediction value, comprising:

Based on the left adjacent reference pixel point and the upper adjacent reference pixel point of each pixel in the initial prediction block, perform intra-frame prediction filtering on each pixel of the initial prediction block to obtain the first type Predictive value;

Based on the preset horizontal distance and the preset vertical distance corresponding to each pixel in the initial prediction block, intra-frame prediction smoothing filtering is performed on the initial prediction block to obtain the second type of prediction value.
The method according to claim 2, wherein, based on the preset horizontal distance and preset vertical distance corresponding to each pixel in the initial prediction block, performing intra prediction smoothing filtering processing on the initial prediction block, Obtain the second type of predicted value, including:

Based on the preset horizontal distance and the preset vertical distance corresponding to each pixel in the initial prediction block, at least two adjacent sets of at least two adjacent reference pixels corresponding to each pixel are determined reference block range;

From the at least two adjacent reference pixel sets, at least two target reference pixel sets belonging to the at least two adjacent reference blocks are correspondingly determined; wherein each target reference pixel set includes at least one target reference pixel;

Based on the at least two target reference pixel sets, intra-frame prediction smoothing filtering is performed on each pixel to obtain the second type of prediction value.
The method according to claim 3, wherein the at least two target reference pixel sets belonging to the range of the at least two adjacent reference blocks are correspondingly determined from the at least two adjacent reference pixel sets, include:

From the at least two adjacent reference pixel sets, pixels at two boundary pixel positions of each adjacent reference block range are determined as a target reference pixel set, thereby determining the at least two target reference pixel sets.
The method according to claim 3, wherein the at least two target reference pixel sets belonging to the range of the at least two adjacent reference blocks are correspondingly determined from the at least two adjacent reference pixel sets, include:

From the at least two adjacent reference pixel sets, M pixels at any pixel position within the range of each adjacent reference block are determined as a target reference pixel set, thereby determining the at least two target reference pixel sets, Wherein, M is greater than 0 and does not exceed the maximum number of pixels in the corresponding adjacent reference pixel set.
The method according to any one of claims 3 to 5, wherein,

The at least two adjacent reference pixel sets include: a left adjacent reference pixel set and an upper adjacent reference pixel set;

The at least two adjacent reference block ranges include: a left adjacent reference pixel range and an upper adjacent reference pixel range;

The at least two target reference pixel sets include: a left target adjacent reference pixel set and an upper target adjacent reference pixel set.
The method according to any one of claims 3 to 6, wherein,

The preset horizontal distance is less than or equal to the width of the initial prediction block; the preset vertical distance is less than or equal to the height of the initial prediction block;

The preset horizontal distance and the preset vertical distance corresponding to the first type of pixels in the position of n rows and m columns of the initial prediction block are the same as those corresponding to other types of pixels except n rows and m columns of the initial prediction block. The preset horizontal distance is different from the preset vertical distance; n and m are natural numbers;

The preset horizontal distance corresponding to each pixel in the first type of pixels is at least partially the same as the preset vertical distance, or both are different;

The preset horizontal distance and the preset vertical distance corresponding to each pixel in the other types of pixels are at least partially the same, or both are different.
The method according to any one of claims 3 to 7, wherein, based on the at least two target reference pixel sets, performing intra-frame prediction smoothing filtering on each pixel to obtain the second type of prediction values, including:

determine the filter coefficient;

Based on the at least two target reference pixel sets and the filter coefficients, intra-frame prediction smoothing filtering is performed on each pixel to obtain the second type of prediction value.
The method of claim 8, wherein,

The filter coefficients corresponding to the first type of pixels in the position of n rows and m columns of the initial prediction block are different from the filter coefficients corresponding to other types of pixels except for n rows and m columns of the initial prediction block; n and m are natural numbers ;

The filter coefficients corresponding to each pixel in the first type of pixels are at least partially the same, or all different;

The filter coefficients corresponding to each pixel in the other types of pixels are at least partially the same, or all different.
The method according to any one of claims 2 to 9, wherein the initial prediction value, the first type prediction value and the second type prediction value are used to calculate the rate-distortion cost with the original pixel value of the current block, respectively, to determine Get the current prediction mode corresponding to the optimal rate-distortion cost, including:

The initial predicted value, the first type of predicted value and the second type of predicted value are used to calculate the rate-distortion cost with the original pixel value of the current block, respectively, to determine the first minimum rate-distortion cost corresponding to the initial predicted value, a second minimum rate-distortion cost corresponding to the first type of prediction value and a third minimum rate-distortion cost corresponding to the second type of prediction value;

determining a minimum optimal rate-distortion cost from among the first minimum rate-distortion cost, the second minimum rate-distortion cost, and the third minimum rate-distortion cost;

The prediction mode corresponding to the optimal rate-distortion cost is obtained as the current prediction mode.
The method according to claim 10, wherein the filtering flags comprise: an intra prediction filtering permission flag, an intra prediction filtering use flag, an intra prediction smoothing filtering permission flag, and an intra prediction smoothing filtering use flag;

The writing of the index information of the current prediction mode and the filter identification into the code stream includes:

writing the index information of the current prediction mode into the code stream;

determining that the intra-frame prediction filtering permission flag is valid, and the intra-frame prediction smoothing filtering permission flag is valid, and writing the intra-frame prediction filtering permission flag and the intra-frame prediction smoothing filtering permission flag into a code stream;

When the optimal rate-distortion cost is the second minimum rate-distortion cost, determining that the intra-frame prediction filtering use flag is valid, and writing the intra-frame prediction filtering use flag into the code stream;

When the optimal rate-distortion cost is the third minimum rate-distortion cost, it is determined that the intra-frame prediction filtering use flag is invalid and the intra-frame prediction smoothing filter use flag is valid, and the intra-frame prediction filtering use flag is valid. The prediction filtering uses the flag and the intra-frame prediction smoothing filtering uses the flag to write the code stream;

When the optimal rate-distortion cost is the first minimum rate-distortion cost, it is determined that the intra-frame prediction filtering use flag is invalid, and the intra-frame prediction smoothing filter use flag is invalid, and the intra-frame prediction filtering use flag is invalid The predictive filtering uses the flag and the intra prediction smoothing filtering uses the flag to write the code stream.
The method according to claim 10, wherein after the traversing the intra-frame prediction mode and determining the initial prediction value of the initial prediction block corresponding to the current block, the current block is framed by using the current prediction mode. Before intra-prediction, the method further includes:

Intra-frame prediction smoothing filtering is performed on the initial prediction block respectively to obtain a second type of prediction value; wherein, the intra-frame prediction smoothing filtering is to use at least two adjacent reference pixel sets in each adjacent reference pixel set. A plurality of adjacent reference pixels of , the process of filtering the current block;

Using the initial predicted value and the second type of predicted value, a rate-distortion cost calculation is performed with the original pixel value of the current block, respectively, to determine the current prediction mode corresponding to the optimal rate-distortion cost.
The method according to claim 1, wherein after the traversing the intra-frame prediction mode and determining the initial prediction value of the initial prediction block corresponding to the current block, the current block is framed by using the current prediction mode. Before intra-prediction, the method further includes:

According to the type of the intra-frame prediction mode, a prediction mode set is determined, and the prediction mode set is a prediction mode enabling the intra-frame prediction smoothing filtering technique;

For the initial predicted value, use the initial sub-predicted value predicted by the prediction mode set to perform intra-frame prediction smoothing filtering to obtain the second type of predicted value;

Using the initial predicted value and the second type of predicted value, a rate-distortion cost calculation is performed with the original pixel value of the current block, respectively, to determine the current prediction mode corresponding to the optimal rate-distortion cost.
The method according to claim 12 or 13, wherein the filtering flags comprise: an intra-frame prediction smoothing filtering permission flag and an intra-frame prediction smoothing filtering use flag;

The writing of the index information of the current prediction mode and the filter identification into the code stream includes:

writing the index information of the current prediction mode into the code stream;

determining that the intra-frame prediction smoothing filtering permission flag is valid, and writing the intra-frame prediction smoothing filtering permission flag into the code stream;

When the optimal rate-distortion cost is the third minimum rate-distortion cost, determining that the intra-frame prediction smoothing filter usage flag is valid, and writing the intra-frame prediction smoothing filter usage flag into the code stream;

When the optimal rate-distortion cost is the first minimum rate-distortion cost, it is determined that the intra-frame prediction smoothing filter usage flag is invalid, and the intra-frame prediction smoothing filter usage flag is written into the code stream.
The method of claim 11, wherein,

The encoding type of the initial predicted value corresponds to multiple encoding channels;

The filtering identifiers corresponding to each encoding channel use the same identifier, or use one-to-one corresponding identifiers.
The method of claim 1 or 12, wherein the method further comprises:

When the size of the initial prediction block satisfies the prediction area threshold range, it is allowed to use intra prediction smoothing filtering for filtering;

When the size of the initial prediction block is greater than the upper limit of the prediction area threshold range, dividing the initial prediction block to obtain a prediction sub-block that satisfies the prediction area threshold range;

The intra prediction processing flow is implemented for the predicted sub-block.
The method according to claim 1, wherein after the traversing the intra prediction mode and determining the initial prediction value of the initial prediction block corresponding to the current block, the initial prediction value, the first type prediction value and the second type prediction value are used. The predicted value, respectively performing rate-distortion cost calculation with the original pixel value of the current block, and before determining the current prediction mode corresponding to the optimal rate-distortion cost, the method further includes:

Perform intra-frame prediction filtering on the initial prediction block respectively to obtain the first type of prediction value;

Traverse the intra-frame prediction mode, and use intra-frame prediction smoothing filtering to predict and estimate the current block to obtain the second type of prediction value.
The method according to claim 17, wherein the traversing the intra-frame prediction mode, using intra-frame prediction smoothing filtering processing to perform prediction estimation on the current block, to obtain the second type of prediction value, comprising:

Traverse the intra prediction mode, perform intra prediction smoothing filtering processing on the current block based on the preset horizontal distance and preset vertical distance corresponding to each pixel in the current block, and obtain the second type of prediction value.
The method of claim 6, wherein the method further comprises:

When the preset horizontal distance is greater than the width of the initial prediction block, use the leftmost reference pixel in the upper adjacent reference pixel set to fill to the left, until the first area of the upper adjacent reference block range is determined Until the boundary time limit, and/or, use the rightmost reference pixel in the upper adjacent reference pixel set to fill to the right, until the second boundary time limit of the upper adjacent reference block range is determined;

When the preset vertical distance is less than or equal to the height of the initial prediction block, the lowermost reference pixel in the set of adjacent reference pixels on the left side is used to fill down, until the range of the adjacent reference block on the left side is determined. Until the third boundary limit, and/or, use the uppermost reference pixel in the left adjacent reference pixel set to fill upwards until the fourth boundary limit of the left adjacent reference block range is determined.
An intra-frame prediction method, applied to a decoder, comprising:

Obtain the code stream, parse the header information of the code stream, and obtain the filter identification;

When the filtering flag indicates that the intra-frame prediction filtering permission flag is valid and the intra-frame prediction smoothing filtering permission flag is valid, at least obtain the current prediction mode and the intra-frame prediction filtering use flag from the code stream;

When the intra-frame prediction filtering use flag is invalid and a valid intra-frame prediction smoothing filter use flag is obtained from the code stream, based on the current prediction mode, perform intra-frame prediction smoothing filtering processing on the current block to obtain the first a predicted value.
The method according to claim 20, wherein after obtaining at least the current prediction mode and the intra-frame prediction filtering use identifier from the code stream, the method further comprises:

When the intra-frame prediction filtering use flag is valid, obtain the current prediction mode from the code stream;

Based on the current prediction mode, intra prediction filtering is performed on the current block to obtain a second prediction value.
The method according to claim 20, wherein, based on the current prediction mode, performing intra-frame prediction smoothing filtering processing on the current block to obtain the first predicted value, comprising:

Based on the preset horizontal distance and the preset vertical distance corresponding to each pixel in the current block, perform intra prediction smoothing filtering processing on the current block to obtain the first predicted value.
The method according to claim 22, wherein, based on the preset horizontal distance and preset vertical distance corresponding to each pixel in the current block, performing intra-frame prediction smoothing filtering processing on the current block to obtain the The first predicted value, including:

Based on the preset horizontal distance and the preset vertical distance corresponding to each pixel in the current block, at least two adjacent reference pixels of at least two adjacent reference pixel sets corresponding to each pixel are determined block range;

From the at least two adjacent reference pixel sets, at least two target reference pixel sets belonging to the at least two adjacent reference blocks are correspondingly determined; wherein, each target reference pixel set includes at least one target reference pixel;

Based on the at least two target reference pixel sets, intra-frame prediction smoothing filtering is performed on each pixel to obtain the first predicted value.
The method according to claim 23, wherein the at least two target reference pixel sets belonging to the range of the at least two adjacent reference blocks are correspondingly determined from the at least two adjacent reference pixel sets, include:

From the at least two adjacent reference pixel sets, pixels at two boundary pixel positions of each adjacent reference block range are determined as a target reference pixel set, thereby determining the at least two target reference pixel sets.
The method according to claim 23, wherein the at least two target reference pixel sets belonging to the range of the at least two adjacent reference blocks are correspondingly determined from the at least two adjacent reference pixel sets, include:

From the at least two adjacent reference pixel sets, M pixels at any pixel position within the range of each adjacent reference block are determined as a target reference pixel set, thereby determining the at least two target reference pixel sets, Wherein, M is greater than 0 and does not exceed the maximum number of pixels in the corresponding adjacent reference pixel set.
The method of any one of claims 23 to 25, wherein,

The at least two adjacent reference pixel sets include: a left adjacent reference pixel set and an upper adjacent reference pixel set;

The at least two adjacent reference block ranges include: a left adjacent reference pixel range and an upper adjacent reference pixel range;

The at least two target reference pixel sets include: a left target adjacent reference pixel set and an upper target adjacent reference pixel set.
The method of any one of claims 23 to 26, wherein,

The preset horizontal distance is less than or equal to the width of the initial prediction block; the preset vertical distance is less than or equal to the height of the initial prediction block;

The preset horizontal distance and the preset vertical distance corresponding to the first type of pixels in the position of n rows and m columns of the initial prediction block are the same as those corresponding to other types of pixels except n rows and m columns of the initial prediction block. The preset horizontal distance is different from the preset vertical distance; n and m are natural numbers;

The preset horizontal distance corresponding to each pixel in the first type of pixels is at least partially the same as the preset vertical distance, or both are different;

The preset horizontal distance and the preset vertical distance corresponding to each of the other types of pixels are at least partially the same, or both are different.
The method according to any one of claims 23 to 27, wherein, based on the at least two target reference pixel sets, performing intra-frame prediction smoothing filtering processing on each pixel to obtain the second type of prediction values, including:

determine the filter coefficient;

Based on the at least two target reference pixel sets and the filter coefficients, intra-frame prediction smoothing filtering is performed on each pixel to obtain the second type of prediction value.
The method of claim 22, wherein,

The filter coefficients corresponding to the first type of pixels in the position of n rows and m columns of the initial prediction block are different from the filter coefficients corresponding to other types of pixels except for n rows and m columns of the initial prediction block; n and m are natural numbers ;

The filter coefficients corresponding to each pixel in the first type of pixels are at least partially the same, or all different;

The filter coefficients corresponding to each pixel in the other types of pixels are at least partially the same, or all different.
The method of claim 26, wherein,

When the preset horizontal distance is greater than the width of the initial prediction block, use the leftmost reference pixel in the upper adjacent reference pixel set to fill to the left, until the first area of the upper adjacent reference block range is determined Until the boundary time limit, and/or, use the rightmost reference pixel in the upper adjacent reference pixel set to fill to the right, until the second boundary time limit of the upper adjacent reference block range is determined;

When the preset vertical distance is less than or equal to the height of the initial prediction block, the lowermost reference pixel in the set of adjacent reference pixels on the left side is used to fill down, until the range of the adjacent reference block on the left side is determined. Until the third boundary limit, and/or, use the uppermost reference pixel in the left adjacent reference pixel set to fill upwards until the fourth boundary limit of the left adjacent reference block range is determined.
An encoder comprising:

The first determining part is configured to traverse the intra prediction mode, and determine the initial prediction value of the initial prediction block corresponding to the current block;

The first filtering part is configured to perform intra-frame prediction filtering and intra-frame prediction smoothing filtering respectively on the initial prediction block to obtain a first-type prediction value and a second-type prediction value; wherein, the intra-frame prediction smoothing filtering process A process of filtering the current block for using a plurality of adjacent reference pixels in each adjacent reference pixel set in at least two adjacent reference pixel sets;

The first determining part is further configured to use the initial prediction value, the first type prediction value and the second type prediction value to calculate the rate-distortion cost with the original pixel value of the current block, respectively, and determine the corresponding optimal rate-distortion cost. The current forecast mode of ;

a first prediction part configured to use the current prediction mode to perform intra prediction on the current block;

The writing part is configured to write the index information of the current prediction mode and the filter identifier into the code stream, wherein the filter identifier represents an identifier corresponding to intra-frame prediction filtering and/or intra-frame prediction smoothing filtering.
A decoder including:

The parsing part is configured to obtain the code stream, parse the header information of the code stream, and obtain the filtering identifier; and when the filtering identifier is an indication that the intra-frame prediction filtering permission flag is valid, and the intra-frame prediction smoothing filtering permission flag is valid, Obtain at least the current prediction mode and intra-frame prediction filtering use identifier from the code stream;

The second prediction part is configured to, based on the current prediction mode, frame the current block when the intra-frame prediction filtering use flag is invalid and a valid intra-frame prediction smoothing filter use flag is obtained from the code stream The intra-prediction smoothing filtering process is performed to obtain the first predicted value.
An encoder comprising:

a first memory and a first processor;

The first memory stores a computer program executable on the first processor, and the first processor implements the intra prediction method of any one of claims 1 to 19 when the first processor executes the program.
A decoder including:

a second memory and a second processor;

The second memory stores a computer program executable on the second processor, and when the second processor executes the program, the intra prediction method of any one of claims 20 to 30 is implemented.
A storage medium on which a computer program is stored, and when the computer program is executed by the first processor, the intra-frame prediction method of any one of claims 1 to 19 is implemented; or, the computer program is executed by the second processor When the intra-frame prediction method described in any one of claims 20 to 30 is implemented.