CN100586184C - Infra-frame prediction method - Google Patents

Abstract

The invention relates to an intra-frame prediction method. The intra-frame prediction method includes selecting a 4×4 pixel brightness block to be predicted as a current block; judging whether the current block is a central block, and if so, calculating the prediction block and the current block of the current block in all prediction modes The energy function of the residual, and determine the candidate prediction mode according to the energy function in different prediction modes, otherwise, use the available prediction mode as the candidate prediction mode; calculate the rate-distortion cost parameter of the candidate prediction mode, and select the smallest rate-distortion cost parameter The candidate prediction mode is the optimal prediction mode of the current block. The present invention determines the candidate prediction mode according to the residual energy function, and determines the optimal prediction mode by calculating the rate-distortion cost parameter of the candidate prediction mode, so as to reduce the number of prediction modes that need to calculate the rate-distortion cost parameter, thereby reducing the calculation amount.

Description

Intra Prediction Method

technical field

The present invention relates to an intra-frame prediction method, in particular to an intra-frame prediction method capable of effectively improving operation speed.

Background technique

The new-generation video coding standard H.264/AVC has good network performance and is suitable for interactive and non-interactive application environments. It has received great attention since its release, and its notable advantage is high compression efficiency. Figure 1 shows the frame diagram of the encoder of the H.264/AVC standard, and its working process can be divided into a forward channel and a reconstruction channel according to the data flow direction. The encoding of the current frame F _n is to encode the macroblock of 16×16 pixels in the original image. The macroblock encoding is divided into intra-frame coding and inter-frame coding. In intra-frame coding and inter-frame coding, the predicted macroblock P is composed of The reconstructed frame is obtained. In the intra-frame coding mode, P is decoded and reconstructed from the coded macroblock in the current frame, and predicted from the reconstructed macroblock _uF'n ; in the inter-frame coding mode, P is obtained by one or more reference frames F ' _n-1 is obtained by motion compensation prediction. The difference between the predicted macroblock P and the current macroblock F _n is the residual macroblock D _n , and D _n is transformed and quantized to obtain a series of transformation parameters X. The parameters X need to be processed in two aspects: one is reordering and entropy Coding processing, the whole process has no feedback component, so it is called the forward channel; the second is inverse quantization and inverse transformation processing, generating macroblock D' _n , D' _n is added to predicted macroblock P to obtain reconstructed macroblock _uF'n , the reconstructed macroblock _uF'n undergoes a series of processing to obtain a reconstructed reference frame _F'n , which is used for motion estimation of the next frame, so it is called a reconstruction channel. The improvement of H.264/AVC performance is inseparable from some new technologies adopted, such as the use of intra-frame prediction method based on space domain, integer discrete cosine transform (Discrete Cosine Transform, hereinafter referred to as DCT) for transformation, motion estimation/motion compensation based on 4×4 pixel block, reconstruction channel adopts loop filter, new entropy coding method, etc.

Among them, the intra prediction based on the space domain is an important factor for the performance improvement of H.264/AVC. The intra prediction uses the spatial correlation of the image to predict the information of the current block according to the information of the decoded and reconstructed adjacent pixel blocks, and obtains the current The prediction block of the current block, and then transform, quantize, and encode the residual of the current block and the prediction block of the current block. In order to better represent the current block, H.264/AVC uses Rate Distortion Optimization (Rate Distortion Optimization, hereinafter referred to as RDO ) techniques to optimize encoding quality and minimize encoding bits.

In the H.264/AVC standard, a 16×16 pixel macroblock is used as a unit. A macroblock includes a luma block and two corresponding chrominance blocks, where the luma block is one 16×16 or 16 4×4 , the chroma block is 8×8.

For a 4×4 luma block, there are 9 prediction modes for intra prediction, and these prediction modes are:

Mode 0: Vertical prediction mode

Mode 1: Horizontal prediction mode

Mode 2: DC prediction mode

Mode 3: 45-degree direction prediction (diagonal down/left prediction) mode

Mode 4: 135-degree direction prediction (diagonal down/right prediction) mode

Mode 5: 112.5-degree direction prediction (vertical-right prediction) mode

Mode 6: 157.5 degree direction prediction (horizontal-down prediction) mode

Mode 7: 67.5-degree direction prediction (vertical-left prediction) mode

Mode 8: 22.5 degree direction prediction (horizontal-up prediction) mode

Among them, in addition to the DC prediction mode, the remaining 8 prediction modes are called directional prediction modes. Figure 2 shows the relationship between the reference pixel and the current block pixel to be predicted, where capital letters represent reference pixels, and small letters represent is the current block pixel. The numbers in Fig. 3 indicate the directions of the prediction modes in each direction.

For a 16×16 luma block, there are 4 prediction modes for intra prediction, these prediction modes are:

vertical prediction mode

Horizontal prediction mode

DC prediction mode

Plane prediction mode

For 8×8 chrominance blocks, there are 4 prediction modes for intra prediction, these prediction modes are:

DC prediction mode

Horizontal prediction mode

vertical prediction mode

Plane prediction mode

From the above analysis, it can be seen that if M8 represents the number of prediction modes for chrominance blocks, M4 represents the number of prediction modes for 4×4 luma blocks, and M16 represents the number of prediction modes for 16×16 luma blocks, in one macroblock The number of combinations of luma block and chrominance block prediction modes is M8×(M4×16+M16)=592, that is, in order to determine the optimal mode of intra prediction of a macroblock, the encoder needs to perform 592 RDO calculations. Therefore, the improvement of H.264/AVC compression rate is at the cost of increased computational complexity, which is difficult to apply in real time.

Contents of the invention

The purpose of the present invention is to provide an intra-frame prediction method for the defects of the prior art, to solve the problem of high computational complexity of the existing intra-frame prediction method, improve the speed of intra-frame prediction encoding, and facilitate the real-time application.

The present invention provides following technical scheme through embodiment:

A method of intra-frame prediction, comprising:

Step 1: Select a 4×4 pixel brightness block to be predicted as the current block;

Step 2: Determine whether the current block is the central block, if so, perform step 3, otherwise, perform step 4;

Step 3: Calculate the energy function of the prediction block of the current block and the residual of the current block in all prediction modes, and determine candidate prediction modes according to the energy functions in different prediction modes, and perform step 5;

Wherein step 3 specifically includes:

Step 31: Calculating prediction blocks of the current block in different prediction modes;

Step 32: According to the prediction block of the current block in different prediction modes and the current block, obtain the prediction block of the current block in different prediction modes and the residual of the current block;

Step 33: performing discrete cosine transform on the residuals in different prediction modes to obtain energy functions of the residuals in different prediction modes;

The calculation formula of the energy function is:

$\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; 4</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\underset{\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; 4</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; w</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; DCT</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

in, $w(\mathrm{the\; y},x)=\left\{\begin{array}{cc}0.2& (\mathrm{the\; y}=1,x=1),(\mathrm{the\; y}=1,x=2),(\mathrm{the\; y}=2,x=1)\\ 0.4& (\mathrm{the\; y}=1,x=3),(\mathrm{the\; y}=2,x=2),(\mathrm{the\; y}=3,x=1)\\ 0.6& (\mathrm{the\; y}=1,x=4),(\mathrm{the\; y}=2,x=3),(\mathrm{the\; y}=3,x=2),(\mathrm{the\; y}=4,x=1)\\ 0.8& (\mathrm{the\; y}=2,x=4),(\mathrm{the\; y}=3,x=3),(\mathrm{the\; y}=4,x=2)\\ 1& (\mathrm{the\; y}=3,x=4),(\mathrm{the\; y}=4,x=3),(\mathrm{the\; y}=4,x=4)\end{array}\right.$

$\mathrm{<span\; class="notranslate">\; DCT</span>}<span\; class="notranslate">\; =</span>\left(\begin{array}{c}\left[\begin{array}{cccc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 2</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]\mathrm{<span\; class="notranslate">\; x</span>}\left[\begin{array}{cccc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 2</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]\end{array}\right)<span\; class="notranslate">\; \&CircleTimes;</span>\left[\begin{array}{cccc}{\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}}\\ {\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}}\end{array}\right]$

y is the coordinate of the width direction of the current block, x is the coordinate of the height direction of the current block, X is the residual difference between the prediction block of the current block and the current block, and DCT(y, x) is the The coordinates of the current block are the DCT coefficients of (y, x), $a=\frac{1}{2},$ $b=\sqrt{\frac{2}{5}};$

Step 34: Arrange the energy functions of residuals in different prediction modes in ascending order of energy values, and obtain the first energy, second energy, third energy, and fourth energy with smaller energy values in sequence, and obtain the corresponding a first prediction mode corresponding to the first energy, a second prediction mode corresponding to the second energy, a third prediction mode corresponding to the third energy and a fourth prediction mode corresponding to the fourth energy ;

Step 35: Determine a candidate prediction mode according to the first energy, the second energy, the third energy and the fourth energy and the first prediction mode, the second prediction mode, the third prediction mode and the fourth prediction mode;

Step 4: Use the available prediction modes as candidate prediction modes, and perform step 5;

Step 5: Calculate the rate-distortion cost parameters of the candidate prediction modes, and select the candidate prediction mode with the smallest rate-distortion cost parameter as the optimal prediction mode of the current block.

The embodiment of the present invention selects the candidate prediction mode of the 4×4 pixel brightness block according to the energy function, and obtains the optimal prediction mode by calculating the rate-distortion cost parameters of the candidate prediction modes, instead of obtaining the optimal prediction according to the rate-distortion cost parameters of all prediction modes mode, due to the reduction in the number of prediction modes, the calculation amount of the rate-distortion cost parameter can be reduced, and the calculation speed can be improved, which is suitable for real-time encoding.

Description of drawings

Fig. 1 is the encoder frame diagram of prior art H.264/AVC standard;

2 is a schematic diagram of the relationship between a reference pixel and a current block pixel to be predicted in the prior art H.264/AVC standard;

FIG. 3 is a pointing diagram of a direction prediction mode in the prior art H.264/AVC standard;

FIG. 4 is a flow chart of Embodiment 1 of the intra prediction method of the present invention;

FIG. 5 is a flow chart of Embodiment 2 of the intra prediction method of the present invention.

Detailed ways

The technical solutions of the present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

FIG. 4 is a flow chart of Embodiment 1 of the intra prediction method of the present invention. The intra prediction method includes:

Step 41: Select a 4×4 pixel brightness block to be predicted as the current block;

Step 42: judging whether the current block is a central block, if so, execute step 43, otherwise, execute step 44;

Step 43: Calculate the energy function of the prediction block of the current block and the residual of the current block in all prediction modes, and determine candidate prediction modes according to the energy functions in different prediction modes, and perform step 45;

Step 44: Use available prediction modes as candidate prediction modes, and perform step 45;

Step 45: Calculate the rate-distortion cost parameters of the candidate prediction modes, and select the candidate prediction mode with the smallest rate-distortion cost parameter as the optimal prediction mode of the current block.

Fig. 5 is a flowchart of Embodiment 2 of the intra prediction method of the present invention, the intra prediction method includes:

Step 501: Divide a 16×16 macroblock into 16 4×4 luma blocks, and execute step 502. In the H.264/AVC standard, a 16×16 pixel macroblock is used as a unit. A macroblock includes a luma block and two chrominance blocks. The luma block is one 16×16 or 16 4×4. The degree block is 8×8, and the pixel value of the current luma block or chrominance block is predicted and coded according to multiple prediction modes using the decoded and reconstructed pixel values of the upper block adjacent to the current block and the adjacent left block. Because the transformation in the H.264/AVC standard is based on a 4×4 block, and 4×4 blocks occupy a large proportion in predictive coding, therefore, this embodiment is aimed at the predictive coding of 4×4 luma blocks. Improvement, the predictive coding method for 16×16 luma block and 8×8 chrominance block will not be changed.

Step 502: Select a 4×4 luma block to be predictively coded as the current block, and execute step 503. The prediction modes of the intra prediction method for the 4×4 luma block include 8 directional predictions and 1 DC prediction. In the traditional H.264/AVC standard, the selection of the prediction mode of the luma block using RDO technology includes the following steps:

(1) For the 4×4 luminance block, calculate the rate-distortion cost parameters in 9 prediction modes respectively, and obtain the minimum value of the rate-distortion cost parameter and the prediction mode corresponding to the minimum rate-distortion cost parameter. The formula for calculating the rate-distortion cost parameter (RDCost) is: RDCost=SSD+λ×bitrate, where RDCost is the rate-distortion cost parameter; SSD is the sum of squares of all elements of the difference block between the current block and the reconstructed block, that is, the reconstructed The square of the residual error; λ is a function of the quantization parameter (QP), λ=0.85×2 ^(QP-12)/3 ; bitrate is the code rate after entropy coding, that is, the number of bits forming the code stream after predictive coding.

(2) Obtain the minimum rate-distortion cost parameters of 16 4×4 luminance blocks according to the above method, add these 16 minimum rate-distortion cost parameters, and use the added sum as the rate-distortion cost of 4×4 luminance blocks parameter.

(3) For the 16×16 brightness block, calculate the sum of all elements after the hasmard transformation of the residual of the current block and the prediction block under the four prediction modes, that is, the value of the residual hasmard transformation sum (STAD), and calculate the STAD value The rate-distortion cost parameter of the smallest prediction mode, and the rate-distortion cost parameter is used as the rate-distortion cost parameter of the 16×16 luma block.

(4) Compare the rate-distortion cost parameters obtained in (2) and (3), and select the prediction mode corresponding to the smallest rate-distortion cost parameter as the prediction mode of the intra-frame luma block.

From the formula of the rate-distortion cost parameter in (1) above, it can be seen that each prediction mode of the current 4×4 luma block must be pre-encoded, and the computational complexity is relatively large. In order to reduce the computational complexity, the RDO can be simplified To calculate or reduce the prediction mode used to calculate the RDO, in order to reduce the complexity of the intra-frame coding while maintaining the efficiency of the intra-frame coding better, this embodiment adopts the method of reducing the prediction mode used to calculate the RDO.

Step 503: Determine whether the current block is the center block, if yes, execute step 504, otherwise, execute step 507. Since video coding uses the order of decoding and reconstruction from left to right and from top to bottom, if the left block adjacent to the current block and the adjacent upper block have been decoded, it is called the left block and the upper block are available, When the left block has been decoded and reconstructed, it is called the left block is available, when the upper block has been decoded and reconstructed, it is called the upper block is available, and when both are not decoded, it is called neither available. When both the left block and the upper block are available, the current block is the center block. At this time, there are 9 prediction modes corresponding to the current block; when the current block is not the center block, the prediction mode corresponding to the current block is called an available prediction mode , there are less than 9 available prediction modes. If only the left block is available, the available prediction modes are horizontal prediction mode (mode 1), DC prediction mode (mode 2), 22.5-degree direction prediction mode (mode 8), and only the upper block is available , the available prediction modes are vertical prediction mode (mode 0), DC prediction mode (mode 2), 45-degree direction prediction mode (mode 3), and 67.5-degree direction prediction (mode 7).

Step 504: Calculate the energy function of the residuals between the current block and the predicted block in all prediction modes, and execute step 505. It can be seen from the above analysis that if the current block is the center block, there are 9 available prediction modes. In order to reduce the calculation amount of RDO calculation, the method of reducing the prediction mode is adopted. Since most areas of natural or synthetic images are composed of low frequencies, most high-frequency information can be coded with 0-length, and high frequencies can roughly reflect the complexity of an image's texture. In practical applications, the information of the adjacent blocks decoded is used to predict the information of the current block according to the texture direction of the current block, and then the energy function is calculated for the corresponding residual. The energy function defined in the present invention is obtained from the multi-dimensional frequency Mapping from domain space to one-dimensional energy space, the high-frequency component of the DCT transformation of the texture residual block reflects the detailed information of the texture residual. In order to highlight the detailed information of the texture residual, the present invention assigns a higher weight to the high frequency The value (w(y,x)) can amplify the influence of high frequency coefficients. According to the definition of the DCT energy function described in the present invention, the richer the details of the texture residual block (the more high-frequency information), the greater its DCT energy. The present invention just utilizes this feature of the energy function, and selects several modes with less complex texture residual blocks (less energy, less high-frequency information) to calculate RDO, thereby reducing The number of calculated modes finally achieves the purpose of speeding up the intra-frame encoding speed. The calculation formula of the energy function is:

${\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; DCT</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\underset{\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; h</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; w</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; DCT</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Here E _DCT is the energy calculation of the DCT transform for the prediction residual, ω and h are the width and height of the current block, DCT(y, x) is the DCT coefficient at (y, x), for H.264/AVC For a 4×4 brightness block, the weight function w(y, x) (x=1, 2, ... h, y = 1, 2, ..., ω) in the above formula is defined as follows:

$\mathrm{<span\; class="notranslate">\; w</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; 0.2</span>}& <span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\\ \mathrm{<span\; class="notranslate">\; 0.4</span>}& <span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\\ \mathrm{<span\; class="notranslate">\; 0.6</span>}& <span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\\ \mathrm{<span\; class="notranslate">\; 0.8</span>}& <span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>\\ \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>\end{array}\right.$

Among them, ω=4, h=4

The DCT transformation in the above formula adopts the integer DCT transformation in the standard, and the formula is as follows:

$\mathrm{<span\; class="notranslate">\; DCT</span>}<span\; class="notranslate">\; =</span>\left(\begin{array}{c}\left[\begin{array}{cccc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 2</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="x"\; filenames="C2008100567760016H2.png"\; state="\{\{state\}\}">x</figure-callout></span>}\left[\begin{array}{cccc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 2</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]\end{array}\right)<span\; class="notranslate">\; \&CircleTimes;</span>\left[\begin{array}{cccc}{\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}}\\ {\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}}\end{array}\right]$

in, $a=\frac{1}{2},$ $b=\sqrt{\frac{2}{5}},$ X is the residual between the current block and the predicted block.

Step 505: Arrange the energy function values of the reconstruction residuals in different prediction modes in ascending order, and select the four smallest energy values, the first energy (E1), the second energy (E2), and the third energy ( E3), the fourth energy (E4) and the corresponding first prediction mode (m1), second prediction mode (m2), third prediction mode (m3), and fourth prediction mode (m4), execute step 506.

Step 506: According to the following screening principles, determine candidate prediction modes from the above four prediction modes m1, m2, m3, and m4, and execute step 508. The screening principles include:

Step 5061: Determine whether E2 is greater than α×E1, if yes, execute step 5062; otherwise, execute step 5063.

Step 5062: The candidate prediction mode is m1.

Step 5063: Determine whether E3 is greater than α×E1, if yes, go to step 5064, otherwise go to step 5065.

Step 5064: The candidate prediction modes are m1 and m2.

Step 5065: Determine whether E4 is greater than α×E1, if yes, go to step 5066, otherwise go to step 5067.

Step 5066: The candidate prediction modes are m1, m2 and m3.

Step 5067: Determine whether E4 is smaller than β×E1, if yes, execute step 5068, otherwise, execute step 5069.

Step 5068: The candidate prediction modes are m1 and m2.

Step 5069: The candidate prediction modes are m1, m2, m3 and m4.

The above-mentioned α and β are constants. After a lot of experiments, the encoding effect is the best when α=1.4 and β=1.05.

Step 507: Use available prediction modes as candidate prediction modes, and execute step 508. When only the left block adjacent to the current block is available, the available prediction modes are mode 1, mode 2, and mode 8; when only the upper block adjacent to the current block is available, the available prediction modes are mode 0, mode 2, and mode 3. Mode 7: when neither the left block nor the upper block adjacent to the current block is available, the candidate prediction mode is the DC prediction mode.

Step 508: Perform RDO calculation on each candidate prediction mode of the current block, select the candidate prediction mode with the smallest rate-distortion cost parameter RDCost as the optimal prediction mode of the current block, and execute step 509. That is, when both the adjacent left block and the adjacent upper block of the current block are available, the RDO calculation is performed according to the candidate prediction mode determined by the energy function, and the candidate prediction mode with the smallest rate-distortion cost parameter is the optimal prediction mode; When one of the adjacent left block or the adjacent upper block is available, RDO calculation is performed on the available prediction modes, and the available prediction mode with the smallest rate-distortion cost parameter is the optimal prediction mode; the adjacent left block and the adjacent When none of the upper blocks is available, the DC prediction mode is the optimal prediction mode.

Step 509: Determine whether all the 16 luminance blocks have been used as the current block for the above calculation, if yes, perform step 510, otherwise, repeat step 502 until all 4×4 luminance blocks have completed predictive coding.

Step 510: Add the minimum rate-distortion cost parameters of the 16 4×4 luma blocks, and use the added sum as the rate-distortion cost parameter RDCost4 of the 4×4 luma blocks, and execute step 512 .

Step 511: Calculate the rate-distortion cost parameters of the four prediction modes for the 16×16 luma block, and obtain the minimum rate-distortion cost parameter and the prediction mode corresponding to the minimum rate-distortion cost parameter, wherein the minimum rate-distortion cost parameter is 16×16 luma The rate-distortion cost parameter RDCost16 of the block corresponds to the optimal prediction mode of the 16×16 luma block where the prediction mode of the minimum rate-distortion cost parameter is, and step 512 is executed.

Step 512: Compare the rate-distortion cost parameter RDcost4 of the 4×4 luma block with the rate-distortion cost parameter RDcost16 of the 16×16 luma block, if RDcost4<RDcost16, go to step 513; otherwise, go to step 514.

Step 513: Select the optimal prediction mode corresponding to each 4*4 luma block as the optimal prediction mode of the luma block of the 16*16 macroblock.

Step 514: Select the optimal prediction mode of the 16×16 luma block corresponding to RDcost16 as the optimal prediction mode of the luma block of the 16×16 macroblock.

In this embodiment, the intra-frame prediction method of the 4×4 luma block in the H.264/AVC standard is improved, and the energy function obtained by the DCT transformation of the residual is used to determine the candidate prediction mode, and the RDO calculation is performed on the candidate prediction mode, and The RDO calculation is not performed on all prediction modes. Since the number of prediction modes for RDO calculation is reduced, the calculation amount of RDO can be effectively reduced, which is beneficial for real-time coding.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent replacements are made to some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims (5)

1. A method for intra-frame prediction, comprising: Step 1: Select a 4×4 pixel brightness block to be predicted as the current block; Step 2: Determine whether the current block is the central block, if so, perform step 3, otherwise, perform step 4; Step 3: Calculate the energy function of the prediction block of the current block and the residual of the current block in all prediction modes, and determine candidate prediction modes according to the energy functions in different prediction modes, and perform step 5; Wherein step 3 specifically includes: Step 31: Calculating prediction blocks of the current block in different prediction modes; Step 32: According to the prediction block of the current block in different prediction modes and the current block, obtain the prediction block of the current block in different prediction modes and the residual of the current block; Step 33: performing discrete cosine transform on the residuals in different prediction modes to obtain energy functions of the residuals in different prediction modes; The calculation formula of the energy function is: $\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; 4</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\underset{\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; 4</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; w</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; DCT</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$ in, $w(\mathrm{the\; y},x)=\left\{\begin{array}{cc}0.2& (\mathrm{the\; y}=1,x=1),(\mathrm{the\; y}=1,x=2),(\mathrm{the\; y}=2,x=1)\\ 0.4& (\mathrm{the\; y}=1,x=3),(\mathrm{the\; y}=2,x=2),(\mathrm{the\; y}=3,x=1)\\ 0.6& (\mathrm{the\; y}=1,x=4),(\mathrm{the\; y}=2,x=3),(\mathrm{the\; y}=3,x=2),(\mathrm{the\; y}=4,x=1)\\ 0.8& (\mathrm{the\; y}=2,x=4),(\mathrm{the\; y}=3,x=3),(\mathrm{the\; y}=4,x=2)\\ 1& (\mathrm{the\; y}=3,x=4),(\mathrm{the\; y}=4,x=3),(\mathrm{the\; y}=4,x=4)\end{array}\right.$ $\mathrm{<span\; class="notranslate">\; DCT</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; (</span>\left[\begin{array}{cccc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 2</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]\mathrm{<span\; class="notranslate">\; x</span>}\left[\begin{array}{cccc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 2</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&CircleTimes;</span>\left[\begin{array}{cccc}{\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}}\\ {\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; a</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}}& \frac{\mathrm{<span\; class="notranslate">\; ab</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; 4</span>}}\end{array}\right]$ y is the coordinate of the width direction of the current block, x is the coordinate of the height direction of the current block, X is the residual difference between the prediction block of the current block and the current block, and DCT(y, x) is the The coordinates of the current block are the DCT coefficients of (y, x), $a=\frac{1}{2},b=\sqrt{\frac{2}{5}};$ Step 34: Arrange the energy functions of residuals in different prediction modes in ascending order of energy values, and obtain the first energy, second energy, third energy, and fourth energy with smaller energy values in sequence, and obtain the corresponding a first prediction mode corresponding to the first energy, a second prediction mode corresponding to the second energy, a third prediction mode corresponding to the third energy and a fourth prediction mode corresponding to the fourth energy ; Step 35: Determine a candidate prediction mode according to the first energy, the second energy, the third energy and the fourth energy and the first prediction mode, the second prediction mode, the third prediction mode and the fourth prediction mode; Step 4: Use the available prediction modes as candidate prediction modes, and perform step 5; Step 5: Calculate the rate-distortion cost parameters of the candidate prediction modes, and select the candidate prediction mode with the smallest rate-distortion cost parameter as the optimal prediction mode of the current block. 2. The intra-frame prediction method according to claim 1, wherein the step 2 specifically includes: pixel brightness blocks are decoded in order from left to right and from top to bottom, and when adjacent to the current block When both the left pixel luminance block and the upper pixel luminance block of have been decoded, the current block is the central block. 3. The intra-frame prediction method according to claim 1, wherein the step 35 specifically comprises: Step 351: judging whether the second energy is greater than 1.4 times the first energy, if yes, then the first prediction mode is a candidate prediction mode, otherwise, execute step 352; Step 352: Determine whether the third energy is greater than 1.4 times the first energy, if yes, then the first prediction mode and the second prediction mode are candidate prediction modes, otherwise, execute step 353; Step 353: judging whether the fourth energy is greater than 1.4 times the first energy, if yes, then the first prediction mode, the second prediction mode and the third prediction mode are candidate prediction modes, otherwise, execute step 354; Step 354: Judging whether the fourth energy is less than 1.05 times the first energy, if yes, the first prediction mode and the second prediction mode are candidate prediction modes, otherwise, the first prediction mode, the second prediction mode The prediction mode, the third prediction mode and the fourth prediction mode are candidate prediction modes. 4. The intra-frame prediction method according to claim 1, wherein the step 4 specifically comprises: When the upper pixel brightness block adjacent to the current block has been decoded, use vertical prediction mode, DC prediction mode, 45-degree direction prediction mode, and 67.5-degree direction prediction mode as candidate prediction modes; When the left pixel brightness block adjacent to the current block has been decoded, the horizontal prediction mode, the DC prediction mode, and the 22.5-degree direction prediction mode are used as candidate prediction modes; When neither the upper pixel brightness block nor the left pixel brightness block adjacent to the current block has been decoded, the DC prediction mode is used as the candidate prediction mode. 5. The intra prediction method according to claim 1, further comprising: Step 6: The sum of the rate-distortion cost parameters of the optimal prediction mode of the 16 4×4 pixel brightness blocks is used as the rate-distortion cost parameter of the 4×4 pixel brightness block; Step 7: Calculate the value of the residual hasmard transform sum of the 16×16 pixel brightness block in different prediction modes; Step 8: using the rate-distortion cost parameter of the prediction mode with the minimum value of the residual hasmard transform sum as the rate-distortion cost parameter of the 16×16 pixel luma block; Step 9: Determine whether the rate-distortion cost parameter of the 4×4 pixel brightness block is smaller than the rate-distortion cost parameter of the 16×16 pixel brightness block, and if so, use the optimal prediction mode corresponding to the 4×4 pixel brightness block as The optimal prediction mode of the pixel brightness block, otherwise, the optimal prediction mode corresponding to the 16×16 pixel brightness block is used as the optimal prediction mode of the pixel brightness block.

Images