CN103517069B - A kind of HEVC intra-frame prediction quick mode selection method based on texture analysis - Google Patents

Abstract

The present invention discloses a fast mode selection method for HEVC intra-frame prediction based on texture analysis. Before performing intra-frame prediction on the coding tree unit, the method uses the absolute value of the gradient in the four directions of horizontal, vertical, left-bottom, and right-bottom. and determine the main texture direction and texture complexity of each 4×4 unit in the coding tree unit, and determine the division of the current coding tree unit according to the principle that a larger coding unit is used in the texture smooth area and a smaller coding unit is used in the texture complex area. When predicting, according to the main texture direction of the prediction unit, several least likely prediction modes are excluded, and then rough mode selection and rate-distortion optimization mode selection are performed according to the HEVC coding standard. The fast mode selection method for HEVC intra-frame prediction based on texture analysis proposed by the present invention can significantly improve the encoding speed under the premise of ensuring the encoding quality.

Description

A Fast Mode Selection Method for HEVC Intra Prediction Based on Texture Analysis

technical field

The invention belongs to the field of multimedia coding, and in particular relates to a fast mode selection method for intra-frame prediction aimed at the HEVC video coding standard, which is a method for dividing coding units and fast mode selection for prediction directions according to image texture analysis.

Background technique

With the release of the next-generation video coding standard HEVC (High Efficiency Video Coding), video coding efficiency has been further improved. Compared with the previous video coding standard H.264/AVC, under the same coding quality, it can save nearly half of the code rate, but this is at the expense of coding complexity. How to minimize the complexity of the algorithm and significantly improve the encoding speed without changing the code stream structure and maintaining the original code rate is the key to the large-scale promotion and application of HEVC.

Intra-frame coding can provide reference information for inter-frame coding, and can prevent error drift caused by bit errors, so intra-frame coding is essential in video coding. There are mainly two parts that affect the time-consuming of HEVC intra-frame encoding: the division of coding units and the selection of prediction directions.

There are mainly two methods for fast division of coding units. One is to use Bayesian and other statistical learning methods to determine the division of coding units according to the rate-distortion cost of the coded units. However, the off-line learning results cannot be used well for all For video sequences, the online learning method will introduce additional computational complexity; the second is to determine the division of the coding unit according to the image content complexity of the current coding unit. The key to this method is to calculate the image complexity and select an appropriate threshold.

There are three main methods for fast mode selection of the prediction direction. One is to use the information of the surrounding coding units to reduce the prediction direction candidate set of the current coding unit according to the strong correlation between adjacent regions or adjacent frames. However, in practical applications , there are often scene switching or texture-rich videos. In this case, the spatio-temporal correlation is weak, resulting in a significant drop in coding performance; the second is to use the high correlation between the best prediction direction and the texture direction of the coding unit to reduce candidate The number of prediction directions; the third is to estimate the code rate and distortion of each prediction mode in a low-complexity way to reduce the computational complexity of the rate-distortion cost. This method retains all prediction directions and does not add additional Computational complexity, but due to the strong randomness of natural image video sequences, it is difficult to construct a rate-distortion estimation model suitable for all video sequences. Compared with the second method, the coding performance of this method has dropped significantly.

In the new generation of video coding standards, HEVC introduces a new coding tree unit (Coding Tree Unit, CTU) concept, which is used to replace the concept of macroblocks in existing video coding standards. In addition, there are coding units, prediction units, and transformation units. and other concepts. The coding tree unit is a recursive quadtree structure, and the depth can range from 0 to 3, that is, the size of the coding unit can be from 64×64 to 8×8. For the smallest coding unit, HEVC intra-frame coding can divide it into 4×4 prediction units, and the size of the prediction unit can range from 64×64 to 4×4. For intra prediction, HEVC provides 35 prediction modes for each prediction unit. Compared with the existing video coding standard H.264/AVC, the optimization problems of coding unit division and prediction direction selection are more complicated. The invention constructs the image complexity description of the current coding tree unit according to the texture analysis, and quickly determines the division of the coding unit according to the texture analysis; utilizes the strong correlation between the texture direction obtained by the texture analysis and the prediction direction, and reduces the number of prediction modes.

Contents of the invention

In order to overcome the above-mentioned defects of the prior art, the present invention proposes a fast mode selection method for HEVC intra prediction based on texture analysis, which determines the division of coding tree units according to the complexity of the video content, and reduces Candidate set for small prediction directions. Experiments prove that this method can well adapt to the real-time compression application of high-definition video images, and greatly improves the encoding speed.

To achieve the above object, the technical solution adopted in the present invention is:

A fast mode selection method for HEVC intra-frame prediction based on texture analysis, the method steps are as follows:

Step (1), for the current coding tree unit, calculate the image complexity and main texture direction of each possible coding unit;

Step (2), for the current coding tree unit, repeat step (3) to step (6) from top to bottom according to the Z scanning order until the current coding tree unit is scanned;

Step (3), if the image complexity of the current coding unit is less than or equal to the threshold, it is determined that the coding unit is a smooth unit, otherwise it is determined as a complex unit;

Step (4), for a smooth coding unit, use step (5) to calculate the rate-distortion cost of the current coding unit;

Step (5), for each prediction unit in the current coding unit, retain the 8 prediction directions around the main texture direction of the unit, planar prediction and DC prediction mode, and perform rough mode selection and rate-distortion optimization selection according to the HEVC video coding standard ;

Step (6), for complex coding units, follow steps (3) to (6) to recursively scan four sub-units in sequence, and if all four sub-units are determined to be smooth units, use step (5) to calculate the rate of the current coding unit Distortion cost, and determine whether the current coding unit is divided according to the principle of minimum rate-distortion cost, and update the threshold at the same time.

The step (1) specifically includes the following steps:

Step (11), for each 4×4 unit in the current coding tree unit, calculate the gradient along the horizontal, vertical, lower left, and lower right directions and calculate the absolute value of the gradient and SAG;

Step (12), taking the gradient direction of the minimum gradient absolute value sum as the main texture direction of the unit, and the main texture direction refers to the texture direction of most pixels of the unit;

Step (13), define the absolute value of the SAG difference between the main texture direction and its orthogonal direction as the texture complexity of the unit, and use the product of the quantization step size and the unit size as the initial threshold.

The step (5) specifically includes the following steps:

Step (51), defining an array of candidate prediction modes;

Step (52), obtain the main texture direction of the current prediction unit according to step (1), and use the main texture direction and eight surrounding prediction directions as candidate prediction directions;

Step (53), mapping the candidate prediction direction to the angle prediction mode in the HEVC coding standard, and adding the selected angle prediction mode, planar and DC mode to the candidate prediction mode array;

Step (54), for each prediction mode in the candidate prediction mode array, perform rough mode selection and rate-distortion optimization mode selection according to the HEVC coding standard.

The step (6) specifically includes the following steps:

Step (61), recursively scan the four sub-units according to step (3) to step (6), if all four sub-units are determined to be smooth units, use step (5) to calculate the rate-distortion cost Cost _coarse of the current coding unit, otherwise Continue to recursively scan the subunits of the next level;

Step (62): If the rate-distortion cost Cost _coarse of the current coding unit is less than or equal to the rate-distortion cost and Cost _fine of the four sub-units, increase the threshold; otherwise, decrease the threshold.

The update strategy for increasing the threshold value in the step (62) is:

TH'＝TH+α×(Cost _fine -Cost _coarse )/2

where TH is the threshold before updating, TH' is the threshold after updating, and α is the multiplication factor of the linear model between texture complexity and rate-distortion cost.

The update strategy for reducing the threshold value in the step (62) is:

TH'=TH-a

In the formula, TH is the threshold before updating, TH' is the threshold after updating, and a is a constant.

The advantage of the present invention compared with prior art is:

1. From the perspective of texture analysis, the present invention considers the division of coding tree units and the selection of prediction direction, and uses the gradient to extract both texture complexity and texture direction. On the premise of ensuring the coding quality, it can save More coding time.

2. The present invention timely updates the threshold for determining whether to divide a coding unit according to the final rate-distortion cost, and can make full use of the strong correlation of adjacent regions in the coding tree unit to weaken the influence of threshold selection on coding quality.

Description of drawings

Fig. 1 is the flow chart of HEVC intra prediction fast mode selection method based on texture analysis;

FIG. 2 is a structural diagram of coding tree unit division in the HEVC video coding standard;

Fig. 3 is a schematic diagram of the method for calculating the direction gradient;

FIG. 4 is a detailed flow chart of gradient-based fast division of coding units.

detailed description

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

In order to meet the real-time compression application requirements of high-definition video images and improve the encoding speed of HEVC video encoders, the present invention proposes a fast mode selection method for HEVC intra-frame prediction based on texture analysis. The method mainly includes two parts: fast partitioning of coding units based on texture complexity and fast selection of prediction mode based on texture direction. The fast partitioning of coding units involves steps (1) to (4) and (6) in the summary of the invention. , the quick selection of the prediction mode involves step (5) in the summary of the invention. First, use the gradient operator to calculate the complexity of each sub-unit in the current coding tree unit, determine a finer coding unit division, and then determine a coarser coding unit division according to the division, and finally according to the rate-distortion cost of the subsequent intra-frame prediction Determine the final coding unit division; in the process of selecting the best prediction direction, exclude the prediction modes that are far from the main texture direction, so as to save the time of intra prediction. The following is a detailed description.

1. Fast division of coding units based on texture complexity

The new-generation video coding standard HEVC introduces the concept of Coding Tree Unit (CTU). A frame of video consists of multiple non-overlapping coding tree units. By default, the size of the coding tree unit is 64×64. The coding tree unit is a recursive structure in the form of a quadtree, which can be flexibly divided into multiple levels of coding units, as shown in FIG. 1 . The size of a coding unit can range from 64×64 to 8×8. In the reference encoder of the HEVC standard, the method of using rate-distortion optimization selection performs a process of prediction, transformation, quantization, entropy coding, inverse quantization, inverse transformation, and reconstruction for each possible division. Compared with H. 264/AVC, the coding complexity is greatly increased, which seriously restricts the practical application of HEVC.

Through experimental statistics, it is found that the size of the coding unit is closely related to the texture complexity of the image content. Regions with simple textures adopt larger coding units, while regions with complex textures adopt smaller coding units. There are many ways to measure texture complexity, such as variance, gradient, entropy, etc. The gradient contains not only the texture complexity information of the image, but also the texture direction information of the image, which can help to quickly determine the prediction mode of the prediction unit. Based on this, the present invention proposes a gradient-based fast coding unit division method.

Before introducing the implementation of fast division of coding units, the calculation method of gradient used in the present invention and the definition of main texture direction and texture complexity are firstly given. The gradient in the present invention refers to the directional derivative of the pixel brightness value obtained along the horizontal, vertical, lower left, lower right directions, etc., as shown in FIG. 3 . The calculation formula is shown in formula (1)-(4).

$\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; )</span>\\ {\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; h</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">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</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>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span>\\ {\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; h</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</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><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</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><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>\\ {\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; h</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</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>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

$\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</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">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; v</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">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>\\ {\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</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><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>\\ {\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; v</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; p</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>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</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>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

$\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; ld</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\\ {\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; ld</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">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</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>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\\ {\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; ld</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</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><span\; class="notranslate">\; =</span>\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</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><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\\ {\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; ld</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</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>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

$\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; rd</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</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><span\; class="notranslate">\; ]</span>\\ {\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; rd</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">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; p</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>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\\ {\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; rd</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</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><span\; class="notranslate">\; =</span>\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</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><span\; class="notranslate">\; ]</span>\\ {\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; rd</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; p</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>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</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">\; the\; y</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

where g _h ( ), g _v ( ), g _ld ( ), and g _rd ( ) represent the gradient filters in the horizontal, vertical, lower-left, and lower-right directions, respectively, and p( ) represents the current coding tree unit The pixel value of the luminance component in , (x, y) represents the coordinates of the upper left corner pixel of each 4×4 unit. In the present invention, SAG is used to represent the absolute sum of gradients along a certain direction, and the calculation formula is shown in formula (5).

SAG _dir =∑|g _dir (x,y)|,dir∈{h,v,ld,rd} (5)

The direction of the minimum gradient absolute value sum is taken as the main texture direction, the gradient absolute value sum of the main texture direction is denoted as SAG _min , and the gradient absolute value sum of its orthogonal direction is denoted as SAG _org . The texture complexity complexity is defined as the difference between SAG _org and SAG _min , namely:

complexity = SAG _org - SAG _min (6)

The detailed process of partitioning is shown in Figure 4. For each coding tree unit, it is firstly divided into N×N 4×4 units, and the gradient values in the horizontal, vertical, lower left, and lower right directions are calculated for each unit. Then recursively calculate the texture complexity and division of each coding tree unit from 64×64 to 8×8, and finally determine the division of the entire coding tree unit. The recursive process is described as follows:

(1) For the current coding unit, use formula (5) to calculate the absolute value of the gradient in the four directions and SAG _h , SAG _v , SAG _ld , SAG _rd , and obtain the main texture direction and texture complexity complexity of the current coding unit .

(2) If the texture complexity complexity of the current coding unit is less than the threshold TH, directly set the current coding unit as a smooth area, and calculate the rate-distortion cost, no longer divide, and return to the previous layer of recursion.

(3) Otherwise, check whether the current coding unit can be divided, if not, calculate the rate-distortion cost and return to the previous layer of recursion.

(4) Otherwise, divide the current coding unit into four subunits, and enter the next layer of recursion in turn.

(5) Check whether all four sub-units are determined to be smooth areas, if they are smooth areas, calculate the rate-distortion cost of the current coding unit, otherwise go to step (8).

(6) If the rate-distortion cost of the current coding unit is less than or equal to the sum of the rate-distortion costs of the four sub-units, the current coding unit is not divided, and the threshold TH is increased.

(7) If the rate-distortion cost of the current coding unit is greater than the sum of the rate-distortion costs of the four sub-units, the current coding unit is divided, and the threshold TH is fine-tuned at the same time.

(8) Detect whether there are unencoded units, if so, return to the previous level of recursion, otherwise exit the recursion.

Generally speaking, the distortion produced by the encoder corresponding to the existing video coding standards is mainly caused by quantization. If the difference between adjacent pixels is smaller than the quantization step size, it means that the difference will not exist after quantization. Therefore, the threshold of texture complexity can be set through the quantization step size, but the quantization step size in the video coding standard is for the transform domain. According to the law of energy conservation between space domain and transform domain, the present invention defines the initial threshold in the recursive process as the product of the size of the current coding unit and the quantization step, ie TH=CUsize×QPstep.

The update strategy for increasing the threshold TH is: TH′=TH+α×(Cost _fine -Cost _coarse )/2, where α is the multiplication factor of the linear model between the texture complexity and the rate-distortion cost.

The update strategy for fine-tuning the threshold TH is as follows: TH' is TH'=TH-a, where a is a constant.

In this way, the division of coding tree units can be obtained according to texture complexity and rate-distortion cost.

2. Fast selection of prediction mode based on texture direction

Compared with H.264/AVC, HEVC, a new generation of video coding standard, introduces the concept of prediction unit (Prediction Unit, PU). For intra prediction, there are two ways to divide the prediction unit. For the smallest coding unit, the prediction unit can be the same as the coding unit, or can be further divided into four branches on the basis of the coding unit; while for the remaining coding units, the prediction unit can only be the same as the coding unit. At the same time, HEVC also provides a more refined prediction direction. Each prediction unit has 35 prediction modes, which greatly increases the coding complexity.

The root of improving the coding efficiency is to minimize the number of candidate prediction modes under the premise of ensuring the coding quality. Usually, the final selected prediction mode is close to the texture direction. Table 1 gives the probability that the best prediction mode is close to the texture direction in various prediction units. It can be seen from the table that the probability that the best prediction mode is similar to the texture direction can reach more than 80%. In addition, in the first step of fast division of coding units in the present invention, the main texture direction of each coding unit has been obtained. Therefore, this intermediate result can be used to minimize the number of candidate prediction modes, thereby improving coding efficiency. It should be noted that, for all prediction units, the planar and DC prediction modes provided by HEVC are reserved, and the present invention only considers the candidate prediction directions of reduced angle prediction.

Table 1 The probability that the best prediction mode of each prediction unit is similar to the texture direction under different quantization steps

Quantization step size 64×64 32×32 16×16 8×8 4×4 twenty two 100% 79.63% 80.76% 84.32% 87.45% 27 86.36% 78.80% 83.64% 87.67% 89.24% 32 80.43% 82.21% 89.47% 88.88% 89.86% 37 78.67% 86.53% 89.83% 89.20% 89.06% average 86.37% 81.79% 85.93% 87.52% 88.90%

Since the probability that the best prediction mode is close to the texture direction is high, the present invention uses several prediction directions around the main texture direction for angle prediction. The specific implementation is as follows:

(1) Define an array of candidate prediction modes with a size of 11, which is used to store the selected candidate prediction modes, including planar mode, DC mode, main texture direction and several prediction directions around the main texture direction. In the specific implementation of the present invention, four prediction directions are taken on both sides of the main texture direction.

(2) Obtain the prediction mode number mode _texture corresponding to the main texture direction of the current prediction unit, and use the formula (7) to calculate the angle prediction range [mode _left t, mode _right ].

$\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; mode</span>}}_{\mathrm{<span\; class="notranslate">\; left</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; mode</span>}}_{\mathrm{<span\; class="notranslate">\; texture</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 4</span>}\\ {\mathrm{<span\; class="notranslate">\; mode</span>}}_{\mathrm{<span\; class="notranslate">\; right</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; mode</span>}}_{\mathrm{<span\; class="notranslate">\; texture</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 4</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; )</span>$

(3) Use formula (8) to obtain the prediction mode on the left side of the main texture direction.

$\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; mode</span>}\mathrm{<span\; class="notranslate">\; Idx</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; mode</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 33</span>}& \mathrm{<span\; class="notranslate">\; mode</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; 2</span>}\\ \mathrm{<span\; class="notranslate">\; mode</span>}\mathrm{<span\; class="notranslate">\; Idx</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; mode</span>}& \mathrm{<span\; class="notranslate">\; mode</span>}<span\; class="notranslate">\; \&Greater\; Equal;</span>\mathrm{<span\; class="notranslate">\; 2</span>}\end{array}\right.<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; mode</span>}<span\; class="notranslate">\; \&Element;</span><span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; mode</span>}}_{\mathrm{<span\; class="notranslate">\; left</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; mode</span>}}_{\mathrm{<span\; class="notranslate">\; texture</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 8</span>}<span\; class="notranslate">\; )</span>$

(4) Use formula (9) to obtain the prediction mode on the right side of the main texture direction.

$\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; mode</span>}\mathrm{<span\; class="notranslate">\; Idx</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; mode</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 33</span>}& \mathrm{<span\; class="notranslate">\; mode</span>}<span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; 34</span>}\\ \mathrm{<span\; class="notranslate">\; mode</span>}\mathrm{<span\; class="notranslate">\; Idx</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; mode</span>}& \mathrm{<span\; class="notranslate">\; mode</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; 34</span>}\end{array}\right.<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; mode</span>}<span\; class="notranslate">\; \&Element;</span><span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; mode</span>}}_{\mathrm{<span\; class="notranslate">\; texture</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; mode</span>}}_{\mathrm{<span\; class="notranslate">\; right</span>}}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 9</span>}<span\; class="notranslate">\; )</span>$

(5) Store the filtered candidate prediction mode set {modeIdx} in the candidate prediction mode array. Select the best prediction mode according to the coarse mode selection and rate-distortion optimization selection in the HEVC standard.

The above disclosures are only specific embodiments of the present invention. According to the technical ideas provided by the present invention, all changes conceivable by those skilled in the art shall fall within the protection scope of the present invention.

Claims (4)

1. a HEVC intra-frame prediction quick mode selection method based on texture analysis, it is characterised in that: the method step As follows:

Step (1), to present encoding tree unit, calculate the image complexity of each possible coding unit and main grain direction；

Step (2), to present encoding tree unit, according to Z scanning sequency repeated execution of steps from top to bottom (3) to step (6), Until scanning through present encoding tree unit；

Step (3) if the image complexity of current coded unit is less than or equal to threshold value, then judges that this coding unit is as smooth single Unit, is otherwise judged to complex unit；

Step (4), for smooth coding unit, utilize step (5) to calculate the rate distortion costs of current coded unit；

Step (5), for each predicting unit in current coded unit, only retain 8 around the main grain direction of this unit Individual prediction direction and planar prediction and DC predictive mode, according to HEVC video encoding standard carry out coarse mode select and Rate-distortion optimization selects；

Step (6), for complicated coding unit, according to step (3) to four subelements of step (6) successively recursive scanning, If four subelements are all judged to smooth unit, then step (5) is utilized to calculate the rate distortion costs of current coded unit, and root Judge whether current coded unit divides, and updates threshold value simultaneously according to rate distortion costs minimum principle；

Described step (6) specifically includes following steps:

Step (61), according to step (3) to four subelements of step (6) successively recursive scanning, if four subelements are all sentenced It is set to smooth unit, then utilizes step (5) to calculate rate distortion costs Cost of current coded unit_coarse, otherwise continue recurrence Scan the subelement of next level；

Step (62) if rate distortion costs Cost of current coded unit_coarseRate distortion generation less than or equal to four subelements Valency and Cost_fine, then threshold value is tuned up, otherwise threshold value is turned down；

The more New Policy that in described step (62), threshold value tunes up is:

TH '=TH+ α × (Cost_fine-Cost_coarse)/2

In formula, TH is the threshold value before updating, and TH ' is the threshold value after updating, and α is the line between Texture complication and rate distortion costs The multiplication factor of property model.

2. HEVC intra-frame prediction quick mode selection method based on texture analysis as claimed in claim 1, its feature exists Following steps are specifically included in: described step (1):

Step (11), to each unit of 4 × 4 in present encoding tree unit, respectively along level, vertical, lower-left, the right side Lower four direction calculates gradient and calculates absolute value and the SAG of gradient；

Step (12), using the gradient direction of minimal gradient absolute value sum as the main grain direction of this code tree unit, described master Grain direction refers to the grain direction of this code tree unit major part pixel；

Step (13), the absolute value of SAG difference on orthogonal to that for main grain direction direction is defined as the Texture complication of this unit, And utilize the quantization step product with this cell size as initial threshold.

3. HEVC intra-frame prediction quick mode selection method based on texture analysis as claimed in claim 1, its feature exists Following steps are specifically included in: described step (5):

Step (51), one candidate modes array of definition；

Step (52), obtain the main grain direction of current prediction unit according to step (1), and by main grain direction and about 8 prediction direction as candidate prediction direction；

Step (53), the angle predictive mode that candidate prediction direction is mapped in HEVC coding standard, and the angle that will select Degree predictive mode and planar and DC pattern add candidate modes array to；

Step (54), to each predictive mode in candidate modes array, carry out roughly according to HEVC coding standard Model selection and R-D optimized mode selection.

4. HEVC intra-frame prediction quick mode selection method based on texture analysis as claimed in claim 1, its feature exists In: the more New Policy that in described step (62), threshold value is turned down is:

TH '=TH-a

In formula, TH is the threshold value before updating, and TH ' is the threshold value after updating, and a is constant.