WO2018010492A1 - Rapid decision making method for intra-frame prediction mode in video coding - Google Patents

Rapid decision making method for intra-frame prediction mode in video coding Download PDF

Info

Publication number
WO2018010492A1
WO2018010492A1 PCT/CN2017/085605 CN2017085605W WO2018010492A1 WO 2018010492 A1 WO2018010492 A1 WO 2018010492A1 CN 2017085605 W CN2017085605 W CN 2017085605W WO 2018010492 A1 WO2018010492 A1 WO 2018010492A1
Authority
WO
WIPO (PCT)
Prior art keywords
mode
prediction mode
pml
prediction
rate distortion
Prior art date
Application number
PCT/CN2017/085605
Other languages
French (fr)
Chinese (zh)
Inventor
董胜富
王萌
王文敏
王振宇
王荣刚
李英
赵辉
Original Assignee
北京大学深圳研究生院
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京大学深圳研究生院 filed Critical 北京大学深圳研究生院
Publication of WO2018010492A1 publication Critical patent/WO2018010492A1/en

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
    • H04N19/103—Selection of coding mode or of prediction mode
    • H04N19/11—Selection of coding mode or of prediction mode among a plurality of spatial predictive coding modes
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/136—Incoming video signal characteristics or properties
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/146—Data rate or code amount at the encoder output
    • H04N19/147—Data rate or code amount at the encoder output according to rate distortion criteria
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
    • H04N19/176—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/593—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving spatial prediction techniques

Definitions

  • the present invention relates to a video coding method, and more particularly to a method for fast decision making in an intra prediction mode in video coding.
  • Predictive coding is one of the core technologies in video coding.
  • the coding algorithm uses the spatial correlation between adjacent pixels of the video signal and the temporal correlation between adjacent images to predictively encode, eliminate redundancy in the video signal, and achieve efficient compression.
  • the new generation of video coding technology uses two methods of intra prediction and inter prediction. Among them, intra prediction uses spatial correlation to predict uncoded pixels using coded pixels, eliminating spatial redundancy in the video signal. By analyzing the information of the current coded block and the adjacent coded block, the best prediction of the current block is obtained by using the angle mode and the plane mode.
  • the New Generation Video Coding adopts a recursive coding tree structure and proposes the concepts of Coding Unit (CU), Prediction Unit (PU) and Transform Unit.
  • the complex block partitioning mechanism allows the encoder to adaptively select the encoding mode based on the video content characteristics.
  • the prediction of the luminance block uses up to 35 prediction modes, including 33 angle modes and 2 kinds. Non-angle mode.
  • the improvement of coding performance brought by the above two aspects is based on increasing the complexity of the algorithm, and the rate-distortion optimization (RDO) of all prediction modes is performed for each block-sized unit. huge.
  • the proposed JCTVC-D283 standard complements the Rough Mode Design (RMD) algorithm for intra prediction and the fast mode decision algorithm for selecting the Most Possible Mode (MPM).
  • RMD Rough Mode Design
  • MPM Most Possible Mode
  • the present invention provides a new method for intra prediction fast mode decision making, which fully utilizes the main feature information of the coding block, reduces the computational complexity of intra prediction, and improves on the premise of ensuring coding performance.
  • the speed of video encoding is a new method for intra prediction fast mode decision making, which fully utilizes the main feature information of the coding block, reduces the computational complexity of intra prediction, and improves on the premise of ensuring coding performance.
  • the principle of the present invention is that in HEVC intra prediction, the coefficients in the 33 angular modes are ⁇ 2, 6, 10, 14, 18, 22, 26,
  • the set of direction patterns of 30, 34 ⁇ can representatively represent the texture information of the current block, and the present invention uses the above-mentioned set of direction patterns as a principal prediction orientation set (Principal Prediction Orientation); two predictions that define an absolute value of the coefficient difference of 4 The orientations are adjacent; the three orientations with the least cost are selected and recorded as M0, M1, and M2, respectively.
  • a prediction mode candidate list (PML) is constructed, and if the least cost is two adjacent prediction directions ⁇ M0, M1 ⁇ (assuming that M0 is a mode with a small cost and a large mode coefficient), then further The three modes separated by 1 between the two modes, the two modes respectively outward, and the two non-angle modes of Planar and DC are added to the candidate list ⁇ M0, M1, 0, 1, M0-1 , M0-2, M0-3, M0+1, M0+2, M1-1, M1-2 ⁇ ; if the least cost is two non-adjacent prediction directions, then ⁇ M0, M1 ⁇ are adjacent A mode with an interval of 1, 2, a mode with a ⁇ M2 ⁇ interval of 2, and two non-angle modes of Planar and DC are added to the candidate list ⁇ M0, M1, 0, 1, M0-1, M0-2, M0+1, M0+2, M1-1, M1-2, M1+1, M1+2, M2, M2+2, M2-2 ⁇ .
  • the number of prediction modes is determined according to the size of the prediction unit PU, that is, the candidate prediction mode in which the rate distortion optimization is finally performed.
  • the candidate prediction mode in which the rate distortion optimization is finally performed For PU blocks larger than 8 ⁇ 8, only the two modes with the least cost in PML are selected for rate-distortion optimization, and the other cases are selected with the least cost-optimized selection.
  • a fast decision method for intra prediction mode in video coding which determines the prediction orientation, uses the correlation of the angle mode and the size of the prediction unit PU to construct a prediction mode candidate list, thereby reducing the number of prediction modes for rate distortion optimization, thereby Reduce the computational complexity of intra prediction and improve the video encoding speed; including the following steps:
  • Step 1 For the current PU block of any size, predict the nine directions of the mode coefficients of the main predicted azimuth set as ⁇ 2, 6, 10, 14, 18, 22, 26, 30, 34 ⁇ , and calculate the rough rate. Distortion cost; sorting the patterns of the main predicted azimuth set according to the rate distortion cost, selecting the first three modes M0, M1 and M2 with the best rate distortion cost, and the rate distortion cost of the M0, M1 and M2 satisfies J(M0) ⁇ J(M1) ⁇ J(M2); the orientation relationship of M0, M1 and M2 is adjacent or not adjacent;
  • Step 2 When M0 is adjacent to M1, the optimal direction of prediction is near M0 and M1. At this time, three modes (M0-1, M0-2, M0-3) and M0 and M1 between M0 and M1 are calculated. The adjacent four modes (M0+1, M0+2, M1-1, M1-2) are added, and the DC mode and the Planar mode are added to the prediction mode candidate list PML; when M0 is not adjacent to M1 , respectively, the ⁇ M0, M1 ⁇ mode with adjacent intervals of 1 and 2, the mode of ⁇ M2 ⁇ interval 2, and the two non-angle modes of Planar and DC are added to the prediction mode candidate list PML;
  • Step 3 Calculate the rate distortion cost in the prediction mode candidate list PML.
  • the block size of the PU block is 4 ⁇ 4 or 8 ⁇ 8, only the two modes with the lowest rate distortion cost in the prediction mode candidate list PML are retained. Optimization; when the block size of the PU block is 16 ⁇ 16, 32 ⁇ 32 or 64 ⁇ 64, three modes with the lowest rate distortion cost in the prediction mode candidate list PML are selected for rate distortion optimization;
  • Step 4 adding the pattern in the most probable prediction mode candidate list MPM to the prediction mode candidate list PML without repetition;
  • Step 5 Perform last rate-distortion optimization on the mode in the prediction mode candidate list PML obtained in step 4; use the mode with the lowest rate-distortion cost as the optimal mode of intra prediction of the current PU block.
  • the main predicted azimuth set in step 1 is the HEVC intra prediction angle mode coefficient is ⁇ 2, 6, 10, 14, 18, 22, 26, 30 , 34 ⁇ the set of direction modes, including nine orientations.
  • the two predicted azimuths are adjacent.
  • step 1 specifically calculates the coarse rate distortion cost by using a Hada code conversion method.
  • step 2 when M0 is adjacent to M1, M0 is set to a larger mode, and the prediction mode candidate list PML is updated to ⁇ M0, M1, 0. , 1M0-1, M0-2, M0-3, M0+1, M0+2, M1-1, M1-2 ⁇ ; when M0 is not adjacent to M1, the prediction mode candidate list PML is updated to ⁇ M0, M1 , 0, 1, M0-1, M0-2, M0+1, M0+2, M1-1, M1-2, M1+1, M1+2, M2, M2+2, M2-2 ⁇ .
  • the coefficients in the range of the angle mode are acquired by loop, and then added to Prediction mode candidate list PML.
  • the rate distortion cost is calculated by using the Hada code conversion method.
  • the final rate distortion optimization performed in step 5 includes the following steps:
  • the invention provides a new intra prediction fast mode decision method, fully utilizes the main feature information of the coding block, reduces the computational complexity of the intra prediction, and improves the video coding speed under the premise of ensuring the coding performance.
  • the invention has the following advantages:
  • the present invention makes full use of the correlation between the angle modes by judging the predicted orientation, and reduces the number of modes for performing the Hadamard transform, thereby reducing the computational complexity of the prediction process.
  • the present invention constructs a prediction mode candidate list, which can make a faster and more accurate judgment on the direction mode, and reduces the number of modes for performing rate distortion optimization for the size of the PU.
  • FIG. 1 is a schematic diagram of prediction point coordinates and reference point selection of HEVC intra prediction
  • each square represents one pixel; Rx, y represents a predicted reference pixel; Px, y represents a predicted pixel; x and y represent the position of the pixel, and the upper left corner is a (0, 0) origin.
  • FIG. 2 is a flow chart of a method for quickly determining an intra prediction mode provided by the present invention.
  • FIG. 3 is a schematic diagram of a direction prediction mode of HEVC and a main predicted orientation set of the present invention
  • the black solid line indicates the main predicted orientation set of the present invention, specifically, the nine prediction mode sets whose coefficients are ⁇ 2, 6, 10, 14, 18, 22, 26, 30, 34 ⁇ .
  • FIG. 4 is a schematic diagram of candidate prediction modes when primary prediction azimuths are adjacent in an embodiment of the present invention
  • M0 is a larger mode
  • the candidate prediction mode list is ⁇ M0, M1, 0, 1M0-1, M0-2, M0-3, M0+1, M0+2, M1-1, M1-2 ⁇ .
  • FIG. 5 is a schematic diagram of candidate prediction modes when primary prediction azimuths are not adjacent in an embodiment of the present invention.
  • ⁇ M0, M1 ⁇ when M0 and M1 are not adjacent, respectively, a mode in which ⁇ M0, M1 ⁇ is adjacent to interval 1 or 2, a mode in which ⁇ M2 ⁇ interval is 2, and two non-angle modes in Planar and DC are added to the candidate prediction mode list.
  • the invention provides a fast decision method for intra prediction mode in video coding, which determines a prediction mode by using a correlation of an angle mode and a size of a prediction unit PU to reduce a prediction mode for performing rate distortion optimization.
  • the number thereby reducing the computational complexity of intra prediction and improving the video encoding speed.
  • the following embodiment describes an intra prediction implementation process that encodes a current frame luma block, constructing a prediction block that differs the least from the original luma block.
  • CU Coding Unit
  • LCU maximum coding unit
  • SCU minimum coding unit
  • Intra prediction performs mode decision and block partitioning on the current CU, and multiple recursive selection of the best partitioning method and optimal mode.
  • Prediction Unit The same as the size of the current CU. When the current CU is an LCU, it can continue to be divided into four smaller prediction blocks.
  • HEVC Luminance Component Intra Prediction supports five sizes of PUs: 4x4, 8x8, 16x16, 32x32, and 64x64.
  • HEVC intra prediction includes 35 prediction modes, which are different predictions of the current block according to certain rules by the coded pixel values of the left adjacent column and the upper adjacent row of the current block. the way. These include 33 angle prediction modes and 2 non-angle prediction modes.
  • Non-angle prediction includes direct current (DC) prediction and plane (Planar) prediction. The coordinates of the predicted points and the selection of the reference points are shown in Fig. 1.
  • the angle prediction mode has a coefficient of 2 to 34, and the DC mode and the Planar mode have coefficients of 0 and 1, respectively.
  • MCM Probable Prediction Mode
  • Rate Distortion Optimization HEVC uses the rate-distortion optimization process to measure the rate and distortion performance of different prediction modes and select the optimal prediction mode.
  • Intra prediction is to obtain the optimal mode and prediction mode of different divided CUs by mode decision, and complete the prediction process.
  • the flow of the present invention is shown in Figure 2.
  • the core method of intra mode decision includes the following steps:
  • Step 1 predicting 9 kinds of orientations of mode coefficients of the main predicted azimuth set of ⁇ 2, 6, 10, 14, 18, 22, 26, 30, 34 ⁇ for the current PU block of any size, in order to improve the coding speed,
  • the Hada code transform is used instead of the discrete cosine transform to calculate the coarse rate distortion cost after the Hada code transform.
  • the cost function is shown in Equation 1, where J is the rate distortion cost, MODE is the coefficient (0-35) of all modes of HEVC intra prediction, ⁇ is the Lagrangian daily quantity, and SATD (Sum of Absolute Transform Difference) is The sum of the absolute values of the block prediction mode and the original block pixel value after the Hada code is transformed.
  • R MODE indicates the current mode residual
  • B MODE indicates the number of bits that the mode needs to encode.
  • the rate distortion cost calculated according to Equation 1 is ordered by the mode of the main predicted azimuth set, and the rate distortion cost is selected therefrom.
  • the HEVC intra prediction all mode and the main predicted azimuth set proposed by the present invention are as shown in FIG. 3, and the main predicted azimuth set is 9 coefficients having coefficients of ⁇ 2, 6, 10, 14, 18, 22, 26, 30, 34 ⁇ .
  • the set of prediction modes is indicated by a solid black line in FIG.
  • Step 2 If M0 and M1 are adjacent, that is, the absolute value of the difference is 4, indicating that the predicted optimal direction is near M0 and M1, then the three modes between M0 and M1 are further calculated (M0-1, M0-2, M0-3) and four adjacent modes of M0 and M1 (M0+1, M0+2, M1-1, M1-2), as shown in Figure 4, and DC mode and The Planar mode is added to the prediction mode candidate list PML. Assuming M0 is a larger mode, the list is updated to ⁇ M0, M1, 0, 1M0-1, M0-2, M0-3, M0+1, M0+2, M1-1, M1-2 ⁇ . If M0 and M1 are not adjacent, that is, the absolute value of the difference is greater than 4, as shown in FIG.
  • Step 3 Calculate the rate distortion penalty in PML according to Equation 1. For PU blocks less than or equal to 8 ⁇ 8, that is, the block size is 4 ⁇ 4 or 8 ⁇ 8, only the two modes with the lowest rate distortion cost in the list are retained, otherwise, the block size is 16 ⁇ 16, 32 ⁇ 32 or At 64 ⁇ 64, the choice of the three least-rate-rate-optimized optimizations is selected.
  • Step 4 Using the MPM, add the patterns in its candidate list to the PML list without duplication.
  • step 5 the last rate-distortion optimization is performed on the mode in the PML list, and the mode with the least cost is the optimal mode of intra prediction of the current block, which specifically includes:
  • This step includes the process of transforming, quantifying, and reconstructing
  • Equation 2 Calculate the rate distortion cost of each predicted angle mode in the PML list by Equation 2:
  • Equation 2 J is the rate distortion cost, and D(Mode) and B(Mode) respectively represent the distortion and the number of bits when using different angle modes;
  • the mode in which the rate distortion cost is the smallest among the rate distortion costs obtained in step 52) is taken as the optimal intra prediction direction mode.
  • the specific implementation of the present invention is to implement and test the HEVC general test sequence including Class A, Class B, Class C, Class D and Class E on the HEVC reference software HM16.0, configured as a full I frame, and the HEVC standard specifies 52 quantization steps. Long, corresponding to 52 quantization parameters (QP). According to the general condition, the quantization parameter QP is set to 22, 27, 32, 37, respectively.
  • QP quantization parameters
  • Table 1 represents the luminance component
  • U and V represent the chrominance components, wherein the average BD-rate of the luminance portion is slightly increased by 0.65%, and the overall time saving (TS) is 26.76%.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Abstract

Disclosed in the present invention is a rapid decision making method for intra-frame prediction mode in video coding. The number of prediction modes of performing rate-distortion optimization is reduced by means of determining a prediction orientation and constructing a PML using a correlation of an angle mode and a prediction unit (PU) size; comprising: predicting orientations in a principal prediction orientation set, and selecting first three modes of optimal rate-distortion cost; adding a mode in an adjacent direction of the first three modes, a Planar mode and a DC mode into a PML; calculating the rate-distortion cost of the modes in the PML; adding modes in an MPM into the PML without repetition; performing final rate-distortion optimization; and using the mode with the lowest rate-distortion cost as an optimal mode of intra-frame prediction of a current PU block. The present invention may reduce operation complexity of intra-frame prediction, and may improve speed of video coding.

Description

视频编码中帧内预测模式的快速决策方法Fast decision method for intra prediction mode in video coding 技术领域Technical field
本发明涉及视频编码方法，尤其涉及一种视频编码中针对帧内预测模式快速决策的方法。The present invention relates to a video coding method, and more particularly to a method for fast decision making in an intra prediction mode in video coding.
背景技术Background technique
视频编码技术解决的重点问题是数字视频海量数据的编码压缩问题。预测编码是视频编码中的核心技术之一。编码算法利用视频信号相邻像素间的空间相关性和相邻图像之间的时间相关性进行预测编码，消除视频信号中的冗余，实现高效压缩。新一代视频编码技术采用帧内预测和帧间预测两种方法。其中，帧内预测利用空间相关性，使用已编码像素预测未编码像素，消除视频信号中的空间冗余。通过分析当前编码块与相邻已编码块的信息，利用角度模式以及平面模式获取当前块的最佳预测。The key issue solved by video coding technology is the coding compression problem of digital video massive data. Predictive coding is one of the core technologies in video coding. The coding algorithm uses the spatial correlation between adjacent pixels of the video signal and the temporal correlation between adjacent images to predictively encode, eliminate redundancy in the video signal, and achieve efficient compression. The new generation of video coding technology uses two methods of intra prediction and inter prediction. Among them, intra prediction uses spatial correlation to predict uncoded pixels using coded pixels, eliminating spatial redundancy in the video signal. By analyzing the information of the current coded block and the adjacent coded block, the best prediction of the current block is obtained by using the angle mode and the plane mode.
新一代视频编码标准(High Efficiency Video Coding,HEVC)采用了递归编码树结构，并提出了编码单元(Coding Unit,CU)、预测单元(Prediction Unit,PU)和变换单元(Transform Unit)的概念,复杂的块划分机制使得编码器根据视频内容特性自适应选择编码模式。同时，为了更好地匹配视频中的复杂纹理，更好地去除空间冗余，在HEVC的帧内预测中，亮度块的预测采用了多达35种预测模式，包括33种角度模式和2种非角度模式。以上两方面所带来的编码性能的提升，是以增加算法复杂度为代价，为每一个块大小的单元都进行全部预测模式的率失真最优选择(Rate-Distortion optimization，RDO)，计算量巨大。The New Generation Video Coding (HEVC) adopts a recursive coding tree structure and proposes the concepts of Coding Unit (CU), Prediction Unit (PU) and Transform Unit. The complex block partitioning mechanism allows the encoder to adaptively select the encoding mode based on the video content characteristics. At the same time, in order to better match the complex texture in the video and better remove the spatial redundancy, in the intra prediction of HEVC, the prediction of the luminance block uses up to 35 prediction modes, including 33 angle modes and 2 kinds. Non-angle mode. The improvement of coding performance brought by the above two aspects is based on increasing the complexity of the algorithm, and the rate-distortion optimization (RDO) of all prediction modes is performed for each block-sized unit. huge.
该标准已采纳的提案JCTVC-D283，为帧内预测补充了粗略模式决策(Rough Mode Design，RMD)算法以及选取最可能模式(Most Possible Mode，MPM)的快速模式决策算法。首先通过Hadamard(哈达玛)变换及码率估计，根据PU尺寸选择3或8个模式，另外加入相邻块预测模式的方法，来减少率失真优化的模式个数，达到了一定的加速效果。但是，这些方法不能充分利用角度模式之间的相关性，HEVC的帧内预测过程仍然有很大的提速空间。The proposed JCTVC-D283 standard complements the Rough Mode Design (RMD) algorithm for intra prediction and the fast mode decision algorithm for selecting the Most Possible Mode (MPM). Firstly, through Hadamard transform and rate estimation, 3 or 8 modes are selected according to the PU size, and the adjacent block prediction mode is added to reduce the number of modes of rate distortion optimization, and a certain acceleration effect is achieved. However, these methods cannot fully utilize the correlation between the angle modes, and the intra prediction process of HEVC still has a large speed-up space.
发明内容Summary of the invention
为了克服上述现有技术的不足，本发明提供一种新的帧内预测快速模式决策的方法，充分利用编码块主要特征信息，降低帧内预测的运算复杂度，在保证编码性能的前提下提高视频编码的速度。In order to overcome the above deficiencies of the prior art, the present invention provides a new method for intra prediction fast mode decision making, which fully utilizes the main feature information of the coding block, reduces the computational complexity of intra prediction, and improves on the premise of ensuring coding performance. The speed of video encoding.
本发明的原理是：在HEVC帧内预测中，33种角度模式中系数为{2,6,10,14,18,22，26, 30,34}的方向模式集合能够有代表性地表征当前块的纹理信息，本发明将上述方向模式集合作为主要预测方位集(Principal Prediction Orientation)；定义系数差的绝对值为4的两个预测方位是相邻的；选取代价最小的3个方位，分别记为M0、M1、M2。然后，构建预测模式候选列表(Prediction Mode List，PML)，如果代价最小的是两个相邻的预测方位{M0,M1}(假设M0为代价较小、模式系数较大的模式)，则进一步将该两种模式之间的相隔为1的三种模式、其分别向外的两种模式、和Planar、DC两个非角度模式加入候选列表中{M0,M1,0,1,M0-1,M0-2,M0-3,M0+1,M0+2,M1-1,M1-2}；如果代价最小的是两个不相邻的预测方位，则分别将{M0,M1}相邻间隔为1、2的模式、{M2}间隔为2的模式和Planar、DC两个非角度模式加入候选列表中{M0,M1,0,1,M0-1,M0-2,M0+1,M0+2,M1-1,M1-2,M1+1,M1+2,M2,M2+2,M2-2}。之后，根据预测单元PU的尺寸确定预测模式个数，即最终进行率失真优化的候选预测模式。对于大于8×8的PU块，仅选取PML中的代价最小的2种模式进行率失真优化选择，其余情况选取代价最小的3个进行率失真优化选择。The principle of the present invention is that in HEVC intra prediction, the coefficients in the 33 angular modes are {2, 6, 10, 14, 18, 22, 26, The set of direction patterns of 30, 34} can representatively represent the texture information of the current block, and the present invention uses the above-mentioned set of direction patterns as a principal prediction orientation set (Principal Prediction Orientation); two predictions that define an absolute value of the coefficient difference of 4 The orientations are adjacent; the three orientations with the least cost are selected and recorded as M0, M1, and M2, respectively. Then, a prediction mode candidate list (PML) is constructed, and if the least cost is two adjacent prediction directions {M0, M1} (assuming that M0 is a mode with a small cost and a large mode coefficient), then further The three modes separated by 1 between the two modes, the two modes respectively outward, and the two non-angle modes of Planar and DC are added to the candidate list {M0, M1, 0, 1, M0-1 , M0-2, M0-3, M0+1, M0+2, M1-1, M1-2}; if the least cost is two non-adjacent prediction directions, then {M0, M1} are adjacent A mode with an interval of 1, 2, a mode with a {M2} interval of 2, and two non-angle modes of Planar and DC are added to the candidate list {M0, M1, 0, 1, M0-1, M0-2, M0+1, M0+2, M1-1, M1-2, M1+1, M1+2, M2, M2+2, M2-2}. Thereafter, the number of prediction modes is determined according to the size of the prediction unit PU, that is, the candidate prediction mode in which the rate distortion optimization is finally performed. For PU blocks larger than 8×8, only the two modes with the least cost in PML are selected for rate-distortion optimization, and the other cases are selected with the least cost-optimized selection.
本发明提供的技术方案是：The technical solution provided by the invention is:
一种视频编码中帧内预测模式的快速决策方法，该方法通过判断预测方位，利用角度模式的相关性和预测单元PU的尺寸构造预测模式候选列表，减少进行率失真优化的预测模式数目，从而降低帧内预测的运算复杂度，提高视频编码速度；包括如下步骤：A fast decision method for intra prediction mode in video coding, which determines the prediction orientation, uses the correlation of the angle mode and the size of the prediction unit PU to construct a prediction mode candidate list, thereby reducing the number of prediction modes for rate distortion optimization, thereby Reduce the computational complexity of intra prediction and improve the video encoding speed; including the following steps:
步骤1，对当前任意尺寸的PU块，对主要预测方位集中的模式系数为{2,6,10,14,18,22，26,30,34}的九种方位进行预测，计算得到粗略率失真代价；根据率失真代价对主要预测方位集中的模式排序，从中选取率失真代价最优的前三个模式M0、M1和M2，所述M0、M1和M2的率失真代价满足J(M0)<J(M1)<J(M2)；所述M0、M1和M2的方位关系为相邻或不相邻； Step 1. For the current PU block of any size, predict the nine directions of the mode coefficients of the main predicted azimuth set as {2, 6, 10, 14, 18, 22, 26, 30, 34}, and calculate the rough rate. Distortion cost; sorting the patterns of the main predicted azimuth set according to the rate distortion cost, selecting the first three modes M0, M1 and M2 with the best rate distortion cost, and the rate distortion cost of the M0, M1 and M2 satisfies J(M0) <J(M1)<J(M2); the orientation relationship of M0, M1 and M2 is adjacent or not adjacent;
步骤2，当M0与M1相邻时，预测最优方向在M0和M1附近，此时计算M0与M1之间的三个模式(M0-1,M0-2,M0-3)和M0与M1之外的相邻的四个模式(M0+1，M0+2，M1-1，M1-2)，并将DC模式和Planar模式加入预测模式候选列表PML中；当M0与M1不相邻时，分别将{M0,M1}相邻间隔为1和2的模式、{M2}间隔为2的模式和Planar和DC两个非角度模式加入预测模式候选列表PML中；Step 2: When M0 is adjacent to M1, the optimal direction of prediction is near M0 and M1. At this time, three modes (M0-1, M0-2, M0-3) and M0 and M1 between M0 and M1 are calculated. The adjacent four modes (M0+1, M0+2, M1-1, M1-2) are added, and the DC mode and the Planar mode are added to the prediction mode candidate list PML; when M0 is not adjacent to M1 , respectively, the {M0, M1} mode with adjacent intervals of 1 and 2, the mode of {M2} interval 2, and the two non-angle modes of Planar and DC are added to the prediction mode candidate list PML;
步骤3，计算预测模式候选列表PML中的率失真代价，当PU块的块大小为4×4或8×8时，仅保留预测模式候选列表PML中率失真代价最小的两个模式进行率失真优化；当PU块的块大小为16×16、32×32或64×64时，选择预测模式候选列表PML中率失真代价最小的三个模式进行率失真优化； Step 3: Calculate the rate distortion cost in the prediction mode candidate list PML. When the block size of the PU block is 4×4 or 8×8, only the two modes with the lowest rate distortion cost in the prediction mode candidate list PML are retained. Optimization; when the block size of the PU block is 16×16, 32×32 or 64×64, three modes with the lowest rate distortion cost in the prediction mode candidate list PML are selected for rate distortion optimization;
步骤4，将最可能预测模式候选列表MPM中的模式无重复地添加入预测模式候选列表PML中；Step 4, adding the pattern in the most probable prediction mode candidate list MPM to the prediction mode candidate list PML without repetition;
步骤5，对步骤4得到的预测模式候选列表PML中的模式进行最后的率失真优化；将率失真代价最小的模式作为当前PU块的帧内预测的最优模式。Step 5: Perform last rate-distortion optimization on the mode in the prediction mode candidate list PML obtained in step 4; use the mode with the lowest rate-distortion cost as the optimal mode of intra prediction of the current PU block.
针对上述视频编码中帧内预测模式的快速决策方法，进一步地，步骤1所述主要预测方位集是HEVC帧内预测角度模式系数为{2,6,10,14,18,22,26,30,34}的方向模式集合，具体包括九种方位。For the fast decision method of the intra prediction mode in the above video coding, further, the main predicted azimuth set in step 1 is the HEVC intra prediction angle mode coefficient is {2, 6, 10, 14, 18, 22, 26, 30 , 34} the set of direction modes, including nine orientations.
针对上述视频编码中帧内预测模式的快速决策方法，进一步地，在主要预测方位集中，当两个预测方位的模式系数的差的绝对值为4时，所述两个预测方位为相邻。For the fast decision method of the intra prediction mode in the above video coding, further, in the main predicted azimuth set, when the absolute value of the difference of the mode coefficients of the two predicted azimuths is 4, the two predicted azimuths are adjacent.
针对上述视频编码中帧内预测模式的快速决策方法，进一步地，步骤1具体采用哈达码变换方法计算得到所述粗略率失真代价。For the fast decision method of the intra prediction mode in the above video coding, further, step 1 specifically calculates the coarse rate distortion cost by using a Hada code conversion method.
针对上述视频编码中帧内预测模式的快速决策方法，进一步地，步骤2中，当M0与M1相邻时，设定M0为较大模式，预测模式候选列表PML更新为{M0,M1,0,1M0-1,M0-2,M0-3,M0+1,M0+2,M1-1,M1-2}；当M0与M1不相邻时，预测模式候选列表PML更新为{M0,M1,0,1,M0-1,M0-2,M0+1,M0+2,M1-1,M1-2,M1+1,M1+2,M2,M2+2,M2-2}。For the fast decision method of the intra prediction mode in the above video coding, further, in step 2, when M0 is adjacent to M1, M0 is set to a larger mode, and the prediction mode candidate list PML is updated to {M0, M1, 0. , 1M0-1, M0-2, M0-3, M0+1, M0+2, M1-1, M1-2}; when M0 is not adjacent to M1, the prediction mode candidate list PML is updated to {M0, M1 , 0, 1, M0-1, M0-2, M0+1, M0+2, M1-1, M1-2, M1+1, M1+2, M2, M2+2, M2-2}.
针对上述视频编码中帧内预测模式的快速决策方法，进一步地，当加入预测模式候选列表PML的模式超出[2,34]范围时，通过循环的方式获取角度模式范围内的系数，再加入到预测模式候选列表PML中。For the fast decision method of the intra prediction mode in the above video coding, further, when the mode of adding the prediction mode candidate list PML exceeds the range of [2, 34], the coefficients in the range of the angle mode are acquired by loop, and then added to Prediction mode candidate list PML.
针对上述视频编码中帧内预测模式的快速决策方法，进一步地，步骤3具体采用哈达码变换方法计算得到所述率失真代价。For the fast decision method of the intra prediction mode in the above video coding, in step 3, the rate distortion cost is calculated by using the Hada code conversion method.
针对上述视频编码中帧内预测模式的快速决策方法，进一步地，步骤5所述进行最后的率失真优化包括如下步骤：For the fast decision method of the intra prediction mode in the above video coding, further, the final rate distortion optimization performed in step 5 includes the following steps:
51)将预测模式候选列表PML中的每一种预测角度模式作为当前PU的预测模式；51) using each of the prediction mode candidate list PML as the prediction mode of the current PU;
52)计算得到PML列表中每一种预测角度模式的率失真代价；52) calculating a rate distortion cost for each predicted angle mode in the PML list;
53)从步骤52)得到的率失真代价中选取率失真代价最小的模式，作为最优的帧内预测方向模式。 53) A mode in which the rate distortion cost is selected from the rate distortion cost obtained in step 52) as the optimal intra prediction direction mode.
与现有技术相比，本发明的有益效果是：Compared with the prior art, the beneficial effects of the present invention are:
本发明提供一种新的帧内预测快速模式决策的方法，充分利用编码块主要特征信息，降低帧内预测的运算复杂度，在保证编码性能的前提下提高视频编码的速度。本发明具有以下优点：The invention provides a new intra prediction fast mode decision method, fully utilizes the main feature information of the coding block, reduces the computational complexity of the intra prediction, and improves the video coding speed under the premise of ensuring the coding performance. The invention has the following advantages:
(一)本发明通过判断预测方位，更加充分地利用了角度模式之间的相关性，减少了进行哈达玛变换的模式数目，进而降低了预测过程的计算复杂度。(1) The present invention makes full use of the correlation between the angle modes by judging the predicted orientation, and reduces the number of modes for performing the Hadamard transform, thereby reducing the computational complexity of the prediction process.
(二)本发明构造一个预测模式候选列表，能够对方向模式做出更快速准确的判断，针对PU的尺寸，减少了进行率失真优化的模式个数。(2) The present invention constructs a prediction mode candidate list, which can make a faster and more accurate judgment on the direction mode, and reduces the number of modes for performing rate distortion optimization for the size of the PU.
附图说明DRAWINGS
图1是HEVC帧内预测的预测点坐标和参考点选取的示意图；1 is a schematic diagram of prediction point coordinates and reference point selection of HEVC intra prediction;
其中，每个方格代表一个像素点；Rx,y表示预测参考像素点；Px,y表示预测像素点；x和y表示像素点的位置，左上角为(0，0)原点。Wherein, each square represents one pixel; Rx, y represents a predicted reference pixel; Px, y represents a predicted pixel; x and y represent the position of the pixel, and the upper left corner is a (0, 0) origin.
图2是本发明提供的帧内预测模式快速决策方法的流程框图。2 is a flow chart of a method for quickly determining an intra prediction mode provided by the present invention.
图3是HEVC的方向预测模式和本发明的主要预测方位集的示意图；3 is a schematic diagram of a direction prediction mode of HEVC and a main predicted orientation set of the present invention;
其中，黑实线标示本发明的主要预测方位集，具体是系数为{2,6,10,14,18,22，26,30,34}的9个预测模式集合。Wherein, the black solid line indicates the main predicted orientation set of the present invention, specifically, the nine prediction mode sets whose coefficients are {2, 6, 10, 14, 18, 22, 26, 30, 34}.
图4是本发明实施例中当主要预测方位相邻时的候选预测模式的示意图；4 is a schematic diagram of candidate prediction modes when primary prediction azimuths are adjacent in an embodiment of the present invention;
其中，假设M0为较大模式，候选预测模式列表为{M0,M1,0,1M0-1,M0-2,M0-3,M0+1,M0+2,M1-1,M1-2}。Wherein, it is assumed that M0 is a larger mode, and the candidate prediction mode list is {M0, M1, 0, 1M0-1, M0-2, M0-3, M0+1, M0+2, M1-1, M1-2}.
图5是本发明实施例中当主要预测方位不相邻时的候选预测模式的示意图；5 is a schematic diagram of candidate prediction modes when primary prediction azimuths are not adjacent in an embodiment of the present invention;
其中，当M0与M1不相邻时，分别将{M0,M1}相邻间隔为1或2的模式、{M2}间隔为2的模式和Planar、DC两个非角度模式加入候选预测模式列表中{M0,M1,0,1,M0-1,M0-2,M0+1,M0+2,M1-1,M1-2,M1+1,M1+2,M2,M2+2,M2-2}。Wherein, when M0 and M1 are not adjacent, respectively, a mode in which {M0, M1} is adjacent to interval 1 or 2, a mode in which {M2} interval is 2, and two non-angle modes in Planar and DC are added to the candidate prediction mode list. {M0, M1, 0, 1, M0-1, M0-2, M0+1, M0+2, M1-1, M1-2, M1+1, M1+2, M2, M2+2, M2- 2}.
具体实施方式detailed description
下面结合附图，通过实施例进一步描述本发明，但不以任何方式限制本发明的范围。The invention is further described by the following examples in conjunction with the accompanying drawings, but not by way of limitation.
本发明提供一种视频编码中帧内预测模式的快速决策方法，该方法通过判断预测方位，利用角度模式的相关性和预测单元PU的尺寸构造预测模式候选列表，减少进行率失真优化的预测模式数目，从而降低帧内预测的运算复杂度，提高视频编码速度。 The invention provides a fast decision method for intra prediction mode in video coding, which determines a prediction mode by using a correlation of an angle mode and a size of a prediction unit PU to reduce a prediction mode for performing rate distortion optimization. The number, thereby reducing the computational complexity of intra prediction and improving the video encoding speed.
以下实施例描述了编码当前帧亮度块的帧内预测实现过程，构造与原亮度块差别最小的预测块。为了方便理解本发明实施例，首先在此介绍本发明实施例描述中引入的几个要素。The following embodiment describes an intra prediction implementation process that encodes a current frame luma block, constructing a prediction block that differs the least from the original luma block. In order to facilitate the understanding of the embodiments of the present invention, several elements introduced in the description of the embodiments of the present invention are first described herein.
编码单元(Coding Unit，CU)：CU可向下划分进行编码，帧内预测允许最大编码单元(Largest Coding Unit，LCU)大小64×64，最小编码单元(Smallest Coding Unit，SCU)大小为8×8。帧内预测时会对当前CU进行模式决策以及块划分，多次递归选择最佳的划分方法以及最优模式。Coding Unit (CU): The CU can be coded down. The intra prediction allows the maximum coding unit (LCU) size to be 64×64, and the minimum coding unit (SCU) size is 8×. 8. Intra prediction performs mode decision and block partitioning on the current CU, and multiple recursive selection of the best partitioning method and optimal mode.
预测单元(Prediction Unit，PU)：与当前CU的大小相同，当前CU为LCU时，可以继续划分为四个更小预测块。HEVC亮度分量帧内预测支持5种大小PU：4×4、8×8、16×16、32×32和64×64。Prediction Unit (PU): The same as the size of the current CU. When the current CU is an LCU, it can continue to be divided into four smaller prediction blocks. HEVC Luminance Component Intra Prediction supports five sizes of PUs: 4x4, 8x8, 16x16, 32x32, and 64x64.
预测模式(Prediction Mode)：HEVC帧内预测包含35种预测模式，是通过当前块的左侧相邻列与上侧相邻行的已编码像素值按照一定的规则对当前块做出的不同预测方式。其中包括33种角度预测模式和2种非角度预测模式，非角度预测包括直流(DC)预测和平面(Planar)预测。预测点的坐标以及参考点的选取如图1所示。角度预测模式的系数为2～34，DC模式和Planar模式的系数分别为0和1。Prediction Mode: HEVC intra prediction includes 35 prediction modes, which are different predictions of the current block according to certain rules by the coded pixel values of the left adjacent column and the upper adjacent row of the current block. the way. These include 33 angle prediction modes and 2 non-angle prediction modes. Non-angle prediction includes direct current (DC) prediction and plane (Planar) prediction. The coordinates of the predicted points and the selection of the reference points are shown in Fig. 1. The angle prediction mode has a coefficient of 2 to 34, and the DC mode and the Planar mode have coefficients of 0 and 1, respectively.
最可能预测模式(MPM)：HEVC标准利用相邻块之间的较强相关性建立的存储相邻上方及左侧PU预测模式的候选列表。Most Probable Prediction Mode (MPM): The HEVC standard uses a strong correlation between neighboring blocks to establish a candidate list of adjacent upper and left PU prediction modes.
率失真优化(Rate Distortion Optimization，RDO)：HEVC通过率失真优化过程，衡量不同预测模式的码率与失真性能，选择最优的预测模式。Rate Distortion Optimization (RDO): HEVC uses the rate-distortion optimization process to measure the rate and distortion performance of different prediction modes and select the optimal prediction mode.
帧内预测是通过模式决策获得不同划分的CU的最优模式以及预测模式，完成预测过程。本发明的流程如图2所示，帧内模式决策的核心方法包括以下几个步骤：Intra prediction is to obtain the optimal mode and prediction mode of different divided CUs by mode decision, and complete the prediction process. The flow of the present invention is shown in Figure 2. The core method of intra mode decision includes the following steps:
步骤1，对当前任意尺寸的PU块进行主要预测方位集中的模式系数为{2,6,10,14,18,22，26,30,34}的9种方位进行预测，为了提高编码速度，用哈达码变换代替离散余弦变换，计算哈达码变换之后的粗略率失真代价。代价函数如式1所示，其中，J为率失真代价，MODE为HEVC帧内预测所有模式的系数(0-35)，λ为拉格朗日常量，SATD(Sum of Absolute Transform Difference)是经过哈达码变换后该块预测模式与原始块像素值差值的绝对值之和。RMODE表示当前模式残差，BMODE表示该模式需要编码的比特数。Step 1: predicting 9 kinds of orientations of mode coefficients of the main predicted azimuth set of {2, 6, 10, 14, 18, 22, 26, 30, 34} for the current PU block of any size, in order to improve the coding speed, The Hada code transform is used instead of the discrete cosine transform to calculate the coarse rate distortion cost after the Hada code transform. The cost function is shown in Equation 1, where J is the rate distortion cost, MODE is the coefficient (0-35) of all modes of HEVC intra prediction, λ is the Lagrangian daily quantity, and SATD (Sum of Absolute Transform Difference) is The sum of the absolute values of the block prediction mode and the original block pixel value after the Hada code is transformed. R MODE indicates the current mode residual, and B MODE indicates the number of bits that the mode needs to encode.
J＝SATD(RMODE)+λ·BMODE    (式1)J=SATD(R MODE )+λ·B MODE (Equation 1)
根据公式1计算出的率失真代价对主要预测方位集中的模式排序，从中选取率失真代价 最优的前三个模式M0、M1和M2，其中M0、M1和M2的率失真代价关系为J(M0)<J(M1)<J(M2)。HEVC帧内预测全部模式及本发明提出的主要预测方位集如图3所示，主要预测方位集是系数为{2,6,10,14,18,22，26,30,34}的9个预测模式集合，在图3中以黑实线标示。The rate distortion cost calculated according to Equation 1 is ordered by the mode of the main predicted azimuth set, and the rate distortion cost is selected therefrom. The optimal first three modes M0, M1 and M2, wherein the rate distortion cost relationship of M0, M1 and M2 is J(M0)<J(M1)<J(M2). The HEVC intra prediction all mode and the main predicted azimuth set proposed by the present invention are as shown in FIG. 3, and the main predicted azimuth set is 9 coefficients having coefficients of {2, 6, 10, 14, 18, 22, 26, 30, 34}. The set of prediction modes is indicated by a solid black line in FIG.
步骤2，如果M0与M1是相邻的，即其差的绝对值为4，则表示预测最优方向在M0和M1附近，则进一步计算M0与M1之间的三个模式(M0-1,M0-2,M0-3)和M0与M1之外的相邻的四个模式(M0+1,M0+2，M1-1，M1-2)，如图4所示，并将DC模式和Planar模式加入预测模式候选列表PML中。假设M0为较大模式，列表更新为{M0,M1,0,1M0-1,M0-2,M0-3,M0+1,M0+2,M1-1,M1-2}。如果M0与M1不是相邻的，即其差的绝对值大于4，如图5所示，则分别将{M0,M1}相邻间隔为1、2的模式、{M2}间隔为2的模式和Planar、DC两个非角度模式加入候选列表中{M0,M1,0,1,M0-1,M0-2,M0+1,M0+2,M1-1,M1-2,M1+1,M1+2,M2,M2+2,M2-2}。另外，若加入候选列表的模式超出了[2,34]范围，则用循环的方式取角度模式范围内的系数。例如M0＝2，则M0-1取34，M0-2取33。Step 2: If M0 and M1 are adjacent, that is, the absolute value of the difference is 4, indicating that the predicted optimal direction is near M0 and M1, then the three modes between M0 and M1 are further calculated (M0-1, M0-2, M0-3) and four adjacent modes of M0 and M1 (M0+1, M0+2, M1-1, M1-2), as shown in Figure 4, and DC mode and The Planar mode is added to the prediction mode candidate list PML. Assuming M0 is a larger mode, the list is updated to {M0, M1, 0, 1M0-1, M0-2, M0-3, M0+1, M0+2, M1-1, M1-2}. If M0 and M1 are not adjacent, that is, the absolute value of the difference is greater than 4, as shown in FIG. 5, the mode in which the {M0, M1} adjacent interval is 1, 2, and the mode in which the {M2} interval is 2, respectively. And two non-angle modes of Planar and DC are added to the candidate list {M0, M1, 0, 1, M0-1, M0-2, M0+1, M0+2, M1-1, M1-2, M1+1, M1+2, M2, M2+2, M2-2}. In addition, if the mode of adding the candidate list is beyond the range of [2, 34], the coefficients in the range of the angle mode are taken in a cyclic manner. For example, if M0=2, then M0-1 takes 34 and M0-2 takes 33.
步骤3，根据公式1计算PML中的率失真代价。对于小于等于8×8的PU块，即块大小为4×4或8×8，则仅保留列表中率失真代价最小的2个模式，否则，即块大小为16×16、32×32或64×64时，则选择代价最小的3个进行率失真优化的选择。 Step 3. Calculate the rate distortion penalty in PML according to Equation 1. For PU blocks less than or equal to 8×8, that is, the block size is 4×4 or 8×8, only the two modes with the lowest rate distortion cost in the list are retained, otherwise, the block size is 16×16, 32×32 or At 64×64, the choice of the three least-rate-rate-optimized optimizations is selected.
步骤4，使用MPM，将其候选列表中的模式无重复地添加入PML列表。Step 4. Using the MPM, add the patterns in its candidate list to the PML list without duplication.
步骤5，对PML列表中的模式进行最后的率失真优化，代价最小的模式即为当前块的帧内预测的最优模式，具体包括：In step 5, the last rate-distortion optimization is performed on the mode in the PML list, and the mode with the least cost is the optimal mode of intra prediction of the current block, which specifically includes:
51)尝试预测模式候选列表PML中的每一种预测角度模式作为当前PU的预测模式；51) Trying to predict each of the prediction angle patterns in the mode candidate list PML as the prediction mode of the current PU;
此步骤包括进行变换、量化、重构的过程；This step includes the process of transforming, quantifying, and reconstructing;
52)通过式2计算PML列表中每一种预测角度模式的率失真代价：52) Calculate the rate distortion cost of each predicted angle mode in the PML list by Equation 2:
J＝D(Mode)+λ·B(Mode)    (式2)J=D(Mode)+λ·B(Mode) (Equation 2)
式2中，J为率失真代价，D(Mode)和B(Mode)分别表示采用不同角度模式时的失真和比特数；In Equation 2, J is the rate distortion cost, and D(Mode) and B(Mode) respectively represent the distortion and the number of bits when using different angle modes;
53)从步骤52)得到的率失真代价中选取率失真代价最小的模式作为最优的帧内预测方向模式。 53) The mode in which the rate distortion cost is the smallest among the rate distortion costs obtained in step 52) is taken as the optimal intra prediction direction mode.
至此，该实例的所有步骤完成。At this point, all steps of the instance are complete.
本发明具体实施是在HEVC参考软件HM16.0上实现并测试HEVC通测序列包括Class A、Class B、Class C、Class D和Class E，配置为全I帧，HEVC标准规定了52个量化步长，对应于52个量化参数(Quantization Parameter，QP)。根据通测条件，量化参数QP分别设置为22、27、32、37。测试结果如表1所示，Y表示亮度分量，U和V表示色度分量，其中亮度部分的平均BD-rate轻微增加0.65％，整体节省编码时间(Time Save,TS)为26.76％。The specific implementation of the present invention is to implement and test the HEVC general test sequence including Class A, Class B, Class C, Class D and Class E on the HEVC reference software HM16.0, configured as a full I frame, and the HEVC standard specifies 52 quantization steps. Long, corresponding to 52 quantization parameters (QP). According to the general condition, the quantization parameter QP is set to 22, 27, 32, 37, respectively. The test results are shown in Table 1. Y represents the luminance component, and U and V represent the chrominance components, wherein the average BD-rate of the luminance portion is slightly increased by 0.65%, and the overall time saving (TS) is 26.76%.
表1Table 1
Figure PCTCN2017085605-appb-000001
Figure PCTCN2017085605-appb-000001
需要注意的是，公布实施例的目的在于帮助进一步理解本发明，但是本领域的技术人员可以理解：在不脱离本发明及所附权利要求的精神和范围内，各种替换和修改都是可能的。因此，本发明不应局限于实施例所公开的内容，本发明要求保护的范围以权利要求书界定的范围为准。 It is to be noted that the embodiments are disclosed to facilitate a further understanding of the invention, but those skilled in the art can understand that various alternatives and modifications are possible without departing from the spirit and scope of the invention and the appended claims. of. Therefore, the invention should not be limited by the scope of the invention, and the scope of the invention is defined by the scope of the claims.

Claims (9)

  1. 一种视频编码中帧内预测模式的快速决策方法，其特征是，通过判断预测方位，利用角度模式的相关性和预测单元PU的尺寸构造预测模式候选列表，减少进行率失真优化的预测模式数目，从而降低帧内预测的运算复杂度，提高视频编码速度；包括如下步骤：A fast decision method for intra prediction mode in video coding, characterized in that by predicting a prediction azimuth, constructing a prediction mode candidate list by using a correlation of an angle mode and a size of a prediction unit PU, reducing the number of prediction modes for performing rate distortion optimization , thereby reducing the computational complexity of intra prediction and improving the video encoding speed; including the following steps:
    步骤1，对当前任意尺寸的PU块，对主要预测方位集中的方位进行预测，计算得到粗略率失真代价；根据率失真代价对主要预测方位集中的模式排序，从中选取率失真代价最优的前三个模式M0、M1和M2，所述M0、M1和M2的率失真代价满足J(M0)<J(M1)<J(M2)；所述M0、M1和M2的方位关系为相邻或不相邻；Step 1: predicting the azimuth of the main predicted azimuth set for the current PU block of any size, and calculating the coarse rate distortion cost; sorting the mode of the main predicted azimuth set according to the rate distortion cost, and selecting the best rate distortion cost before selecting Three modes M0, M1 and M2, the rate distortion cost of M0, M1 and M2 satisfies J(M0)<J(M1)<J(M2); the orientation relationship of M0, M1 and M2 is adjacent or Not adjacent
    步骤2，当M0与M1相邻时，预测最优方向在M0和M1附近，计算M0与M1之间的三个模式和M0与M1之外的相邻的四个模式，并将DC模式和Planar模式加入预测模式候选列表PML中；当M0与M1不相邻时，分别将{M0,M1}相邻间隔为1和2的模式、{M2}间隔为2的模式、Planar模式和DC模式加入预测模式候选列表PML中；Step 2: When M0 is adjacent to M1, the optimal direction is predicted to be near M0 and M1, three modes between M0 and M1, and four adjacent modes other than M0 and M1 are calculated, and the DC mode and Planar mode is added to the prediction mode candidate list PML; when M0 is not adjacent to M1, {M0, M1} adjacent mode is 1 and 2, {M2} interval 2 mode, Planar mode, and DC mode, respectively. Adding to the prediction mode candidate list PML;
    步骤3，计算预测模式候选列表PML中的率失真代价，当PU块的块大小为4×4或8×8时，仅保留预测模式候选列表PML中率失真代价最小的两个模式进行率失真优化；当PU块的块大小为16×16、32×32或64×64时，选择预测模式候选列表PML中率失真代价最小的三个模式进行率失真优化；Step 3: Calculate the rate distortion cost in the prediction mode candidate list PML. When the block size of the PU block is 4×4 or 8×8, only the two modes with the lowest rate distortion cost in the prediction mode candidate list PML are retained. Optimization; when the block size of the PU block is 16×16, 32×32 or 64×64, three modes with the lowest rate distortion cost in the prediction mode candidate list PML are selected for rate distortion optimization;
    步骤4，将最可能预测模式候选列表MPM中的模式无重复地添加入预测模式候选列表PML中；Step 4, adding the pattern in the most probable prediction mode candidate list MPM to the prediction mode candidate list PML without repetition;
    步骤5，对步骤4得到的预测模式候选列表PML中的模式进行最后的率失真优化；将率失真代价最小的模式作为当前PU块的帧内预测的最优模式。Step 5: Perform last rate-distortion optimization on the mode in the prediction mode candidate list PML obtained in step 4; use the mode with the lowest rate-distortion cost as the optimal mode of intra prediction of the current PU block.
  2. 如权利要求1所述视频编码中帧内预测模式的快速决策方法，其特征是，步骤1所述主要预测方位集是HEVC帧内预测角度模式系数为{2,6,10,14,18,22，26,30,34}的方向模式集合，具体包括九种方位。The method for quickly determining an intra prediction mode in video coding according to claim 1, wherein the main predicted azimuth set in step 1 is an HEVC intra prediction angle mode coefficient of {2, 6, 10, 14, 18, The set of direction patterns of 22, 26, 30, 34}, including nine orientations.
  3. 如权利要求1所述视频编码中帧内预测模式的快速决策方法，其特征是，步骤1具体采用哈达码变换方法计算得到所述粗略率失真代价。The method for quickly determining an intra prediction mode in video coding according to claim 1, wherein the step 1 specifically calculates the coarse rate distortion cost by using a Hadamard code conversion method.
  4. 如权利要求1所述视频编码中帧内预测模式的快速决策方法，其特征是，在主要预测方位集中，当两个预测方位的模式系数的差的绝对值为4时，所述两个预测方位为相邻。A method for quickly determining an intra prediction mode in video coding according to claim 1, wherein in the main predicted azimuth set, when the absolute value of the difference between the mode coefficients of the two predicted azimuths is 4, the two predictions The orientation is adjacent.
  5. 如权利要求1所述视频编码中帧内预测模式的快速决策方法，其特征是，步骤2中， 当M0与M1相邻时，设定M0为系数较大模式，预测模式候选列表PML更新为{M0,M1,0,1,M0-1,M0-2,M0-3,M0+1,M0+2,M1-1,M1-2}；当M0与M1不相邻时，预测模式候选列表PML更新为{M0,M1,0,1,M0-1,M0-2,M0+1,M0+2,M1-1,M1-2,M1+1,M1+2,M2,M2+2,M2-2}。A method for quickly determining an intra prediction mode in video coding according to claim 1, wherein in step 2, When M0 is adjacent to M1, M0 is set to a larger coefficient mode, and the prediction mode candidate list PML is updated to {M0, M1, 0, 1, M0-1, M0-2, M0-3, M0+1, M0. +2, M1-1, M1-2}; when M0 is not adjacent to M1, the prediction mode candidate list PML is updated to {M0, M1, 0, 1, M0-1, M0-2, M0+1, M0 +2, M1-1, M1-2, M1+1, M1+2, M2, M2+2, M2-2}.
  6. 如权利要求1所述视频编码中帧内预测模式的快速决策方法，其特征是，当加入预测模式候选列表PML的模式超出[2,34]范围时，通过循环的方式获取角度模式范围内的系数，再加入到预测模式候选列表PML中。A method for quickly determining an intra prediction mode in video coding according to claim 1, wherein when the mode of adding the prediction mode candidate list PML exceeds the range of [2, 34], the angle mode is acquired by cyclically The coefficients are then added to the prediction mode candidate list PML.
  7. 如权利要求1所述视频编码中帧内预测模式的快速决策方法，其特征是，步骤3具体采用哈达码变化方法计算得到率失真代价。The fast decision making method of the intra prediction mode in the video coding according to claim 1, wherein the step 3 specifically calculates the rate distortion cost by using the Hada code change method.
  8. 如权利要求1所述视频编码中帧内预测模式的快速决策方法，其特征是，步骤5所述进行最后的率失真优化包括如下步骤：A method for quickly determining an intra prediction mode in video coding according to claim 1, wherein the final rate distortion optimization performed in step 5 comprises the following steps:
    51)将预测模式候选列表PML中的每一种预测角度模式作为当前PU的预测模式；51) using each of the prediction mode candidate list PML as the prediction mode of the current PU;
    52)计算得到PML列表中每一种预测角度模式的率失真代价；52) calculating a rate distortion cost for each predicted angle mode in the PML list;
    53)从步骤52)得到的率失真代价中选取率失真代价最小的模式，作为最优的帧内预测方向模式。53) A mode in which the rate distortion cost is selected from the rate distortion cost obtained in step 52) as the optimal intra prediction direction mode.
  9. 如权利要求1所述视频编码中帧内预测模式的快速决策方法，其特征是，步骤52)具体通过式2计算PML列表中每一种预测角度模式的率失真代价：A method for quickly determining an intra prediction mode in a video encoding according to claim 1, wherein the step 52) calculates the rate distortion cost of each of the predicted angle modes in the PML list by using Equation 2:
    J＝D(Mode)+λ·B(Mode)    (式2)J=D(Mode)+λ·B(Mode) (Equation 2)
    式2中，J为率失真代价，D(Mode)和B(Mode)分别表示采用不同角度模式时的失真和比特数。 In Equation 2, J is the rate distortion cost, and D (Mode) and B (Mode) indicate the distortion and the number of bits when using different angle modes, respectively.
PCT/CN2017/085605 2016-07-12 2017-05-24 Rapid decision making method for intra-frame prediction mode in video coding WO2018010492A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610548047.0 2016-07-12
CN201610548047.0A CN106131547B (en) 2016-07-12 2016-07-12 The high-speed decision method of intra prediction mode in Video coding

Publications (1)

Publication Number Publication Date
WO2018010492A1 true WO2018010492A1 (en) 2018-01-18

Family

ID=57283015

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/085605 WO2018010492A1 (en) 2016-07-12 2017-05-24 Rapid decision making method for intra-frame prediction mode in video coding

Country Status (2)

Country Link
CN (1) CN106131547B (en)
WO (1) WO2018010492A1 (en)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109688414A (en) * 2018-12-19 2019-04-26 同济大学 A kind of reduction of VVC intraframe coding unit candidate modes and block, which divide, shifts to an earlier date terminating method
CN110248194A (en) * 2019-06-28 2019-09-17 广东中星微电子有限公司 Angle prediction technique and equipment in a kind of frame
CN111355956A (en) * 2020-03-09 2020-06-30 蔡晓刚 Rate distortion optimization fast decision making system and method based on deep learning in HEVC intra-frame coding
CN111800642A (en) * 2020-07-02 2020-10-20 中实燃气发展(西安)有限公司 HEVC intra-frame angle mode selection method, device and equipment and readable storage medium
CN112118444A (en) * 2019-06-20 2020-12-22 杭州海康威视数字技术股份有限公司 Encoding method and device
CN112235570A (en) * 2020-09-28 2021-01-15 中南大学 Fast prediction method based on precoding
CN112839224A (en) * 2019-11-22 2021-05-25 腾讯科技（深圳）有限公司 Prediction mode selection method and device, video coding equipment and storage medium
CN113596444A (en) * 2021-05-19 2021-11-02 西安邮电大学 360-degree video intra-frame prediction mode decision method
CN114097233A (en) * 2019-06-24 2022-02-25 现代自动车株式会社 Method and apparatus for intra prediction encoding of video data

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106131547B (en) * 2016-07-12 2018-07-03 北京大学深圳研究生院 The high-speed decision method of intra prediction mode in Video coding
CN106534870B (en) * 2016-12-19 2019-12-03 国网新疆电力公司电力科学研究院 A kind of rate-distortion optimal coding method based on RGB source video
WO2019083284A1 (en) 2017-10-24 2019-05-02 주식회사 윌러스표준기술연구소 Video signal processing method and apparatus
CN109246430B (en) * 2018-09-06 2022-02-01 北方工业大学 Virtual reality 360-degree video fast intra prediction and CU partition advance decision
CN111263144B (en) * 2018-11-30 2022-07-01 杭州海康威视数字技术股份有限公司 Motion information determination method and equipment
CN110139098B (en) * 2019-04-09 2023-01-06 中南大学 Decision tree-based intra-frame fast algorithm selection method for high-efficiency video encoder
CN110312127B (en) * 2019-06-25 2022-04-19 浙江大华技术股份有限公司 Method for constructing most probable prediction mode list, image coding method and processing device
CN110365982B (en) * 2019-07-31 2022-01-04 中南大学 Multi-transformation selection accelerating method for intra-frame coding in multipurpose coding
CN111918058B (en) * 2020-07-02 2022-10-28 北京大学深圳研究生院 Hardware-friendly intra prediction mode fast determination method, device and storage medium
CN113038132B (en) * 2021-03-18 2023-01-20 北京奇艺世纪科技有限公司 Prediction mode determination method and device for Coding Unit (CU)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104052994A (en) * 2014-04-14 2014-09-17 嘉兴职业技术学院 Method for HEVC intra-frame prediction mode quick decision based on grading self-adaption
US20150245068A1 (en) * 2014-02-21 2015-08-27 Industry-Academic Cooperation Foundation, Yonsei University Tsm rate-distortion optimizing method, encoding method and device using the same, and apparatus for processing picture
CN104883565A (en) * 2014-12-31 2015-09-02 乐视网信息技术（北京）股份有限公司 Decision-making method and device for intra-frame prediction mode of high efficiency video coding
KR101621358B1 (en) * 2015-04-16 2016-05-17 아주대학교 산학협력단 Hevc encoding device and method for deciding intra prediction mode for the same
CN106131547A (en) * 2016-07-12 2016-11-16 北京大学深圳研究生院 The high-speed decision method of intra prediction mode in Video coding

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150245068A1 (en) * 2014-02-21 2015-08-27 Industry-Academic Cooperation Foundation, Yonsei University Tsm rate-distortion optimizing method, encoding method and device using the same, and apparatus for processing picture
CN104052994A (en) * 2014-04-14 2014-09-17 嘉兴职业技术学院 Method for HEVC intra-frame prediction mode quick decision based on grading self-adaption
CN104883565A (en) * 2014-12-31 2015-09-02 乐视网信息技术（北京）股份有限公司 Decision-making method and device for intra-frame prediction mode of high efficiency video coding
KR101621358B1 (en) * 2015-04-16 2016-05-17 아주대학교 산학협력단 Hevc encoding device and method for deciding intra prediction mode for the same
CN106131547A (en) * 2016-07-12 2016-11-16 北京大学深圳研究生院 The high-speed decision method of intra prediction mode in Video coding

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109688414A (en) * 2018-12-19 2019-04-26 同济大学 A kind of reduction of VVC intraframe coding unit candidate modes and block, which divide, shifts to an earlier date terminating method
CN112118444A (en) * 2019-06-20 2020-12-22 杭州海康威视数字技术股份有限公司 Encoding method and device
CN112118444B (en) * 2019-06-20 2022-11-25 杭州海康威视数字技术股份有限公司 Encoding method and device
CN114097233A (en) * 2019-06-24 2022-02-25 现代自动车株式会社 Method and apparatus for intra prediction encoding of video data
CN110248194B (en) * 2019-06-28 2022-07-29 广东中星微电子有限公司 Intra-frame angle prediction method and device
CN110248194A (en) * 2019-06-28 2019-09-17 广东中星微电子有限公司 Angle prediction technique and equipment in a kind of frame
CN112839224B (en) * 2019-11-22 2023-10-10 腾讯科技（深圳）有限公司 Prediction mode selection method and device, video coding equipment and storage medium
CN112839224A (en) * 2019-11-22 2021-05-25 腾讯科技（深圳）有限公司 Prediction mode selection method and device, video coding equipment and storage medium
CN111355956A (en) * 2020-03-09 2020-06-30 蔡晓刚 Rate distortion optimization fast decision making system and method based on deep learning in HEVC intra-frame coding
CN111800642A (en) * 2020-07-02 2020-10-20 中实燃气发展(西安)有限公司 HEVC intra-frame angle mode selection method, device and equipment and readable storage medium
CN111800642B (en) * 2020-07-02 2023-05-26 中实燃气发展(西安)有限公司 HEVC intra-frame intra-angle mode selection method, device, equipment and readable storage medium
CN112235570A (en) * 2020-09-28 2021-01-15 中南大学 Fast prediction method based on precoding
CN112235570B (en) * 2020-09-28 2024-04-02 中南大学 Depth prediction method based on precoding and intra-frame direction prediction method
CN113596444B (en) * 2021-05-19 2022-07-19 西安邮电大学 360-degree video intra-frame prediction mode decision method
CN113596444A (en) * 2021-05-19 2021-11-02 西安邮电大学 360-degree video intra-frame prediction mode decision method

Also Published As

Publication number Publication date
CN106131547B (en) 2018-07-03
CN106131547A (en) 2016-11-16

Similar Documents

Publication Publication Date Title
WO2018010492A1 (en) Rapid decision making method for intra-frame prediction mode in video coding
RU2665237C1 (en) Moving image encoding device, moving image decoding device, moving image encoding method and moving image decoding method
JP6851429B2 (en) Image decoding device, image decoding method, image coding device and image coding method
RU2700396C1 (en) Image encoding device, an image decoding device, an image encoding method, an image decoding method and an image prediction device
TWI711296B (en) Image encoding device, image decoding device, image encoding method, image decoding method and storage media
RU2680194C1 (en) Image encoding device, image decoding device, image encoding method and image decoding method
CN105681808A (en) Rapid decision-making method for SCC interframe coding unit mode
CN104581152A (en) HEVC intra-frame prediction mode decision accelerating method
CN104702959A (en) Intra-frame prediction method and system of video coding
CN103442229A (en) Bit rate estimation method of SAO mode decision applied to encoder of HEVC standard
RU2562414C1 (en) Method for fast selection of spatial prediction mode in hevc coding system
CN113992911A (en) Intra-frame prediction mode determination method and device for panoramic video H264 coding
KR20120061035A (en) Method and apparatus for image intra prediction
JP2012023609A (en) Moving image encoding device, moving image decoding device, moving image encoding method and moving image decoding method
CN113674372A (en) Preprocessing algorithm for encoding and decoding remote sensing image
KR20130055316A (en) Apparatus and method for encoding/decoding of video for intra prediction

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17826831

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17826831

Country of ref document: EP

Kind code of ref document: A1