基于亮度残差的EVC视频编码的快速CU划分算法  

Residual-Based Fast CU Partitioning Algorithm for EVC

作　　者：陈思佳 李跃 林文斌 CHEN Sijia;LI Yue;LIN Wenbin(School of Mathematics and Physics,University of South China,Hengyang,Hunan 421001,China;School of Computer Science,University of South China,Hengyang,Hunan 421001,China)

机构地区：[1]南华大学数理学院,湖南衡阳421001 [2]南华大学计算机学院,湖南衡阳421001

出　　处：《南华大学学报（自然科学版）》2023年第3期52-58,共7页Journal of University of South China：Science and Technology

基　　金：国家自然科学基金项目(62001209)。

摘　　要：基本视频编码标准(essential video coding,EVC)采用了新的四叉树与嵌套多叉树(quadtree plus multi-type tree,QTMT)分区结构,使视频编码的效率大大提高,其中QTMT包括四叉树分区、水平和竖直二叉树分区以及水平和竖直三叉树分区。然而,随着编码效率的提高,编码的复杂度也在增加。根据EVC编码单元(coding unit,CU)的亮度残差值设计了一种全新的CU快速划分算法,CU的亮度残差值是由原始亮度值和预测亮度值之间的差异得到的,利用残差值,建立了一个CU提前终止模型,有效地降低了编码的复杂性。大量的实验结果表明,本文提出的方法平均减少了31.56%的编码时间,而Bj?ntegaard delta rate的增加也是较小的。Essential video coding(EVC) uses a new quadtree with a nested multitype tree(QTMT) partition structure,resulting in much more efficient video encoding,including Quadro tree partitions,horizontal and vertical binary tree partitions,and horizontal and vertical trinomial tree partitions.However,as the coding efficiency increases,the coding complexity also increases.This study proposes an EVC coding unit(CU) size partitioning method based on CU residual values.The distortion of the CU is obtained from the difference between the original luminance pixels and the predicted pixels.Using the distortion,a CU partitioning early termination model is established,which effectively reduces the coding complexity.Experimental results show that the proposed method in this paper reduces the coding time by 31.56% on average,and the Bj?ntegaard delta rate increases little.

关 键 词：EVC CU大小 四叉树划分 二叉树划分 三叉树划分 

分 类 号：TN919.81[电子电信—通信与信息系统]