应用主成分分析和KL变换的体育视频压缩算法  

Sports Video Compression Algorithm by Using Principal Component Analysis and KL Transform

作　　者：熊辉 王文雯[2] XIONG Hui;WANG Wenwen(College of Engineering and Technology, Yangtze University College of Arts and Sciences, Jingzhou 434020, China;Public Basic Course Department, Jingzhou Institute of Technology, Jingzhou 434020, China)

机构地区：[1]长江大学文理学院体育系,湖北荆州434020 [2]荆州职业技术学院公共基础课部,湖北荆州434020

出　　处：《微型电脑应用》2022年第1期126-129,共4页Microcomputer Applications

摘　　要：当前体育视频压缩算法存在压缩率低,丢失信息严重等缺陷,为了获得理想的体育视频压缩效果,提出了基于主成分分析和KL变换的体育视频压缩算法。首先分析当前体育视频压缩的研究现状,找到各种体育视频压缩算法的不足,然后采用主成分分析算法对体育视频关键帧投影,得到多个子空间,并采用KL变换对体育视频关键帧的子空间进行编码,使子空间的KL变换系数相关性最小,从而最大限度的实现体育视频压缩,最后与其它体育视频压缩算法进行仿真测试,测试结果表明,主成分分析和KL变换提升了体育视频压缩率,不仅能够去掉体育视频关键帧之间的冗余信息,而且体育视频压缩效率高,获得了比传统算法更好的体育视频压缩效果,具有更加广泛的实际应用价值。The current sport video compression algorithm has the defects of low compression rate and serious loss of information.In order to obtain ideal compression effect of sports video,a sport video compression algorithm based on principal component analysis and KL transform is proposed.Firstly,this paper analyzes the current research status of sports video compression,finds out the shortcomings of various sports video compression algorithms,then uses principal component analysis algorithm to project sports video key frame to obtain multiple subspaces,and uses KL transform to code the subspace of sports video key frame,so as to minimize the correlation of KL transform coefficient in subspace,and realize sports video compression to the maximum extent.Finally,simulation tests are carried out with other sports video compression algorithms.The test results show that the principal component analysis and KL transform can improve the compression rate of sports video,remove the redundant information between the key points of sports video,and the sports video compression efficiency is high,and the sports video compression effect is better than the traditional algorithm,which has more extensive practical application value.

关 键 词：体育视频 压缩算法 主成分分析 关键帧 冗余信息 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN911.73[自动化与计算机技术—计算机科学与技术]