融合像素串匹配的AVS2全色度屏幕与混合内容视频编码算法  

Pixel String Matching for Full-Chroma Screen and Mixed Content Coding in AVS2

作　　者：赵利平[1,2] 周开伦[2] 郭靖[2] 蔡文婷[2] 林涛[2] 朱蓉[1] ZHAO Li-Ping;ZHOU Kai-Lun;GUO Jing;CAI Weng-Ting;LIN Tao;ZHU Rong(College of Mathematics,Physics and Information Engineering,Jiaxing University,Jiaxing,Zhejiang 314000;Institute of Very Large Scale Integration,Tongji University,Shanghai 200092)

机构地区：[1]嘉兴学院数理与信息工程学院,嘉兴314000 [2]同济大学超大规模集成电路研究所,上海200092

出　　处：《计算机学报》2018年第11期2482-2495,共14页Chinese Journal of Computers

基　　金：国家自然科学基金(61871289;61601200);浙江省自然科学基金(LY15F020039;LY18F020021;LQ18F020006;LY19F020015);上海市自然科学基金(18ZR1440600)资助~~

摘　　要：新一代主流云计算与信息处理模式及平台中的一个主要问题是客户端与云服务器端(或客户端)之间的数据传输瓶颈问题,屏幕与混合内容编码技术以带宽最省、安全性高等优势,成为解决该问题的核心技术之一.该文在我国第二代数字音视频编解码技术标准(AVS2)的屏幕与混合内容编码(SMCC)扩展版中,提出了一种融合像素串匹配的全色度屏幕与混合内容编码技术,该技术已经被AVS2-SMCC扩展版工作草案稿采纳.针对目前AVS2不支持全色度4∶4∶4格式的特点,作者设计与实现了一种融合4∶4∶4格式的像素串匹配编码算法和4∶2∶0格式的帧内、帧间预测编码算法的全色度屏幕混合内容编解码框架与算法.在此框架下,首先提出了一种4∶2∶0格式的帧内、帧间编码算法与4∶4∶4格式的串匹配编码算法的率失真代价值计算方式;然后给出了一种改进的像素串匹配算法及其关键技术,包括最佳参考串确定算法、基于N元组的快速哈希表搜索算法、基于分段截断二进制码的长度参数编码方案和基于N元组的位移参数编码方案.实验结果表明,对于AVS2-SMCC通用测试序列中的移动的文字和图形类别,提出的算法与AVS2-P2基准算法集成了色度上、下采样后支持YUV4∶4∶4格式的编解码器相比,在编码复杂度增加较少的情况下,对于有损全帧内(AI)配置Y分量、U分量、V分量平均BD-rate降低率分别可达56.1%,97.2%,99.2%;对于有损低延迟P(LDP)配置Y分量、U分量、V分量平均BD-rate降低率分别可达41.8%,70.9%,72.9%,能大幅提高屏幕与混合内容的编码效率.With the development of the new generation of cloud computing and information processing mode and platform,Screen and Mixed Content Coding(SMCC or SCC)becomes a key technology with the lowest bandwidth and high security to solve the data transmission bottleneck problem from the cloud to the client or the client to the client.The screen contents with their diversity and comprehensiveness show significant differences from the traditional video contents.Therefore,how to build efficient screen content coding technology has become a challenging issue in video coding area.In recent years,SCC has attracted increasing attention of researchers from both academia and industry.In response to the demand on SCC,ISO/IEC MPEG and ITU-T VCEG are jointly developing an SCC version of high efficiency video coding(HEVC)standard.Although HEVC SCC has made significant progress in the past few years,continuous improvement of performance is still indispensable and very challenging.Two of the major tools adopted in HEVC SCC draft up to now are Intra Block Copy(IBC)coding and palette coding(PLT).IBC is efficient to code repeated identical patterns of a few fixed sizes with rectangle or square shapes in a picture,but is not flexible enough to code repeated identical patterns of varying sizes from a few pixels to a few thousands of pixels with a variety of shapes.PLT can code repeated identical patterns inside a CU using two types of intra-CU pixel-index string matching,but it cannot exploit non-local repeated identical patterns.On the other hands,pixel string matching technique has a generic,flexible and uniform coding system architecture to exploit both local and non-local repeated identical patterns with a variety of sizes and/or shapes in a wide range of generic screen content.A pixel string matching technique is proposed in AVS2(the second generation of the Audio Video Coding Standard)SMCC(AVS2-SMCC)extension to improve coding efficiency of screen and mixed content.The technique has been adopted into AVS2-SMCC extension working draft.I

关 键 词：视频编码 数字音视频编解码技术标准 屏幕混合内容 串匹配 哈希搜索 熵编码 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]