一种新的视频编码二次率失真模型及其性能分析  

A New Quadratic Rate Distortion Model for Video Coding and Its Performance Analysis

作　　者：万帅[1] 常义林[1] 杨付正[1] 罗忠[2] 

机构地区：[1]西安电子科技大学综合业务网国家重点实验室,西安710071 [2]华为技术有限公司,深圳518057

出　　处：《电子与信息学报》2007年第5期1136-1139,共4页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(60172030);华为科技基金(YJCB2003017MU)资助课题

摘　　要：基于对视频编码比特率与内容复杂度之间关系的研究,该文提出了一种新的视频编码二次率失真模型。该模型保留泰勒级数中的常数项,并通过合理引入内容复杂度,提高了描述实际信源率失真特性的准确性。大量实验分析表明该模型具有良好的性能,可广泛应用于各种速率控制算法以及其他使用率失真模型的场合,以提高视频编码器的率失真性能。Based on the study of the relationship between the bit rate for video coding and the content complexity, this paper presents a new quadratic rate distortion model for video coding, which retains the constant term in Taylor series. Through introducing the content complexity term reasonably, this model provides an accurate description of the rate distortion characteristic of the practical source. The performance of the model is evaluated through analyzing large amount of experimental data. This model can be used in rate control schemes for video coding and other applications where rate distortion model is in demand, by which the rate distortion performance of video coding systems can be improved.

关 键 词：视频编码 二次率失真模型 方差分析 速率控制 

分 类 号：TN919.8[电子电信—通信与信息系统]