摘要
研究了基于信源量化、信道错误和差错隐藏的视频传输端到端失真问题,提出了一种基于视频编码流结构的端到端失真度估算方法.采用帧间递归,以宏块为单位根据信道平均误比特率和视频编码信息在编码器端实时估算端到端失真.仿真结果表明,对于绝大部分视频测试序列,在不同信道误比特率下,该方法模型估算的平均相对偏差小于8%,平均绝对偏差控制在0.9 dB以内,准确度高,为基于率失真或联合功率率失真优化中的失真估算提供一个有力的工具.
End-to-end distortion estimation for video transmission was studied,in which the quantization,channel error and error concealment was considered.On the basis of coding bit stream,an end-to-end distortion estimation method is proposed,in which the end-to-end distortion of video sequences in MB level cursively was estimated according to channel bit error rate and video coding information obtained from the video encoder.Simulation results demonstrate that the average relative error of the estimation is less than 8 % and the average absolute error is within 0.9 dB for most sequences under different channel bit error rates.The method is applicable to relative optimization based on rate-distortion（R-D） or joint power and rate-distortion（P-R-D）.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2008年第7期26-29,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(60572063)
教育部博士学科点专项基金资助项目(20040487009)
关键词
视频传输
率失真优化
端到端失真
视频编码
video transmission
rate-distortion optimization
end-to-end distortion
video coding
