摘要
视频信号无论从空域还是时域都有很强的相关性,因此传输视频信号除了加纠错码外,在接收端的错误隐藏技术对提高重构图像质量也是十分有效的。该文提出了一个支持错误隐藏技术的MPEG-2(movingpictureexpertgroup)视频解码器的设计方案,用该方案进行的仿真实验证明其是可行的。同时基于图像信号的相关性、光滑性及一致性的假设,分别对MPEG-2的I-帧、P-帧和B-帧图像提出了相应的错误隐藏技术。此外通过分析直流系数的错误结果后,提出了直流系数的错误隐藏方法。该文算法简单,适于实时处理,峰值信噪比PSNR比一般传统方法平均高出2～3dB。
Video signals have high intra-frame and inter-frame correlations, so error concealment techniques in the decoder can be more effective other than error correction codes in video signal transmissions. Simulation of a video decoder design with error concealment have shown that the error concealment scheme is quite effective. Statistical correlations of the smoothness and consistency in images were used to develop error concealment techniques for I-frame, P-frame, and B-frame MPEG-2. Error concealment techniques for direct current (DC) coefficient errors gave higher quality reconstructed images. Test results show that the peak signal-to-noise ratio is 2～3 dB higher than that of some traditional methods and can be used with real-time processing.
出处
《清华大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2004年第7期921-924,共4页
Journal of Tsinghua University(Science and Technology)
基金
天津市自然科学基金资助项目(023600111)
关键词
图像信号处理
错误隐藏
直流系数
视频解码器
image signal processing
error concealment
direct current (DC) coefficient
video decoder
