【摘要】 在H.264/AVC标准中,当出现数据包丢失或缺损时,在解码端进行误码掩盖是一种经济而有效的保证视频质量的方法,因而广受关注。常见的做法是利用正确接收的时域或空域信息对误码区域进行掩盖。然而,对于误码率很高的视频,目前尚无令人满意的技术。该文中提出一种基于无迹Kalman滤波,并辅以改进的双线性插值的新型误码掩盖架构。对若干典型视频的高误码掩盖实验表明,其峰值信噪比高于边界匹配算法和传统双线性插值算法达4dB。更多还原

【Abstract】 H.264/AVC error concealment at the decoder side is an economical approach for ensuring acceptable,stable video quality in case of packet erasure or loss;thus,it has attracted much interest.Avialable spatial or temporal information in various zones is often used to conceal errors.However,there are few methods for heavily corrupted videos.This paper presents a framework based on the unscented Kalman filter accompanied by a modified bilinear motion field interpolation method.Tests on several typical sequences with high macroblock loss rates reveal its superiority in PSNR over the boundary matching algorithm and the conventional MFI algorithm by up to about 4dB.更多还原