基于H.264的时空域联合误码掩盖算法

Spatio-temporal error concealing algorithm based on H.264

为了对视频传输中丢失的信息进行恢复，提高重建视频的质量，提出了一种基于H．264的时空域联合误码掩盖方法。该算法将丢失宏块分成16个4×4子块，通过分析子块的空间特性来确定其恢复的优先权，并依次恢复其运动矢量。仿真结果表明，对于不同运动程度序列和不同宏块丢失率，该算法性能均稳定地优于H．264标准的算法。特别是在20％的宏块丢失率下，恢复视频峰值信噪比（PSNR）相对H．264标准算法可提高1～2dB。

Video communications receivers conceal error to mask damaged images to improve the quality of reconstruction video. This paper presents a spatio-temporal error concealing algorithm based on H. 264. The algorithm divides a macro-block into sixteen sub-blocks, then analyzes the spatial features of each sub-block to determine its priority. The motion vectors in the sub-blocks are recovered in order depending on their priorities. The recovered motion vector is then used to get the sub-block in the reference frame. Test results show that the peak signal-to-noise ratio （PSNR） is improved by 1-2 dB over the traditional H. 264 algorithm when the macro-block lost rate per frame is 20%.