Abstract: By analyzing the end-to-end distortion, the influence of motion vector selection on error propagation is studied. Allowing for that traditional motion estimation algorithm only takes into account the prediction error within source encoding framework, which is no longer accurate in lossy environment, an end-to-end prediction error calculation method is proposed. This prediction error is used in the rate-distortion optimized selection of motion vectors. Simulations have been done based on H.264 standard. The results show that for low-bit-rate video transmission over Internet, the robustness of video stream and the decoded video quality are improved by appling the proposed algorithm.

Key words: rate-distortion optimization, motion estimation, low-bit-rate, Internet

摘要： 通过对端到端失真的成因进行分析，研究运动矢量的选择对误码扩散大小的影响。针对传统运动估计算法只考虑信源端的情况，提出一种基于端到端预测误差的计算方法并据此对运动矢量进行率失真优化选择。基于H.264编码平台的仿真试验表明，采用所提出的率失真优化运动估计算法可以进一步提高在Internet误码条件下的低码率视频流鲁棒性及解码质量。

关键词: 率失真优化, 运动估计, 低码率, 互联网

CLC Number: 

• TP301.6