首先提出了一种与H.264标准兼容的多描述视频编码方法。通过空域滤波把原始的视频序列分解为多个具有低空间分辨率的子序列,然后采用H.264编码器对每个子序列进行高效的压缩编码,得到与H.264标准兼容的压缩码流.在接收端,解码器根据信道误码所造成的描述丢失或受损的具体情况,利用正确接收的描述对丢失的描述进行恢复,提高视频的重建质量.另外,为了抑制解码端的误码扩散,该方法对丢失描述解码器的参考帧进行刷新,以提高重建视频的主观质量.实验结果表明:与H.264单描述编码方法相比,在信道误码率高的情况下,该方法可以提供更好的传输鲁棒性.
In this paper, a multiple description video coding scheme algorithm compatible with H.264 standard has been proposed. First, original video sequence is spatially filtered into several sub-sequences with lower definition. Then every sub-sequence is encoded by H.264 encoder, generating the H. 264 standardcompatible compressed bit-streams. At the receiver, according to the various cases that the descriptions are corruptednecriptions decoder in by channel errors, different methods are used to recover lost description(s) by correctly received one(s) to increase the decoded video quality. Furthermore, to control error propagation at the receiver, we refresh the reference frame of decoder for the lost description(s), to improve the subjective quality of reconstructed video. Experimental results show that, compared with the H.264 single description coding method, our proposed multiple description coding method can provide better transmission robustness.