丢包环境中H.264视频编码的快速模式判决

发布时间：2018-07-20 12:06

【摘要】：目的视频编码中传统的快速模式判决算法通常基于对视频源特性的分析,但如果考虑到信道传输中的差错,编码模式的率失真特性就会随之改变,快速模式判决算法的性能也会随之下降,需要考虑丢包环境对快速模式判决算法进行优化。方法为了解决这一问题,首先分析了丢包环境下各种编码模式的端到端率失真特性,在此基础上提出了一个分层结构的快速模式判决算法。通过快速估计出丢包环境下skip和intra模式的编码率失真代价,进而将模式判决的路径划分为non-intra和non-skip。结果将本文算法与基于遍历算法的容错视频编码算法进行对比实验,本文算法平均可以降低50%左右的编码时间,同时几乎不会降低率失真性能。结论实验结果表明,在对丢包环境下端到端率失真代价进行估计的基础上,所提出的分层结构快速模式判决的算法,可以在保证解码端图像质量的同时,显著节省编码时间,满足实时视频通信中对低复杂度和鲁棒性的要求。
[Abstract]:The traditional fast mode decision algorithm in video coding is usually based on the analysis of the characteristics of the video source, but if the errors in the channel transmission are taken into account, the rate-distortion characteristic of the coding mode will change. The performance of the fast pattern decision algorithm will also decline, so it is necessary to optimize the fast pattern decision algorithm considering the packet loss environment. Methods in order to solve this problem, the end-to-end rate-distortion characteristics of various coding modes in packet loss environment are analyzed, and a hierarchical fast pattern decision algorithm is proposed. The coding rate-distortion cost of skip and intra modes in packet loss environment is estimated quickly, and the path of pattern decision is divided into non-intra and non-skipp. Results compared with the ergodic algorithm, the proposed algorithm can reduce the coding time by about 50% on average, and the performance of rate-distortion can hardly be reduced. Conclusion the experimental results show that on the basis of estimating the cost of end-to-end rate-distortion in packet loss environment, the proposed fast pattern decision algorithm with hierarchical structure can significantly save coding time while ensuring the image quality at the decoding end. It meets the requirements of low complexity and robustness in real time video communication.
【作者单位】：中国传媒大学信息工程学院;
【基金】：国家自然科学基金项目(61001177,61472389)
【分类号】：TN919.81

【参考文献】