基于遗传算法的SVC非均等差错保护算法研究  

Unequal Error Protection with Genetic Algorithm for Scalable Video Coding

作　　者：安华[1] 张雄[1] 

机构地区：[1]太原科技大学电子信息工程学院,太原030024

出　　处：《太原科技大学学报》2015年第6期416-423,共8页Journal of Taiyuan University of Science and Technology

基　　金：太原科技大学研究生科技创新项目(20134011);太原科技大学同州电子科技创新基金项目(TZ201307)

摘　　要：由于网络传输环境的不稳定性,为了将可伸缩视频编码(Scalable video coding,SVC)的码流能够在异构网络上实现可靠高效的传输质量,采用了联合前向纠错技术的不均等差错保护(Unequal Error Protection).不均等差错保护根据数据重要度的不同对各层给予不同强度的保护量,然而如何合理分配各层所需的保护数据,使恢复的视频质量达到最优是研究中的难点。通过建立信道差错所引起的失真模型,利用遗传算法(Genetic Algorithms)得到了最优码率分配方案,通过实验与传统方法比较,该方法有效提高了解码端的重建视频质量,特别是在高丢包的网络环境下表现的更为出色。Abstract： Due to the instability of network transmission environment,in order to transmit the bit stream of SVC in heter- ogeneous network and realize the transmission quality, the scalable video coding combining unequal error protection scheme with forward error correction technology was adopted. The unequal error protection scheme gives each layer the protection of different intensity according to the importance degree of data. However, it is difficult to allocate protection date that is needed for each layer to restore the most optimal video quality. Through the establishment of distortion model and genetic algorithms, a new rate allocation scheme was proposed. Comparing with traditional approach in the unstable network environment, the results show that the proposed scheme effectively improves the reconstructed video quality of scalable video coding at the decoder, especially in high packet loss environment.

关 键 词：可伸缩视频编码(SVC) 不均等差错保护(UEP) 失真模型 遗传算法(GA) 优化 

分 类 号：TN919.72[电子电信—通信与信息系统]