基于场景参考帧的预测帧刷新  

Predictive Frame Refreshment Based on Scene-Context Reference Frame

作　　者：艾育华[1] 叶梧[1] 冯穗力[1] 胡兵[1] 谢明[1] 

机构地区：[1]华南理工大学电子与信息学院,广东广州510640

出　　处：《华南理工大学学报（自然科学版）》2007年第5期10-14,共5页Journal of South China University of Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(69972015);广东省自然科学基金资助项目(31391)

摘　　要：为了阻止无线网络视频传输中由于丢包和误码而引发的差错扩散,提出了一种基于场景参考帧的预测帧刷新的抗差错策略.通过对视频序列的分割,对场景内需要同步刷新的视频帧进行基于场景参考帧的预测编码,获得相应预测帧,在编解码端同步场景标志和刷新标志,从而实现预测帧刷新.试验结果表明,与传统的I帧刷新策略相比,在需要高I帧刷新率的不可靠无线网络中,文中策略能够在保持相似重建视频质量的前提下,使帧比特率显著下降.In order to avoid the error propagation due to packet loss and bit error during the video transmission in wireless network, a novel error-resilient scheme with predictive frame refreshment is proposed based on scene-context reference frame. By segmenting the video sequences, the corresponding predictive frames locating at the original frames and needing to be intra-frame refreshed are predictably coded based on the scene-context reference frame within the same scene scope. The predictive frame refreshment is then implemented by means of the scene-flag and refresh-signal synchronization between the encoder and the decoder. As compared with the traditional intra-frame refreshment method, the proposed scheme can obviously reduce the frame bit rate with a similar reconstructed video quality when the video transmits over an unreliable wireless network that requires high intra-frame refreshment rate.

关 键 词：抗差错 预测帧刷新 场景参考帧 视频传输 无线网络 

分 类 号：TN919.81[电子电信—通信与信息系统]