基于压缩域互补分割的视频多描述编码研究  

Multiple description video coding research based on complementary segmentation in compressed domain

作　　者：江玉珍[1] 朱映辉[1] 欧阳春娟[2] 

机构地区：[1]韩山师范学院数学与信息技术系,广东潮州521041 [2]井冈山大学信息科学与传媒学院,江西吉安343009

出　　处：《计算机工程与设计》2009年第14期3362-3364,共3页Computer Engineering and Design

基　　金：江西省教育厅2008科技基金项目(GJJ08417)

摘　　要：鉴于数字视频在空间域及时间域上的信息相关性,提出在视频压缩过程中实现多描述编码的方法。算法以H.263为参考标准,在图块及运动估计的DCT域上进行系数划分,通过对少量重要信号的重复分配来保证单个描述的接收质量。同时,系数域的对角分割又使各个描述间具有较强的信号互补性。实验结果表明,算法具有描述冗余率低、编/解码直接快速、压缩率高等良好的MDC特性,是保障视频传输业务可靠性和实时性的有效方法。Due to the information relativity of digital video in space domain and time domain, a new multiple description video coding method is presented integrating the video compression process. Based on H.263 criterion, the algorithm is realized by distributing the DCT domain coefficient of image blocks and motion estimation. With redistributing a few important coefficients, the receiving quality of single description is guaranteed. Meanwhile, the diagonal distribution of coefficient domain can improve their signal complementary. Experimental results show that the algorithm has good MDC characteristics such as low redundancy, rapid coding and decoding, and high compression ratio. It is an effective approach to ensure the security and real-time character in video communication.

关 键 词：多描述编码 视频压缩 H.263 运动估计 DCT域 

分 类 号：TN919[电子电信—通信与信息系统]