基于ROI运动区域的分布式视频编码方法  

Distributed video coding method based on ROI motion-region

作　　者：李峰[1] 程晓钢[1] 

机构地区：[1]江苏大学计算机科学与通信工程学院,江苏镇江212013

出　　处：《计算机工程与设计》2012年第1期269-272,共4页Computer Engineering and Design

摘　　要：传统的分布式视频编码方法中,视频帧的解码通常采用内插或外推的方法,然而对于运动剧烈视频帧的解码效果却并不理想,会产生重影、鬼脸等现象。因此,提出了一种基于ROI运动区域的分布式视频编码解决方案,通过在编码端提取ROI辅助信息,然后在解码端将其与边信息做融合处理来得到更加准确的边信息,从而提高视频解码的质量。实验结果表明,与传统的分布式视频编码方法相比,基于ROI运动区域的分布式视频编码解决方案在性能上平均有3.1db的提高。In the traditional method of distributed video coding, the decoding of the video frames use the way of interpolation or extrapolation. However, the decoding effect of motion video frames is unsatisfied, which may along with ghosting, faees phenomenons, etc. Therefor, a ROI motion-region based distributed video coding solution is proposed, which extracts the ROI information at the eneoder, and then the side information at the decoder is merged to obtain more accurate side information and to improve the quality of the decoded video. The experimental result indicate that compared with the traditional distributed video coding method, the ROI motion-region based distributed video coding have an average of 3. ldb promotion.

关 键 词：无线多媒体传感器 分布式视频编码 边信息 感兴趣区域 运动区域 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]