视频码流转换是支持在异质网络环境中进行视频通信的关键技术之一。提出了一种充分利用人类视觉系统特性的MPEG视频码流码率转换方法。与原有方法相比,该方法在转换过程中巧妙地使用Foveation视觉模型,有效地提高转换效率,产生视觉质量更好的低码率码流,并减少了计算量。

Video transcoding is a key technology to support video communications over heterogeneous networks.In this paper,we propose a new bit-rate transcoding method for MPEG video streams,which can utilize human visual system characteristics adequately.Compared with the conventional method,this method uses skillfully foveation visual model to improve transcoding efficiency,and it can produce new lower bit-streams with better visual quality while achieving computation reduction.