基于运动估计的视频质量评价  

Video Quality Assessment Based on Motion Estimation

作　　者：姚杰[1] 谭建明[1] 唐超[1] 陈婧[1] 

机构地区：[1]重庆通信学院,重庆400035

出　　处：《重庆理工大学学报（自然科学）》2012年第5期74-78,共5页Journal of Chongqing University of Technology：Natural Science

基　　金：国家创新基金资助项目(11c26215115768)

摘　　要：目前大多数视频质量评价算法是将所有帧的图像质量平均值作为整个视频的质量,但视频中的物体具有视觉运动特性,而简单地取所有帧的质量平均值无法准确反映出这种特性。针对该情况,在基于结构相似性的图像质量评价的基础上提出了一种基于运动估计的帧加权方式,通过运动向量对当前帧的运动情况进行估计,并赋予相应的权值,从而将图像质量评价扩展为视频质量评价。实验证明:本算法与目前流行的几个视频质量评价算法相比具有较高的评价准确性。Most current video quality assessment algorithms always consider the average quality scores of all frames in the video as the quality of the entire video. However, objects in video have visual too- tion features, and they can not be reflected accurately when simply get the average of all the frame quality scores. In this view, a frame-weighting method based on motion estimation is proposed, which estimates the motion condition of the faced frame through the motion vector, weights it appropriately, and extends the image quality assessment to video quality assessment. The experiments show that the proposed algorithm is more accurate than several modern popular algorithms

关 键 词：图像质量评价 视频质量评价 运动向量 结构相似性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]