对视觉敏感区域进行加权的视频质量评价方法

Criterion of video quality assessment based on property of HVS

最近Guan-Hao Chen在Zhou Wang提出的MSSIM方法基础上,提出了一种基于边缘结构失真的测量方法(ESSIM),但是这种方法不能够很好的直接应用到视频领域。就人类视觉系统(HVS)做了一定的研究,提出了一种考虑人眼视觉特性,考虑结构信息和对运动场景分类的视频质量评价方法(HCSSIM)。实验结果表明,在压缩损失、模糊和部分椒盐噪声损失上,其评价结果要好于现有的几类通用的评价方法。

Zhou Wang proposed the Structural Similarity theory（MSSIM）.Based on it,Guan-Hao Chen proposed a marginal texture distorts theory.But the theory isn＇t applied directly on the video field.In this dissertation,based on the HVS,the structural distortion measurement and motion information,this paper proposes a theory-HCSSIM.From lots of simulation results,HCSSIM outperforms MSSIM and the weighting SSIM based on variance proposed by Wang Zhou et al,in the compress distortion,smooth distortion and partial salt-and-pepper noise.