灰色复数矩阵SVD的无参考模糊图像质量评价

摘要

为了进一步突出重要的图像结构特征,采用复数矩阵表示图像,提出了基于灰色复数奇异值分解的无参考模糊图像质量评价方法.该方法首先将原始模糊图像经点扩散函数生成二次模糊图像,再采用复数矩阵的形式表示原始图像和二次模糊图像的结构特征,在此基础上,对原始模糊图像和二次模糊图像进行分块复数矩阵奇异值分解,获得区域相关度,采用灰色关联度评价模糊图像质量.在3个数据库上的实验结果表明,该方法评价结果合理,与主观评价具有较好的一致性.%In order to accentuate the complicated information that human eyes are sensitive to in an image,complex ma-trix was used to describe image structure,the method of no-reference blur image quality assessment based on grey weighted complex singular value decomposition was proposed.In this paper,based on point spread function,second blur image was generated by original blur image.The original fuzzy image and second blur image was respectively represented by the form of complex matrix,and then the original blur image and the second blur image were decomposed by the singular value decom-position in local region,the numerical quality assessment results were achieved by computing grey correlation of local singu-lar value.Experimental results on third open blur image databases show that the proposed method is more reasonable and has good agreement with the subjective score.