基于黎曼度量结构相似度的图像质量评价方法

Image Quality Assessment Based on Riemannian Metric Structural Similarity

图像最主要的是其结构特性,人类的视觉系统是从图像结构上获得图像信息,为了更好的对图像的质量进行评价,本文在SSM模型的基础上提出一种基于黎曼度量的结构相似度评价,把图像置于黎曼流形中进行讨论.实验结果表明本文提出的方法比SSM模型更有区分度,更符合人类视觉系统的特性.

An image is mainly defined by its structural characteristics. The image information in human visual system is based on the image structure. In order to better evaluate the quality of an image, a structural similarity assessment based on Riemann metric is proposed based on DDM model. The image is studied in Riemannian manifold. Experimental results show that the proposed method shows better discrimination than the SSM model and better adaptation to hu- man visual system.