基于几何结构失真模型的图像质量评价研究

客观图像质量评价研究的目的是设计一种和视觉感知保持一致,且适用于各种失真模型的质量评价方法. 传统的结构相似度量质量评价方法忽视了自然图像本身的特点,不能很好地评判某些失真类型图像. 本文根据人眼视觉系统(Human visual system, HVS)在感知图像质量过程中的特点,探索自然图像的本征几何结构特征, 考虑像素点的方向失真、幅度失真和方差失真,提出了一种新型的基于图像几何结构失真模型的完全参考质量评价方法. 在标准数据库上的实验结果表明,本文方法适用于所有失真模型图像数据的质量评价, 计算复杂度相对较低,得到的图像客观评价结果和主观评价方法具有更好的一致性, 能够很好地反映人眼对图像质量的主观感受.

Image Quality Assessment Based on Geometric Structural Distortion Model

The goal of objective image quality assessment research is to automatically predict perceived image quality in a perceptually consistent manner, which should generally include all distortion types. It is found there is some deficiencies in traditional structural similarity index for image quality assessment by ignoring the geometric features of natural images, which fails to measure some particular distortion types. Inspired by the researches of quality prediction of human visual system (HVS) and the intrinsically geometric structural features of natural images, a novel geometric structural distortion model based full reference image quality assessment method is proposed in this paper. The distortion comparisons of direction, magnitude, and covariance are considered in our proposed method. The experimental results on standard image database show that our method is generally good for all distortion types and has relatively low computational complexity. It has a good consistency with the subjective assessment of human beings, thus, can be used to describe the visual perception of image quality.