基于视觉感兴趣度和视觉系统特性融合的图像质量评价方法

Image Quality Assessment Based on Fusion of Visual Interest and HVS Characteristics

在对视觉感兴趣度和人类视觉系统特性深入分析研究的基础上,提出了一种基于视觉感兴趣度和视觉系统特性相融合的图像质量评价新方法。该方法首先根据影响视觉感兴趣度的五个因子进行综合得到感兴趣度因子,结合结构相似性方法得到感兴趣度模型。然后利用小波分解后的子频带图像进行质量评价,对结果进行视觉敏感度加权,在此基础上得到评价算法。通过实验验证所提出的算法的正确性,其测试结果并与视觉主观感受保持很好的一致性。

Based on in-depth analysis of the visual interest and the characteristics of human visual system, a new method of image quality assessment combining visual interest and visual system is put forward. Firstly this method get interest factor by integrating five factors of interest, combine with the structural similarity method to get interested degree model. Then assess the quality of wavelet sub-band images and weight the results to evaluate the assessment algorithm. A large number of experiments verify the correctness of the algorithm proposed in this paper. The test results keep a good consistency with subjective feelings.