基于最优色空间和视觉掩蔽的彩色图像评价算法

目的针对当前色差算法在评价彩色图像时未考虑图像中像素之间颜色在视觉上的空间效应,提出基于最优色空间和视觉掩蔽效应的彩色图像质量评价算法。方法通过分析色空间通道间的相关性,选取最优正交、对立空间作为评价的工作色空间,在此基础上,利用色空间各颜色通道的掩蔽函数,去除图像颜色与颜色之间在视觉上的空间关联性,最后构建图像颜色差别公式,以评价彩色图像质量。结果在验证实验中,通过利用Pearson相关系数、Spearman等级相关系数以及Kendall等级相关系数,分析各算法评价与图像主观评价之间的关系发现,该算法评价与主观评价的Pearson相关系数、Spearman等级相关系数和Kendall等级相关系数分别可达到0.3948,0.5840和0.4814,且分别大于现有其他色差算法评价与主观评价的相关系数。结论该算法评价结果与人眼视觉主观评价相对一致。

Color Quality Assessment Algorithm Based on Optimal Color Space and Visual Masking

Objective The spatial visual effects of image colors among different pixels are not considered when using the typical color difference formula to assess the quality of color images. Targeting at this problem, a color image quality assessment algorithm based on optimal color space and visual masking effects was presented. Methods In this algorithm, an orthogonal and opponent color space, determined by analyzing correlation coefficient matrices of color spaces, was selected as its optimal working space. Visual masking functions for each channel of the color space were then used to de-correlate the spatial correlation caused by the interplay among colors in human vision. Finally, an image color difference formulae was constructed to assess the quality of color images. Results In the testing experiment, Pearson Correlation Coefficient (PCC), Spearman Rank-Order Correlation Coefficient (SROCC) and Kendall Rank-Order Correlation Coefficient (KROCC) were used to figure out the relationship between the objective assessment and the subjective visual assessment. In relation to the subjective assessment values, the PCC, SROCC and KROCC values of the objective assessment algorithms were 0.3948, 0.5840 and 0.4814, respectively, which were higher than those for other assessment algorithms. Conclusion The final results of the experiment showed that the proposed algorithm was essentially consistent with the subjective visual assessment.