中文摘要：

针对视觉感知和结构化图像质量评价方法中均存在与视觉感受不吻合的现象，在建立亮度空间视觉相似分解方法的基础上，依据韦伯定律的基本概念并将其扩展到RGB三基色彩色空间，由此提出了一种基于彩色空间视觉相似分解的图像质量评价方法．该方法在保证误差图像和标准图像间的形状和色彩均完全相似的基础上，将误差图像分解为均值分量、交变相似分量和交变无关分量．由于这三个分量在视觉感受上截然不同，从而依据对视觉感知差异起到主导作用的交变无关分量来评价彩色图像质量．实验结果表明，该彩色空间图像质量评价模型与主观视觉感受吻合较好．

英文摘要：

To solve the inconsistency between human visual system （HVb）, structure similarlty （SSIM） and image quality assessment results, a method was proposed to expand the space similarity decomposition model into RGB color space according to Weber-Fechner law. Keeping the shape and color of error and reference images unchanged, the method decomposed the error image into background mean luminance component, alternation similarity component and alternation independence component. The visual perceptions of these three components are different from each other. The experiments show that the present image quality assessment results based on alternation independence components are more consistent with visual perception.