摘要
为构建符合主观视觉感知的图像质量评价指标,文中提出了一种利用突出边缘结构相似的图像质量评价方法.通过计算每一个像素点位置上的高阶统计来局部化突出边缘.根据退化图像与参考图像的突出边缘强度相似性,结合亮度相似性和对比度相似性定义图像质量指标.数值实验结果表明,利用文中方法对TID2008数据库中的图像进行评价的平均运算时间为78.43s,评价结果与主观评价接近.
To establish an image quality assessment index in accordance with the subjective visual perception , a Structural Similarity(SSIM) image quality assessment method is presented based on the structure similarity of the salient edge .The higher order statistics (HOS) was calculated in every pixel to localize the salient edge .Then the image quality metric was defined based on the similarity of the salient edge strength between the distortion image and the reference image .The similarity of lightness and contrast were also incorporated .Numerical experiment shows that it at takes 78 .43 s on average to assess the images in TID2008 database by the proposed method .And the assessment result is close to the result obtained by the subjective method .
出处
《西安工业大学学报》
CAS
2016年第3期173-176,共4页
Journal of Xi’an Technological University
基金
陕西省教育厅专项科研计划项目(15JK1371)
陕西省自然科学基础研究计划项目(2014JM1018)