中文说明：结构化图像相似性计算函数，SSIM = Structural SIMilarity（结构相似性），这是一种用来评测图像质量的一种方法。由于人类视觉很容易从图像中抽取出结构信息,因此计算两幅图像结构信息的相似性就可以用来作为一种检测图像质量的好坏. 首先结构信息不应该受到照明的影响,因此在计算结构信息时需要去掉亮度信息,即需要减掉图像的均值;其次结构信息不应该受到图像对比度的影响,因此计算结构信息时需要归一化图像的方差;最后我们就可以对图像求取结构信息了,通常我们可以简单地计算一下这两幅处理后的图像的相关系数. 然而图像质量的好坏也受到亮度信息和对比度信息的制约,因此在计算图像质量好坏时,在考虑结构信息的同时也需要考虑这两者的影响.
English Description：
Structured image similarity calculation functions,SSIM = Structural SIMilarity (structural similarity), which is used as a method of evaluating image quality. Because human vision is easy to extract structured information from the image, and therefore calculate the similarity of image structural information can be used as a method of testing the image quality is good or bad.First structure information not should by lighting of effects, so in calculation structure information Shi needs removed brightness information, that needs reduction off image of are value; second structure information not should by image compared degrees of effects, so calculation structure information Shi needs owned a of image of variance; last we on can on image seeking take structure information has, usually we can simple to calculation about this two site processing Hou of image of