摘要
基于结构相似度的图像质量评价方法考虑图像的结构特点,因此相比传统的客观评价方法有改进。但这种方法不足的地方是:没有区别对待图像的边缘块、细节块和平滑块,而是将这些块的结构相似度SSIM简单地取平均而得出整幅图像的平均结构相似度MSSIM。提出一种基于分类图像块的加权结构相似度(WSSIM)的方法,并且在实验中取得比较满意的结果。
The image quality assessment method based on structural similarity makes some progress com- pared with the traditional methods, since it takes into account of the structure of the image. But it still has some drawbacks. For example, it doesn＇t divide an image into edge pieces, detail pieces, and smooth pieces, and just takes a mean value of every piece＇s SSIM to calculate MSSIM. Develops a method called Weighted Structural Similarity （WSSIM） which is based on the different pieces, and has a satisfying result in the experiment.
关键词
图像质量评价
结构相似度
加权结构相似度
Image Quality Assessment
Structural Similarity
Weighted Structural Similarity（WSSIM）
