首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于块分类的图像加权结构相似度
引用本文:杨春玲;何流.基于块分类的图像加权结构相似度[J].华南理工大学学报(自然科学版),2009,37(1).
作者姓名:杨春玲;何流
作者单位:华南理工大学电信学院
摘    要:图像质量评价方法用来评价图像质量或图像处理算法的优劣,在图像处理领域至关重要。Zhou Wang等人提出的结构相似度图像质量评价方法,具有计算简单、性能优越的特性。但其不足之一是没有区别对待图像的边缘块,细节块和平滑块,仅简单地取子块结构相似度(SSIM)的平均值而得到整幅图像的平均结构相似度MSSIM。基于此,本文提出了一种基于图像块分类的加权平均结构相似度(WSSIM)的图像质量评价算法,并进行了大量的仿真实验,实验结果证明,本文所提算法明显比MSSIM更加符合人眼视觉系统特性(HVS)。

关 键 词:图像质量评价  人眼视觉系统  结构相似度  块分类  加权平均结构相似度  
收稿时间:2007-11-28
修稿时间:2008-1-18

Weighted Structural Similarity Based on Block Classification for Image Quality Assessment
Liu He he.Weighted Structural Similarity Based on Block Classification for Image Quality Assessment[J].Journal of South China University of Technology(Natural Science Edition),2009,37(1).
Authors:Liu He he
Abstract:Image quality assessment, evaluating the quality of images and image processing algorithms, is very important in image processing field. The metric Structural Similarity proposed by Zhou Wang et al. performs well with low computational complexity. But the Mean SSIM (MSSIM) obtained by averaging all the sub-blocks’ SSIMs is used as the whole image quality, without distinguishing the edge areas、detailed areas and smooth areas of the image. In this paper, we develop a Weighted Structural Similarity (WSSIM) based on block classification. From a lot of simulation results, it is obvious that the proposed WSSIM is more consistent with HVS than MSSIM.
Keywords:image quality assessment  human visual system  structural similarity  classified block  weighted structural similarity
点击此处可从《华南理工大学学报(自然科学版)》浏览原始摘要信息
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号