摘要
图像质量评价已成为图像科学的重要分支。从真彩色图像的角度分析,利用人眼对于颜色信息较为敏感这一特性,将颜色失真加入到图像质量评价中。分析了适合人眼视觉感知的HSI颜色空间,提出了一种度量颜色相似度的方法,及分别度量颜色的色度和亮度的相似度,使用带有空间距离作参数的指数函数模拟颜色降质。在此基础上,提出了基于结构和颜色相似度的客观质量评价算法CS_SSIM,将颜色相似性融合到结构相似性中,并详细描述了算法实现过程。实验表明,在真彩色图像上,CS_SSIM具有更好的单调性和一致性。
The image quality assessment has become an important branch of image science. The real color images is analysed. Color distortion is applied to quality assessment because that our eyes are sensitive to colors is found. A method is composed to evaluate the color similarity on HSI space. Combine the SSIM and color similarity into the evaluation of image quality, we proposed a new method to evaluate image quality called CS_SSIM. The effectiveness of the algorithm by experiments is proved.
出处
《科学技术与工程》
北大核心
2014年第7期233-235,共3页
Science Technology and Engineering
关键词
颜色相似性
客观质量评价
结构相似性(SSIM)
color similarity objective quality assessment structure similarity image index （SSIM）
