Computer Engineering and Applications ›› 2013, Vol. 49 ›› Issue (8): 174-177.

Previous Articles     Next Articles

Singular value decomposition low-pass-filter for normalization of retinal image

YI Yao, ZHANG Dongbo, LI Xiong   

  1. College of Information Engineering, Xiangtan University, Xiangtan, Hunan 411105, China
  • Online:2013-04-15 Published:2013-04-15

奇异值分解低通滤波眼底图像归一化

易  瑶,张东波,李  雄   

  1. 湘潭大学 信息工程学院,湖南 湘潭 411105

Abstract: A novel Singular Value Decomposition(SVD) method to estimate the fundus images background is presented. The feasibility of background estimation by using SVD is clarified in theory, and a SVD low-pass-filter is designed to realize the estimation of background to normalize the fundus image. Based on the normalization of fundus images, the new method reduces the computation time of background estimation during traditional fundus image normalization, which is meaningful to the post-processing of fundus image.

Key words: background estimation, Singular Value Decomposition(SVD), non-uniform illumination and contrast, normalization

摘要: 针对视网膜眼底图像归一化中的背景估计问题,提出了基于奇异值分解(Singular Value Decomposition,SVD)背景估计新方法。从原理上阐述了SVD估计背景的可行性,并在此基础上设计SVD低通滤波器实现了眼底图像的背景估计,最终实现图像的归一化。新方法在完成图像归一化的基础上,克服了传统眼底图像归一化中背景估计环节计算速度慢的特点,这对视网膜眼底图像后续处理具有重要意义。

关键词: 背景估计, 奇异值分解(SVD), 非均匀照度和对比度, 归一化