期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2015
卷号：4
期号：8
页码：3363-3369
出版社：Shri Pannalal Research Institute of Technolgy
摘要：Accurate extraction of retinal blood vessels is very important task in the diagnosis of various retinopathies. Directional filtering is a simple yet effective method for vessel extraction. Number of direction templates used for filtering effects the results. More the number of directional templates used, more precise the results are. In this paper, we propose a novel extension to directional filters method .The proposed method is composed of the Local entropy based thresholding method to extract thin vessels from the retinal image and the eight directional filter templates are provides filtering. The purposed method is very simple, however it significantly reduce the false detection rate and detect fine vessels that were missed by the directional filters based method.
关键词：Retinal images; Local Entropy based Thresholding; ; Directional filters; DRIVE dataset
