Affiliations: [a] School of Electrical Engineering and Intelligentization, Dongguan University of Technology, Dongguan, Guangdong, China | [b] School of Computer Science and Technology, Dongguan University of Technology, Dongguan, Guangdong, China | [c] School of Information Engineering Guangdong University of Technology, Guangdong, Guangdong, China | [d] Shanghai Dafeng Technology Co., Ltd, Shanghai, China
Correspondence:
[*]
Corresponding author: Xuefang Chen, School of Computer Science and Technology, Dongguan University of Technology, Dongguan, Guangdong 523808, China. E-mail: yuli_vip@yeah.net.
Abstract: The crossing compression of retinal artery and vein is closely related to retinal vein occlusion, so detecting the contraction angle of the crossed vein blood vessel can assist to diagnose the retinal vein occlusion diseases. Through pretreating methods such as filtering, enhancement and edge extraction, the binary edge images can be extracted. The candidate feature points can be obtained by the corner point detection method based on chord-to-point distance accumulation (CPDA). The self-adaptive rectangular filter is used to screen out the crossing point of candidate angle, so as to fit the edge curves and figure out the contraction degree of vein. The experimental results show that this algorithm can better detect the contraction degree of crossed vein blood vessel, with an average error remaining at ± 1∘ under different resolutions.
Keywords: Retina, chord-to-point distance accumulation (CPDA), crossing compression angle of artery and vein
