An Efficient Fuzzy Based Edge Estimation for Iris Localization and Pupil Detection in Human Eye for Automated Cataract Detection System

Abstract

Presently Digital Image Processing based automated diagnosis of medical images is very popular area of research. This technique cuts short the diagnosis time and produces results with high degree of accuracy with least variations in diagnostic opinion. For such applications, basic parameter mapping and evolving image processing techniques play a very crucial role. Automated Cataract detection through digital eye images requires circular pupil extraction from eye image. Conventional Hough transform fails to extract the circular pupil area due to presence of similar ring structure in Iris and in between region. This paper explains the distinct approach in iris localization and pupil extraction from eye images by extraction of edges using fuzzy logic approach resulting into effective edge estimation technique and later use of these inputs to circular Hough Transform for detection of both. The results of proposed approach has been compared on MMU1, UTIRIS, and IIT-Delhi eye biometric databases. Experimental results explain effective iris localization and pupil extraction with greater accuracy and lower computational time. Results show that the proposed algorithm can be effectively applied in automated cataract detection where circular regions: Iris Localization and subsequent pupil extraction are key to the detection methodology.