Design and Development of New Algorithm for Person Identification Based on Iris Statistical Features and Retinal Blood Vessels Bifurcation Points

Abstract

Biometric are the trait, measured features used to tag and describe persons. Physical distinctiveness is interconnected to the structure of the body. Biometric recognition are, face recognition, fingerprint, DNA, palm print, iris, and retina. We describe a novel algorithm for the identification and measurement of iris statistical features and identify the bifurcation points of retinal blood vessels for person recognition, by applying DIP techniques. Iris algorithm is performed on CASIA database and local database collected from KVKR (Department of CS and IT, Dr. B.A.M.U, Aurangabad) research lab, overall 100 iris image databases. For localization and extraction of inner iris is done by using different image processing techniques. After extraction of inner iris, statistical features are calculated such as, area, diameter, length, thickness, and mean. Performance analysis of the algorithm is done by using ROC curve. This algorithm achieves sensitivity of 94.92% and specificity of 100%. Afterwards retinal blood vessels bifurcations points are extracted. Retinal image database is collected by Dr. Manoj Saswade (Ophthalmologist, Saswade Netra Rugnalaya, Aurangabad (MH)), overall 500 retinal image databases is collected. Then-after apply DIP techniques, like image enhancement and otsu’s method, minutia techniques etc. Retinal blood vessels bifurcations points achieves a TP rate of 98%, FP rate of 20%, and overall accuracy score of 0.9702.