Wireless Body Area Sensor Network Authentication Using Voronoi Diagram of Retinal Vascular Pattern

This paper focuses on the problem of human authentication in Body Area Sensor Network using retina. In this proposed method, Voronoi Diagram (VD), a well known technique in computational geometry, is generated from the topological structure of the bifurcation points, considered as vertices, obtained from the blood vessels found in the retina which can further be used in the process of identification and verification. Since the structure formed by bifurcation points is unique in every retina, hence the calculated VD is also unique and provides the foundation of developing the system of retina based identification. The approach presented in this paper rejects any non-similar retina instantly while maintaining excellent accuracy and performance. Another advantage of using this approach is that it does not require the localization of Optic Disc and the Fovea, which most of the existing algorithms have required, and, experimental results proved that VD is efficient in template matching and storage requirements. Additionally, our proposed algorithm is invariant against any geometric transformation (i.e. scaling, translation and rotation).