Robust extraction of blood vessels for retinal recognition

In this competitive era, biometric systems provide more reliable security than traditional methods like passwords etc. Biometric systems perform person´s authentication based on his physical traits. A number of biometric systems has been developed in the last few years such as fingerprints, hand and palm geometry, retina etc. Due to stability, uniqueness and non-replicable nature of vascular pattern, retinal recognition is the most stable biometric system. Retinal recognition performs person´s identification based on the unique vasculature of retina. Generally, it is a three-step process, which includes pre-processing, segmentation and matching. Segmentation is the fundamental step, which becomes crucial in the presence of different pathological signs like exudates, lesions. If they are not removed in segmentation, then they produce false positives, hence leads to misclassification. To address this problem, this paper presents an efficient segmentation algorithm which aims to remove pathological effects from the diseased retinal images and improve matching results by reducing false recognition rate. Experimental results demonstrate the efficiency of proposed system.