We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
A Robust Automatic Vascular and Non-Vascular based Retina Recognition system Using PSO.
- Authors
Kaur, Rupinder; Jindal, Sonika
- Abstract
Retinal recognition is the most accurate, stable and reliable identification method in biometric system. Unique vascular pattern of retinal fundus image is used to perform person's identification. In this paper, two different retinal recognition methods, vascular-based feature extraction and non-vascular based feature extraction with a PSO-based feature selection approach is presented. In our first method bifurcation or minutiae points are extracted from the retinal image and in second method structural features like luminance, contrast and structure are extracted instead of taking bifurcation points. After extraction of features, feature selection process has been applied to select a subset of features that efficiently represents original extracted features. To perform classification, the optimized feature vector is passed to the support vector machine (SVM) classifier to find the fitness values. Our proposed work is implemented on RIDB and DRIDB retinal image databases. Experimental results for both vascular and non-vascular methods are measured with 99% accuracy with false acceptance rate (FAR) and false rejection rate (FRR) parameters on both databases.
- Publication
International Journal of Advanced Research in Computer Science, 2017, Vol 8, Issue 5, p553
- ISSN
0976-5697
- Publication type
Academic Journal