Retinal Verification Using Point Set Matching

In this paper, we propose a new retinal verification method based on point set matching. First, image processing techniques are applied to segment the retinal image's optic disc. The blood vessel map is generated inside the segmented optic disc region. We observe that optic disc region blood vessels are more stable and exhibit unique variation for a particular person. The edge map of the optic disc blood vessels is computed and used as the feature for similarity measurement. The partial Hausdorff measure is used to compute similarity measure between two edge based feature maps. Experimental results on a set of retinal images in the publicly available VARIA database show promising verification (FAR, FRR and EER) performance.