FINGER VEIN IDENTIFICATION BASED ON CORNER DETECTION

Abstract

authentication based on the physical features of the vein patterns in the finger. In this paper, computer vision algorithms (FAST and Harris) were used to extract features (corner points) from the finger vein image, while the matching of patterns was based on the differences between corners represented in the form of points using the Manhattan distance. The False Matching Rate (FMR) and False Non-Matching Rate (FNMR) were minimized to obtain the optimum threshold value. This threshold were used to make the final decision. The proposed system proved that use of two algorithms together reduces the error rate and aids to produce a reliable system of finger vein identification.