Features extraction in images on finger veins with hybrid curves

Abstract

This paper presents a research on the practical and advanced processing, which is supported by the geometric analysis for the features extraction in images of finger vein patterns in order to contribute to a system that minimizes the degree of phishing. This article develops a method of hybrid curves, which are supported on the tangent and secant curves, based on the Hough transform. A standardized system is formulated with a subset of data that are operated and extracted automatically from the vein image for index finger to save the characteristic information in terms of the angle. Proposal has some advantages, over other methods, as a low approach in terms of computational cost and a robust system test with the obtained percentage in the results.