Biometric system based in the feature of hand palm

Abstract

In this paper, we are going to use the drawings that form the characteristics of the hand palm, as are fold (commonly called lines of the hand), cracks, scars, and even fold of the fingers to verify and identify the persons. We propose a biometric system based on the hand palm image. On this way, a database of 500 hands belonging to 50 users has been built (10 hands by user). It has been studied two different methods for obtaining the parameters of the hand, by measures of textures and by measures of phalanges. The classifier system is a neural network, with back-propagation training algorithm on the Multilayer Perceptron. The found results are over 99% in verification system, and over 98% in identification system.