A New Biornetorics Using Footprint

This paper proposes a new biometrics method based on footprint for personal classification. We introduce a normalization procedure to improve the accuracy rate of classification. The normalization procedure is composed of two main steps: 1) left and right footprints are separated automatically; 2) each footprint is normalized in direction and in position. In the experiment, the pressure distribution of the footprint was measured with a pressure-sensing mat. Ten volunteers contributed their footprints for testing the proposed method. The obtained results show that a) our normalization procedure makes classification stable, that b) Euclidean distance between the pressure distributions of input and registered footprints gives 82.64% accuracy rate, and that c) Euclidean distance used together with the subject's weight information and the geometric information of footprint achieves the highest rate 86.55%.