The GA_NN_FL Associated Model for Authenticating Fingerprints

Abstract

The problem of “fingerprint authentication” is described briefly as follows: given various patterns of the “original” fingerprint of certain subject A. We need to determine that an “observing” fingerprint image is a form of the “original” given fingerprint or not. If the judgment is true, the conclusion is the fingerprint of subject A; otherwise conclusion is not the fingerprint of A. In this paper, we propose a model combining three techniques in intelligent computing–genetic algorithms, neural network, and fuzzy logic–in solving the above problem. The proposed hybrid model includes two networks: the classification neural network and the fuzzy neural network for authenticating fingerprints. The experimental results obtained from the locking system which permit entering/exiting security places show the feasibility of the proposed method.