Fingerprint Identification System Using Neural Networks

Abstract

The use of fingerprint in biometric identification has been the most widely used authentication system. The uniqueness of the fingerprint for every human provides us with all we need for faultless identification. However, during the fingerprint scanning process, the image generated by the scanner may be slightly different during each scan. This paper puts the implementation of Artificial Neural Networks to provide an efficient matching algorithm for fingerprint authentication. Using the Back-Propagation technique, the algorithm works to match twelve fingerprint parameters and relate them to a unique number provided for each authorized user. Upon matching, the algorithm returns the best match for the given fingerprint parameters.