Estimation of Fingerprint detection and Authentication Using the Neural Network Command Line

Abstract

Fingerprint authentication/identification is a good of biometric technique for identity recognition. It is good
and very important technique used in forensic research for investigation purpose. Here in my proposed work I have
implemented an algorithm for personal authentication of a person by his/her fingerprints using Neural Network and
compare it with SVM technique on the basis of CCR. Here I have considered the fingerprint of a person as an image and
after the fingerprints of many persons we have created a database for all the images. Whenever we have two check the
authenticity of a person then he/she has to scan his/her finger by a scanner. Then this input fingerprints will be matched
with the databases for matching for recognition. Because, now a fingerprint is an image, that’s why nprtool has been
used. The experiment result of this research work will achieve a good performance on these databases. After completed
all the required steps, the result shows improvements in MSE and PSNR and also perform cumulative cost results over
the SVM technique