EFFECTIVE TECHNIQUE FOR FINGERPRINT LIVENESS DETECTION USING PROBABILISTIC NEURAL NETWORKS
Author(s):
Mr. PRABURAM G, Mrs. NISHA PRIYA P
Keywords:
Abstract
Fingerprint Detection and recognition is the most Challenging and widely used Biometric technologies. Today’s modern world, apparently it is used in many real applications. The real images of human identification characteristics are spoofed by Putty, Play-doh, Fingerprint mold, etc. Here we obtain a Probabilistic Neural Networks (PNN) used to oversee training set to develop probability density functions intense a pattern layer to fingerprint liveliness detection. This is a model based the core concept and based on multivariate probability estimation.Yield state-of-the-art results for architecture or hyper parameter selection is not needed for pre-trained PNNs., Not only for extreme architectures but also for requiring ones used by Dataset Augmentation is to improve the performance. More advantaged accuracy on very small training sets using these large pre- exercise networks. Our best model achieves an overall rate of 95.1% of correctly exercised classified samples.
Article Details
Unique Paper ID: 146436
Publication Volume & Issue: Volume 4, Issue 12
Page(s): 773 - 777
Article Preview & Download
Share This Article
Join our RMS
Conference Alert
NCSEM 2024
National Conference on Sustainable Engineering and Management - 2024