PROJECT TITLE :

Fingerprint liveness detection using convolution Neural networks - 2016

ABSTRACT:

With the growing use of biometric authentication systems in the recent years, spoof fingerprint detection has become increasingly necessary. In this paper, we tend to use convolutional neural networks (CNNs) for fingerprint liveness detection. Our system is evaluated on the info sets employed in the liveness detection competition of the years 2009, 201one, and 2013, which contains nearly fifty 000 real and pretend fingerprints images. We have a tendency to compare four different models: two CNNs pretrained on natural images and fine-tuned with the fingerprint pictures, CNN with random weights, and a classical native binary pattern approach. We have a tendency to show that pretrained CNNs can yield the state-of-the-art results without having for design or hyperparameter selection. Knowledge set augmentation is used to increase the classifiers performance, not solely for deep architectures however also for shallow ones. We tend to additionally report sensible accuracy on terribly tiny training sets (400 samples) using these large pretrained networks. Our best model achieves an overall rate of 97.one% of properly classified samples-a relative improvement of sixteen% in check error in comparison with the simplest previously printed results. This model won the primary prize in the fingerprint liveness detection competition 2015 with an overall accuracy of 95.5%.