Abstract
Liveness authentication has been a serious concern in various biometric types specially where sensitive body is hard to expose to sensors. In this paper, a detailed analysis has been carried out to test the liveness of iris template in comparison with vital signs of human body. A deep
structured learning approach has been applied with set of supervised, partially-supervised and unsupervised approaches on Chinese CASIA data set. The main purpose is to test the liveness of the iris template to avoid fraudulent cases at sensitive places where liveness authentication is mandatory.
There are basically two existing methods for liveness authentication i.e., static and dynamic. A fortified hybrid approach is proposed in this paper, which is rigorously tested with existing static and dynamic methods as benchmark. Twenty features were defined to authenticate the liveness
authentication using fortified hybrid approach. The detailed analysis of the deep learning method of liveness authentication phenomenon has been graphically visualized in results section as proof of concept. It is concluded that liveliness of iris templates with proposed twenty features can
significantly identify liveness of iris template in comparison of static or dynamic methods individually.