Efficient two stage approach to detect face liveness : Motion based and Deep learning based

Abstract

Face liveness detection is a big challenge for the researcher. Face recognition based security system suffer from spoofing attack, because of lacking of proper face liveness detection system. In this paper, we proposed a new approach to prevent spoofing attack with a two stage approach, one is motion based and another is deep learning based. The network is train on ROSE-Youtu Face Liveness Detection Database. The whole model is test on real time videos from webcam. This combine approach gives a better performance than other approaches in ROSE-Youtu Face Liveness Detection Database. Our proposed model gives an accuracy of 95.44% and error rate of 4.56% which is better than existing models on ROSE-Youtu Face Liveness Detection Database.