Face Liveness Detection using Deep Learning and Support Vector Machine
Face recognition systems have pervaded our daily lives rapidly, from being present in our mobile phones to being used for authentication in sensitive areas. Despite its enthusiastic acceptance, there still remain some flaws of such systems. Face recognition systems are vulnerable to varying environmental factors and may be deceived by an attacker pretending to be a legitimate user. Face liveness detection systems that work to prevent such ill-usage are either too computationally complex and time consuming or expensive. The use of Convolutional Neural Network (CNN) helps give better results in less time. This paper proposes a method that utilizes the convolution neural network along with Support Vector Machine (SVM) classifier implemented using SVM loss function. This leads to better results than those produced when using only CNN with it’s built in classifier, SoftMax.