Admissions Open
Admissions Open
As the uniqueness property of face, face recognition is widely used in area which needs security. But it has some spoof attack by the unauthorized person to gain access of the authorized person. These attacks are the big challenge for current time. As the deep learning approaches gives the promising results in image processing so, in this study, a liveness detection technique for face recognition is proposed which gives robust and accurate results for different percentage of training set of images by using multilayer deep network DBN. The proposed DeBNet is used for extracting deep features and classification of real and spoof faces. Photograph imposter NUAA data set of images is used in this study.