Face recognition, which is security-critical, has been widely deployed in our daily life. However, traditional face recognition technologies in practice can be spoofed easily, for example, by using a simple printed ph...
详细信息
ISBN:
(纸本)9781479903108
Face recognition, which is security-critical, has been widely deployed in our daily life. However, traditional face recognition technologies in practice can be spoofed easily, for example, by using a simple printed photo. In this paper, we propose a novel face liveness detection approach to counter spoofing attacks by recovering sparse 3D facial structure. Given a face video or several images captured from more than two viewpoints, we detect facial landmarks and select key frames. then, the sparse 3D facial structure can be recovered from the selected key frames. Finally, an Support Vector Machine (SVM) classifier is trained to distinguish the genuine and fake faces. Compared withthe previous works, the proposed method has the following advantages. First, it gives perfect liveness detection results, which meets the security requirement of face biometric systems. Second, it is independent on cameras or systems, which works well on different devices. Experiments with genuine faces versus planar photo faces and warped photo faces demonstrate the superiority of the proposed method over the state-of-the-art liveness detection methods.
暂无评论