A robust anti-spoofing technique for face liveness detection with morphological operations

Face liveness detection is an advanced research topic nowadays due to its significant security applications in various fields and is of utmost paramountcy to ascertain the physical presence of person. The spoofing problem is a ferocious threat to security of the face recognition systems and it can be minimized by detecting the face liveness. In this paper, a robust anti-spoofing technique for face liveness detection with morphological operations has been proposed by considering eyeblink and mouth movements for procuring maximum reliability during face liveness detection. ZJU Eyeblink dataset, Print-Attack Replay dataset and in-house dataset created in our university have been used for experimental purpose. ZJU Eyeblink dataset has been used to capture eyeblink, Print - Attack Replay dataset has been used to detect photo and video attack based on eyeblink while both eyeblink and mouth movements have been detected simultaneously using in-house dataset. The experimental results show that the proposed anti-spoofing technique significantly improves the security of a face recognition system by detecting face liveness. The efficiency of the proposed technique has been prosperously evaluated by detecting photo and video spoofing attacks.