Face Recognition Systems (FRS) are vulnerable to various attacks performed directly and indirectly. Among these attacks, face morphing attacks are highly potential in deceiving automatic FRS and human observers and indicate a severe security threat, especially in the border control scenario. This work presents a face morphing attack detection, especially in the On-The-Fly (OTF) Automatic Border Control (ABC) scenario. We present a novel Differential-MAD (D-MAD) algorithm based on the spherical interpolation and hierarchical fusion of deep features computed from six different pre-trained deep Convolutional Neural Networks (CNNs). Extensive experiments are carried out on the newly generated face morphing dataset (SCFace-Morph) based on the publicly available SCFace dataset by considering the real-life scenario of Automatic Border Control (ABC) gates. Experimental protocols are designed to benchmark the proposed and state-of-the-art (SOTA) D-MAD techniques for different camera resolutions and capture distances. Obtained results have indicated the superior performance of the proposed D-MAD method compared to the existing methods.
翻译:在这些攻击中,面部变形攻击极有可能欺骗自动FRS和人类观察者,并显示严重的安全威胁,特别是在边界控制情景中,这项工作展示了面部变形攻击探测,特别是在On-Fly(OTF)自动边境控制(ABC)情景中。我们根据从六种经过预先训练的深层神经网络(CNNs)中计算出的深层特征的球形内插和等级组合,提出了一种新的差异MAD(D-MAD)算法。通过考虑自动边境控制(ABC)门的真实生活情景,对新生成的面部变形数据集(SCFace-Morph)进行了广泛的实验。我们设计的实验协议旨在为不同摄像分辨率和距离确定拟议和状态DMAD(SOTA)技术的基准。获得的结果显示,与现有方法相比,拟议的DMAD方法的性能优于现有方法。