Detection of Face Morphing Attacks Based on Patch-Level Features and Lightweight Networks
Min Long, Xuan Zhao, Le-Bing Zhang, Fei Peng · Security and Communication Networks · 2022
Aiming at the problem of face morphing attack detection under mobile and resource-constrained conditions, a face morphing detection method based on patch-level features and lightweight networks is proposed. It utilizes the combination of three blocks’ structures for learning. By outputting the probability of each bona fide or morphed face patches, the whole face features are integrated for recognition. Experimental results and analysis show this method can significantly improve the detection accuracy of face morphing attacks. Compared with the existing methods, this method has the characteristics of high computational efficiency and strong robustness. It has great application potential in enhancing the security of the face recognition system.