Face Morph Detection Using Deep Learning with Inception V3

Abstract

The hazards to identification and security presented by sophisticated face morphing techniques that can fool biometric systems are examined in this study. It offers a dependable technique for identifying changed face photos using deep learning, more especially the InceptionV3 model. The model outperformed conventional techniques in identifying actual and changed photos after being trained on a sizable dataset of real and morphing faces. It also obtained good accuracy, precision, recall, and F1-score. By lowering false positives and negatives, our study improves biometric security and lays the groundwork for future deep learning advancements for useful security applications.