Hybrid Methods for Detecting Face Morphing Attacks

The face morphing process blends two or more facial images to produce a singular morphed facial image that shows the vulnerabilities of Face Recognition Systems (FRS). The widespread use of facial recognition algorithms, especially in Automatic Border Control (ABC) systems, has elicited concerns about potential attacks, as modified passports pose a significant risk to national security. This research presents a hybrid approach for feature extraction from facial images. The suggested approach involves three stages: The initial phase involves preprocessing the image through resizing and face identification, using the Viola-Jones algorithm to detect and locate the human face in the image, regardless of its size, context, or environment. In the second step, we extract features using three different techniques: Transfer learning using ResNet50, Histogram of Oriented Gradients (HOG), and Local Binary Pattern (LBP); we produce a one-dimensional feature vector that merges the outputs of each technique. The third phase includes the classification process utilizing the Deep Neural Network (DNN) classifier and the Support Vector Machine (SVM) as a secondary classifier. The AMSL dataset that contains real face and morphed face images has been used for training and testing the proposed approach. The DNN classifier achieved an average accuracy of 98.62%, surpassing the SVM, which achieved an accuracy of 97.39%. The results demonstrate a higher accuracy in identifying morphing attacks relative to previous studies.