Robust Ensemble Morph Detection with Domain Generalization

Abstract

Although a substantial amount of studies is dedicated to morph detection, most of them fail to generalize for morph faces outside of their training paradigm. Moreover, recent morph detection methods are highly vulnerable to adversarial attacks. In this paper, we intend to learn a morph detection model with high generalization to a wide range of morphing attacks and high robustness against different adversarial attacks. To this aim, we develop an ensemble of convolutional neural networks (CNNs) and Transformer models to benefit from their capabilities simultaneously. To improve the robust accuracy of the ensemble model, we employ multi-perturbation adversarial training and generate adversarial examples with high transferability for several single models. Our exhaustive evaluations demonstrate that the proposed robust ensemble model generalizes to several morphing attacks and face datasets. In addition, we validate that our robust ensemble model gain better robustness against several adversarial attacks while outperforming the state-of-the-art studies.

A Quality Aware Sample-to-Sample Comparison for Face Recognition

“…Both the backbone and classifier will be trained end-to-end using a back-propagation algorithm. The Softmax training criterion can be formulated as follows [1,20]:…”

Joint Super-Resolution and Head Pose Estimation for Extreme Low-Resolution Faces

“…Among the image characteristics that can affect image quality, the resolution is one of the primary challenges in HPE [9]. It is known that Low-Resolution (LR) synthetic data can mimic real-world LR images and increase the resolution diversity of the training dataset [9], [20], [21]. Therefore, training HPE module with synthetic LR data may narrow the performance gap between the high and low-resolution HPE [9], [11].…”