BENet: A Cross-domain Robust Network for Detecting Face Forgeries via Bias Expansion and Latent-space Attention

View on arXiv ← Back to list

Authors: Weihua Liu, Jianhua Qiu, Said Boumaraf, Chaochao lin, Pan liyuan, Lin Li, Mohammed Bennamoun, Naoufel Werghi

Published: 2024-12-10 11:41:55+00:00

AI Summary

BENet, a Cross-Domain Robust Bias Expansion Network, improves fake face detection by using autoencoders to amplify differences between real and fake face reconstructions, creating a reliable bias for detection. It also incorporates a Latent-Space Attention module to capture inconsistencies in fake faces at various scales and a cross-domain detector module to enhance accuracy on unseen data.

Abstract

In response to the growing threat of deepfake technology, we introduce BENet, a Cross-Domain Robust Bias Expansion Network. BENet enhances the detection of fake faces by addressing limitations in current detectors related to variations across different types of fake face generation techniques, where ``cross-domain refers to the diverse range of these deepfakes, each considered a separate domain. BENet's core feature is a bias expansion module based on autoencoders. This module maintains genuine facial features while enhancing differences in fake reconstructions, creating a reliable bias for detecting fake faces across various deepfake domains. We also introduce a Latent-Space Attention (LSA) module to capture inconsistencies related to fake faces at different scales, ensuring robust defense against advanced deepfake techniques. The enriched LSA feature maps are multiplied with the expanded bias to create a versatile feature space optimized for subtle forgeries detection. To improve its ability to detect fake faces from unknown sources, BENet integrates a cross-domain detector module that enhances recognition accuracy by verifying the facial domain during inference. We train our network end-to-end with a novel bias expansion loss, adopted for the first time, in face forgery detection. Extensive experiments covering both intra and cross-dataset demonstrate BENet's superiority over current state-of-the-art solutions.

Key findings

BENet outperforms state-of-the-art methods in both intra- and cross-dataset evaluations, achieving superior accuracy and AUC scores. Its robustness is demonstrated through its resistance to unseen perturbations and its ability to generalize across various deepfake generation techniques. Ablation studies confirm the contribution of each module to the overall performance.

Approach

BENet uses an autoencoder to reconstruct input faces, highlighting differences between real and fake reconstructions (bias expansion). A Latent-Space Attention module captures inconsistencies across scales, and a cross-domain detector improves generalization to unseen deepfakes. The network is trained end-to-end with a novel bias expansion loss.

Datasets

FaceForensics++ (FF++) (LQ, RAW, HQ), Celeb-DF, Diverse Fake Face Dataset (DFFD), DeepFake Detection Challenge Dataset (DFDC)

Model(s)

Autoencoder-based network with a bias expansion module, Latent-Space Attention (LSA) module, and a cross-domain detector module; uses Xception as a backbone network.

Author countries

China, Unknown, Unknown, China, China

← Previous