RCDN: Real-Centered Detection Network for Robust Face Forgery Identification

TL;DR

RCDN introduces a frequency spatial CNN framework that emphasizes real image consistency to improve robustness and cross-domain generalization in face forgery detection, outperforming existing methods.

Contribution

The paper proposes RCDN, a novel real-centered CNN architecture with a dual-branch design and real centered loss to enhance cross-domain robustness in face forgery detection.

Findings

Achieves state-of-the-art in-domain accuracy on DiFF dataset.

Significantly improves cross-domain generalization and reduces the generalization gap.

Maintains high stability ratio across different forgery types.

Abstract

Image forgery has become a critical threat with the rapid proliferation of AI-based generation tools, which make it increasingly easy to synthesize realistic but fraudulent facial content. Existing detection methods achieve near-perfect performance when training and testing are conducted within the same domain, yet their effectiveness deteriorates substantially in crossdomain scenarios. This limitation is problematic, as new forgery techniques continuously emerge and detectors must remain reliable against unseen manipulations. To address this challenge, we propose the Real-Centered Detection Network (RCDN), a frequency spatial convolutional neural networks(CNN) framework with an Xception backbone that anchors its representation space around authentic facial images. Instead of modeling the diverse and evolving patterns of forgeries, RCDN emphasizes the consistency of real images, leveraging a dual-branch architecture and a real centered loss design to enhance robustness under distribution shifts. Extensive experiments on the DiFF dataset, focusing on three representative forgery types (FE, I2I, T2I), demonstrate that RCDN achieves both state-of-the-art in-domain accuracy and significantly stronger cross-domain generalization. Notably, RCDN reduces the generalization gap compared to leading baselines and achieves the highest cross/in-domain stability ratio, highlighting its potential as a practical solution for defending against evolving and unseen image forgery techniques.