A Sanity Check for Multi-In-Domain Face Forgery Detection in the Real World

2025-12-04

AI Summary

In real-world multi-domain face forgery detection, existing methods exhibit poor generalization under the unknown-domain single-image discrimination task (MID-FFD): although achieving high intra-domain AUC, they suffer from low cross-domain accuracy (ACC), primarily due to domain discrepancies dominating the feature space and weakening genuine/forged discriminative signals. To address this, we propose DevDet—the first model-agnostic framework tailored for MID-FFD. First, we introduce the Face Forgery Developer (FFDev), which synthesizes controllable and diverse forged samples to explicitly sharpen the genuine/forged decision boundary. Second, we design a Dose-Adaptive Fine-Tuning (DAFT) strategy that dynamically suppresses domain shift while enhancing cross-domain robust discrimination. Extensive experiments demonstrate that DevDet significantly improves ACC across multiple unseen domains (average +5.2%), while preserving strong generalization—establishing a novel paradigm for zero-prior-domain forgery detection in practical scenarios.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Application Domains: Misinformation & Fake News
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Existing methods for deepfake detection aim to develop generalizable detectors. Although "generalizable" is the ultimate target once and for all, with limited training forgeries and domains, it appears idealistic to expect generalization that covers entirely unseen variations, especially given the diversity of real-world deepfakes. Therefore, introducing large-scale multi-domain data for training can be feasible and important for real-world applications. However, within such a multi-domain scenario, the differences between multiple domains, rather than the subtle real/fake distinctions, dominate the feature space. As a result, despite detectors being able to relatively separate real and fake within each domain (i.e., high AUC), they struggle with single-image real/fake judgments in domain-unspecified conditions (i.e., low ACC). In this paper, we first define a new research paradigm named Multi-In-Domain Face Forgery Detection (MID-FFD), which includes sufficient volumes of real-fake domains for training. Then, the detector should provide definitive real-fake judgments to the domain-unspecified inputs, which simulate the frame-by-frame independent detection scenario in the real world. Meanwhile, to address the domain-dominant issue, we propose a model-agnostic framework termed DevDet (Developer for Detector) to amplify real/fake differences and make them dominant in the feature space. DevDet consists of a Face Forgery Developer (FFDev) and a Dose-Adaptive detector Fine-Tuning strategy (DAFT). Experiments demonstrate our superiority in predicting real-fake under the MID-FFD scenario while maintaining original generalization ability to unseen data.

Problem

• Addresses domain dominance in multi-in-domain face forgery detection
• Proposes DevDet to amplify real-fake distinctions in feature space
• Enables definitive real-fake judgments for domain-unspecified inputs

Innovation

• Introduces Multi-In-Domain Face Forgery Detection (MID-FFD) paradigm
• Proposes model-agnostic DevDet framework to amplify real-fake differences
• Uses FFDev and DAFT to enhance detection in domain-unspecified conditions