Learning Real Facial Concepts for Independent Deepfake Detection

TL;DR

This paper introduces RealID, a novel deepfake detection approach that improves generalization by learning a comprehensive concept of real faces and making independent class decisions, outperforming existing methods across multiple datasets.

Contribution

RealID's innovative modules capture real face concepts and redefine classification, significantly enhancing deepfake detection generalization and accuracy.

Findings

Achieves 1.74% higher average accuracy than state-of-the-art methods.

Effectively reduces false positives on unseen datasets.

Demonstrates robustness across five widely used datasets.

Abstract

Deepfake detection models often struggle with generalization to unseen datasets, manifesting as misclassifying real instances as fake in target domains. This is primarily due to an overreliance on forgery artifacts and a limited understanding of real faces. To address this challenge, we propose a novel approach RealID to enhance generalization by learning a comprehensive concept of real faces while assessing the probabilities of belonging to the real and fake classes independently. RealID comprises two key modules: the Real Concept Capture Module (RealC2) and the Independent Dual-Decision Classifier (IDC). With the assistance of a MultiReal Memory, RealC2 maintains various prototypes for real faces, allowing the model to capture a comprehensive concept of real class. Meanwhile, IDC redefines the classification strategy by making independent decisions based on the concept of the real class and the presence of forgery artifacts. Through the combined effect of the above modules, the influence of forgery-irrelevant patterns is alleviated, and extensive experiments on five widely used datasets demonstrate that RealID significantly outperforms existing state-of-the-art methods, achieving a 1.74% improvement in average accuracy.