SoK: Facial Deepfake Detectors

2024-01-09

arXiv.org

AI Summary

Existing deepfake detection methods predominantly rely on lab-curated datasets, raising concerns about their generalizability to real-world scenarios. Method: We systematically evaluate 16 state-of-the-art detectors under black-box, white-box, and gray-box adversarial settings to assess cross-scenario robustness. We propose the first standardized taxonomy for deepfake detectors—comprising four high-level categories and thirteen fine-grained subcategories—and a comprehensive evaluation paradigm integrating cross-dataset generalization testing, adversarial attack modeling, and multi-dimensional performance analysis. Contribution/Results: Our empirical study reveals critical factors governing detection efficacy; notably, most detectors suffer substantial performance degradation on emerging or black-box-generated deepfakes. The framework validates the taxonomy’s utility in guiding detector design and defense strategies, and advances deepfake detection from controlled benchmarking toward real-world robustness—establishing both theoretical foundations and practical pathways for next-generation detection technologies.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Adversarial Learning & Robustness
• Multiagent Systems: Adversarial Agents

Application Category

• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Deepfakes have rapidly emerged as a profound and serious threat to society, primarily due to their ease of creation and dissemination. This situation has triggered an accelerated development of deepfake detection technologies. However, many existing detectors rely heavily on lab-generated datasets for validation, which may not effectively prepare them for novel, emerging, and real-world deepfake techniques. In this paper, we conduct an extensive and comprehensive review and analysis of the latest state-of-the-art deepfake detectors, evaluating them against several critical criteria. These criteria facilitate the categorization of these detectors into 4 high-level groups and 13 fine-grained sub-groups, all aligned with a unified standard conceptual framework. This classification and framework offer deep and practical insights into the factors that affect detector efficacy. We assess the generalizability of 16 leading detectors across various standard attack scenarios, including black-box, white-box, and gray-box settings. Our systematized analysis and experimentation lay the groundwork for a deeper understanding of deepfake detectors and their generalizability, paving the way for future research focused on creating detectors adept at countering various attack scenarios. Additionally, this work offers insights for developing more proactive defenses against deepfakes.

Problem

• evaluate deepfake detection technologies
• classify detectors into unified framework
• assess detector generalizability in attack scenarios

Innovation

• Comprehensive review of detectors
• Classification into 13 sub-groups
• Assess generalizability across attack scenarios