ML Security PapersFace2Parts: Exploring Coarse-to-Fine Inter-Regional Facial Dependencies for Generalized Deepfake Detection
Kutub Uddin , Nusrat Tasnim , Byung Tae Oh
Published on arXiv
2603.26036
Output Integrity Attack
OWASP ML Top 10 — ML09
Key Finding
Achieves 98.42% AUC on FF++ and demonstrates strong cross-dataset generalization with 89.41% on DFD and 84.07% on DFDC
Face2Parts
Novel technique introduced
Multimedia data, particularly images and videos, is integral to various applications, including surveillance, visual interaction, biometrics, evidence gathering, and advertising. However, amateur or skilled counterfeiters can simulate them to create deepfakes, often for slanderous motives. To address this challenge, several forensic methods have been developed to ensure the authenticity of the content. The effectiveness of these methods depends on their focus, with challenges arising from the diverse nature of manipulations. In this article, we analyze existing forensic methods and observe that each method has unique strengths in detecting deepfake traces by focusing on specific facial regions, such as the frame, face, lips, eyes, or nose. Considering these insights, we propose a novel hybrid approach called Face2Parts based on hierarchical feature representation ($HFR$) that takes advantage of coarse-to-fine information to improve deepfake detection. The proposed method involves extracting features from the frame, face, and key facial regions (i.e., lips, eyes, and nose) separately to explore the coarse-to-fine relationships. This approach enables us to capture inter-dependencies among facial regions using a channel-attention mechanism and deep triplet learning. We evaluated the proposed method on benchmark deepfake datasets in both intra-, inter-dataset, and inter-manipulation settings. The proposed method achieves an average AUC of 98.42\% on FF++, 79.80\% on CDF1, 85.34\% on CDF2, 89.41\% on DFD, 84.07\% on DFDC, 95.62\% on DTIM, 80.76\% on PDD, and 100\% on WLDR, respectively. The results demonstrate that our approach generalizes effectively and achieves promising performance to outperform the existing methods.
Key Contributions
- Hierarchical feature representation (HFR) approach extracting coarse-to-fine features from frame, face, and facial parts (eyes, lips, nose)
- Channel-attention mechanism with deep triplet learning to capture inter-dependencies among facial regions
- Cross-dataset generalization achieving 98.42% AUC on FF++, 89.41% on DFD, 84.07% on DFDC across 8 benchmark datasets
🛡️ Threat Analysis
Detects AI-generated deepfake videos by analyzing hierarchical facial features — this is AI-generated content detection, a core ML09 use case for output integrity and content authenticity.