Decoupling Bias, Aligning Distributions: Synergistic Fairness Optimization for Deepfake Detection

Fairness is a core element in the trustworthy deployment of deepfake detection models, especially in the field of digital identity security. Biases in detection models toward different demographic groups, such as gender and race, may lead to systemic misjudgments, exacerbating the digital divide and social inequities. However, current fairness-enhanced detectors often improve fairness at the cost of detection accuracy. To address this challenge, we propose a dual-mechanism collaborative optimization framework. Our proposed method innovatively integrates structural fairness decoupling and global distribution alignment: decoupling channels sensitive to demographic groups at the model architectural level, and subsequently reducing the distance between the overall sample distribution and the distributions corresponding to each demographic group at the feature level. Experimental results demonstrate that, compared with other methods, our framework improves both inter-group and intra-group fairness while maintaining overall detection accuracy across domains. The code is available at https://github.com/ywh1093/Fairness-Optimization.

• Publication year

  2025

• Venue

  Unknown venue

• Publication date

  2025-11-13

• Fields of study

  Computer Science

• Identifiers
  arXiv 2511.10150
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Deepfake Detection

  2025cited by this paper
• Fair Deepfake Detectors Can Generalize

  2025cited by this paper
• Breaking Latent Prior Bias in Detectors for Generalizable AIGC Image Detection

  2025cited by this paper
• Generating Higher-Quality Anti-Forensics DeepFakes with Adversarial Sharpening Mask

  2025cited by this paper
• VLForgery Face Triad: Detection, Localization and Attribution via Multimodal Large Language Models

  2025cited by this paper
• FairAdapter: Detecting AI-generated Images with Improved Fairness

  2024cited by this paper
• Disrupting Anti-Spoofing Systems by Images of Consistent Identity

  2024cited by this paper
• Thinking Racial Bias in Fair Forgery Detection: Models, Datasets and Evaluations

  2024cited by this paper
• Decoupling Forgery Semantics for Generalizable Deepfake Detection

  2024cited by this paper
• Preserving Fairness Generalization in Deepfake Detection

  2024influential reference
• FRAPPÉ: A Group Fairness Framework for Post-Processing Everything

  2023cited by this paper
• Improving Fairness in Deepfake Detection

  2023influential reference
• Securing Facial Bioinformation by Eliminating Adversarial Perturbations

  2023cited by this paper
• A review of deep learning‐based approaches for deepfake content detection

  2022cited by this paper
• Anti-Forensics for Face Swapping Videos via Adversarial Training

  2022cited by this paper
• Fairness-aware Adversarial Perturbation Towards Bias Mitigation for Deployed Deep Models

  2022cited by this paper
• Fairness Evaluation in Deepfake Detection Models using Metamorphic Testing

  2022cited by this paper
• To the Fairness Frontier and Beyond: Identifying, Quantifying, and Optimizing the Fairness-Accuracy Pareto Frontier

  2022cited by this paper
• Learning a deep dual-level network for robust DeepFake detection

  2022cited by this paper
• Deepfake noise investigation and detection

  2022cited by this paper
• End-to-End Reconstruction-Classification Learning for Face Forgery Detection

  2022cited by this paper
• GBDF: Gender Balanced DeepFake Dataset Towards Fair DeepFake Detection

  2022cited by this paper
• The Deepfake Detection Dilemma: A Multistakeholder Exploration of Adversarial Dynamics in Synthetic Media

  2021cited by this paper
• Gendered Differences in Face Recognition Accuracy Explained by Hairstyles, Makeup, and Facial Morphology

  2021cited by this paper
• An Experimental Evaluation on Deepfake Detection using Deep Face Recognition

  2021cited by this paper
• Towards Measuring Fairness in AI: The Casual Conversations Dataset

  2021cited by this paper
• Robust Attentive Deep Neural Network for Detecting GAN-Generated Faces

  2021cited by this paper
• An Examination of Fairness of AI Models for Deepfake Detection

  2021cited by this paper
• Deepfakes generation and detection: state-of-the-art, open challenges, countermeasures, and way forward

  2021cited by this paper
• DeepFakes and Beyond: A Survey of Face Manipulation and Fake Detection

  2020cited by this paper
• Understanding Fairness of Gender Classification Algorithms Across Gender-Race Groups

  2020cited by this paper
• Fairness by Learning Orthogonal Disentangled Representations

  2020cited by this paper
• FR-Train: A mutual information-based approach to fair and robust training

  2020cited by this paper
• FaceForensics++: Learning to Detect Manipulated Facial Images

  2019cited by this paper
• Celeb-DF: A Large-Scale Challenging Dataset for DeepFake Forensics

  2019cited by this paper
• On Fairness and Calibration

  2017cited by this paper
• Equality of Opportunity in Supervised Learning

  2016cited by this paper
• Xception: Deep Learning with Depthwise Separable Convolutions

  2016cited by this paper
• Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization

  2016cited by this paper
• Deep Residual Learning for Image Recognition

  2015cited by this paper

• No citing papers are available for this paper.