# Mixture-of-Attack-Experts with Class Regularization for Unified Physical-Digital Face Attack Detection

arXiv: https://arxiv.org/abs/2504.00458 · HTML: https://mlsanity.com/paper/2504.00458/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in Vision mixture-of-experts, one of 12 directions in Mixture-of-experts.
- Builds on 27 of its 41 references that are in the map.
- First posted on arXiv 1 Apr 2025.

## Placed in

- Mixture-of-experts / [Vision mixture-of-experts](https://mlsanity.com/moe/vision-mixture-of-experts/)

## Builds on (most cited first, 27 of 27)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [FaceNet: A Unified Embedding for Face Recognition and Clustering](https://mlsanity.com/paper/1503.03832/) (2015)
- [Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention](https://mlsanity.com/paper/2006.16236/) (2020)
- [FaceForensics++: Learning to Detect Manipulated Facial Images](https://mlsanity.com/paper/1901.08971/) (2019)
- [Thinking in Frequency: Face Forgery Detection by Mining Frequency-aware Clues](https://mlsanity.com/paper/2007.09355/) (2020)
- [Multi-attentional Deepfake Detection](https://mlsanity.com/paper/2103.02406/) (2021)
- [On the Detection of Digital Face Manipulation](https://mlsanity.com/paper/1910.01717/) (2019)
- [From Sparse to Soft Mixtures of Experts](https://mlsanity.com/paper/2308.00951/) (2023)
- [DeepFakes and Beyond: A Survey of Face Manipulation and Fake Detection](https://mlsanity.com/paper/2001.00179/) (2020)
- [Learning Fine-grained Image Similarity with Deep Ranking](https://mlsanity.com/paper/1404.4661/) (2014)
- [Capsule-Forensics: Using Capsule Networks to Detect Forged Images and Videos](https://mlsanity.com/paper/1810.11215/) (2018)
- [Learning Deep Models for Face Anti-Spoofing: Binary or Auxiliary Supervision](https://mlsanity.com/paper/1803.11097/) (2018)
- [NAS-FAS: Static-Dynamic Central Difference Network Search for Face Anti-Spoofing](https://mlsanity.com/paper/2011.02062/) (2020)
- [Biometric Face Presentation Attack Detection with Multi-Channel Convolutional Neural Network](https://mlsanity.com/paper/1909.08848/) (2019)
- [An empirical study of LLaMA3 quantization: from LLMs to MLLMs](https://mlsanity.com/paper/2404.14047/) (2024)
- [CASIA-SURF: A Large-scale Multi-modal Benchmark for Face Anti-spoofing](https://mlsanity.com/paper/1908.10654/) (2019)
- [A Dataset and Benchmark for Large-scale Multi-modal Face Anti-spoofing](https://mlsanity.com/paper/1812.00408/) (2018)
- [Cross Modal Focal Loss for RGBD Face Anti-Spoofing](https://mlsanity.com/paper/2103.00948/) (2021)
- [On Learning Multi-Modal Forgery Representation for Diffusion Generated Video Detection](https://mlsanity.com/paper/2410.23623/) (2024)
- [Multi-domain Learning for Updating Face Anti-spoofing Models](https://mlsanity.com/paper/2208.11148/) (2022)
- [Unified Detection of Digital and Physical Face Attacks](https://mlsanity.com/paper/2104.02156/) (2021)
- [Benchmarking Joint Face Spoofing and Forgery Detection with Visual and Physiological Cues](https://mlsanity.com/paper/2208.05401/) (2022)
- [Flexible-Modal Face Anti-Spoofing: A Benchmark](https://mlsanity.com/paper/2202.08192/) (2022)
- [Unified Physical-Digital Face Attack Detection](https://mlsanity.com/paper/2401.17699/) (2024)
- [Towards Data-Centric Face Anti-Spoofing: Improving Cross-domain Generalization via Physics-based Data Synthesis](https://mlsanity.com/paper/2409.03501/) (2024)
- [La-SoftMoE CLIP for Unified Physical-Digital Face Attack Detection](https://mlsanity.com/paper/2408.12793/) (2024)
- [Flexible-modal Deception Detection with Audio-Visual Adapter](https://mlsanity.com/paper/2302.05727/) (2023)

## Built on by (most cited first, 5 of 5)

- [Benchmarking Unified Face Attack Detection via Hierarchical Prompt Tuning](https://mlsanity.com/paper/2505.13327/) (2025)
- [InstructFLIP: Exploring Unified Vision-Language Model for Face Anti-spoofing](https://mlsanity.com/paper/2507.12060/) (2025)
- [Multi-View Slot Attention Using Paraphrased Texts for Face Anti-Spoofing](https://mlsanity.com/paper/2509.06336/) (2025)
- [Spoofing-aware Prompt Learning for Unified Physical-Digital Facial Attack Detection](https://mlsanity.com/paper/2512.06363/) (2025)
- [DINO-VPT: Hierarchical Visual Prompt Tuning for Joint Physical-Digital Face Anti-Spoofing](https://mlsanity.com/paper/2607.20900/) (2026)

## Abstract (arXiv)

Facial recognition systems in real-world scenarios are susceptible to both digital and physical attacks. Previous methods have attempted to achieve classification by learning a comprehensive feature space. However, these methods have not adequately accounted for the inherent characteristics of physical and digital attack data, particularly the large intra class variation in attacks and the small inter-class variation between live and fake faces. To address these limitations, we propose the Fine-Grained MoE with Class-Aware Regularization CLIP framework (FG-MoE-CLIP-CAR), incorporating key improvements at both the feature and loss levels. At the feature level, we employ a Soft Mixture of Experts (Soft MoE) architecture to leverage different experts for specialized feature processing. Additionally, we refine the Soft MoE to capture more subtle differences among various types of fake faces. At the loss level, we introduce two constraint modules: the Disentanglement Module (DM) and the Cluster Distillation Module (CDM). The DM enhances class separability by increasing the distance between the centers of live and fake face classes. However, center-to-center constraints alone are insufficient to ensure distinctive representations for individual features. Thus, we propose the CDM to further cluster features around their respective class centers while maintaining separation from other classes. Moreover, specific attacks that significantly deviate from common attack patterns are often overlooked. To address this issue, our distance calculation prioritizes more distant features. Experimental results on two unified physical-digital attack datasets demonstrate that the proposed method achieves state-of-the-art (SOTA) performance.