# Interpretable Face Anti-Spoofing: Enhancing Generalization with Multimodal Large Language Models

arXiv: https://arxiv.org/abs/2501.01720 · HTML: https://mlsanity.com/paper/2501.01720/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Placed in MLLM jailbreaks and adversarial robustness, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 4 of its 60 references that are in the map.
- First posted on arXiv 3 Jan 2025.

## Placed in

- Multimodal LLMs / vision-language models / [MLLM jailbreaks and adversarial robustness](https://mlsanity.com/mllm/mllm-jailbreaks-and-adversarial-robustness/)

## Builds on (most cited first, 4 of 4)

- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [From CLIP to DINO: Visual Encoders Shout in Multi-modal Large Language Models](https://mlsanity.com/paper/2310.08825/) (2023)
- [An empirical study of LLaMA3 quantization: from LLMs to MLLMs](https://mlsanity.com/paper/2404.14047/) (2024)
- [SHIELD : An Evaluation Benchmark for Face Spoofing and Forgery Detection with Multimodal Large Language Models](https://mlsanity.com/paper/2402.04178/) (2024)

## Built on by (most cited first, 11 of 11)

- [FaceShield: Explainable Face Anti-Spoofing with Multimodal Large Language Models](https://mlsanity.com/paper/2505.09415/) (2025)
- [Harnessing Chain-of-Thought Reasoning in Multimodal Large Language Models for Face Anti-Spoofing](https://mlsanity.com/paper/2506.01783/) (2025)
- [Benchmarking Vision Foundation Models for Domain-Generalizable Face Anti-Spoofing](https://mlsanity.com/paper/2604.19196/) (2026)
- [From Intuition to Investigation: A Tool-Augmented Reasoning MLLM Framework for Generalizable Face Anti-Spoofing](https://mlsanity.com/paper/2603.01038/) (2026)
- [Learning Unknown Spoof Prompts for Generalized Face Anti-Spoofing Using Only Real Face Images](https://mlsanity.com/paper/2505.03611/) (2025)
- [CPG-PAD: Concept-Informed Prompts Guided Presentation Attack Detection](https://mlsanity.com/paper/2607.01303/) (2026)
- [Foundation Models for Face Presentation Attack Detection: A Unified Linear-Probing Benchmark](https://mlsanity.com/paper/2607.26993/) (2026)
- [DINO-VPT: Hierarchical Visual Prompt Tuning for Joint Physical-Digital Face Anti-Spoofing](https://mlsanity.com/paper/2607.20900/) (2026)
- [LoRA-based Adaptation Alone Is Not Enough: Understanding the Limits of Foundation Models for Face Presentation Attack Detection](https://mlsanity.com/paper/2608.09633/) (2026)
- [Primitive-Driven Compositional Forensic Visual Prompting for Open-World Face Anti-Spoofing](https://mlsanity.com/paper/2608.17351/) (2026)
- [Foundation and Multimodal Large Language Models for Face Presentation and Morph Attack Detection](https://mlsanity.com/paper/2608.29802/) (2026)

## Abstract (arXiv)

Face Anti-Spoofing (FAS) is essential for ensuring the security and reliability of facial recognition systems. Most existing FAS methods are formulated as binary classification tasks, providing confidence scores without interpretation. They exhibit limited generalization in out-of-domain scenarios, such as new environments or unseen spoofing types. In this work, we introduce a multimodal large language model (MLLM) framework for FAS, termed Interpretable Face Anti-Spoofing (I-FAS), which transforms the FAS task into an interpretable visual question answering (VQA) paradigm. Specifically, we propose a Spoof-aware Captioning and Filtering (SCF) strategy to generate high-quality captions for FAS images, enriching the model's supervision with natural language interpretations. To mitigate the impact of noisy captions during training, we develop a Lopsided Language Model (L-LM) loss function that separates loss calculations for judgment and interpretation, prioritizing the optimization of the former. Furthermore, to enhance the model's perception of global visual features, we design a Globally Aware Connector (GAC) to align multi-level visual representations with the language model. Extensive experiments on standard and newly devised One to Eleven cross-domain benchmarks, comprising 12 public datasets, demonstrate that our method significantly outperforms state-of-the-art methods.