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

TL;DR

This paper introduces I-FAS, a multimodal large language model framework for face anti-spoofing that offers interpretable results and improved generalization across diverse datasets by transforming the task into a visual question answering paradigm.

Contribution

The work proposes a novel interpretable FAS framework using multimodal LLMs, with a captioning strategy, a specialized loss function, and a global visual feature alignment technique.

Findings

Outperforms state-of-the-art methods on multiple benchmarks

Achieves better generalization in cross-domain scenarios

Provides interpretable visual question answering outputs

Abstract

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.