Learning Knowledge-based Prompts for Robust 3D Mask Presentation Attack Detection

2025-05-06

AI Summary

To address poor generalization and high deployment costs in 3D mask attack detection, this paper pioneers the integration of vision-language multimodal prompt learning into this task, proposing a novel knowledge graph (KG)-driven detection framework. Methodologically, it synergizes KG embeddings, attention-based vision–knowledge alignment, text-guided local patch filtering, and prompt-tuning of vision-language models. Key contributions include: (1) an explicit prompt generation mechanism grounded in KG entities and triples; (2) a vision-perception-aware knowledge filtering module; and (3) a causal graph-guided training paradigm that mitigates spurious correlations. Evaluated on multiple benchmarks, the method achieves state-of-the-art performance, demonstrating significant improvements in generalization to unseen attack types and robustness across diverse acquisition conditions—including cross-device setups and varying illumination.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

3D mask presentation attack detection is crucial for protecting face recognition systems against the rising threat of 3D mask attacks. While most existing methods utilize multimodal features or remote photoplethysmography (rPPG) signals to distinguish between real faces and 3D masks, they face significant challenges, such as the high costs associated with multimodal sensors and limited generalization ability. Detection-related text descriptions offer concise, universal information and are cost-effective to obtain. However, the potential of vision-language multimodal features for 3D mask presentation attack detection remains unexplored. In this paper, we propose a novel knowledge-based prompt learning framework to explore the strong generalization capability of vision-language models for 3D mask presentation attack detection. Specifically, our approach incorporates entities and triples from knowledge graphs into the prompt learning process, generating fine-grained, task-specific explicit prompts that effectively harness the knowledge embedded in pre-trained vision-language models. Furthermore, considering different input images may emphasize distinct knowledge graph elements, we introduce a visual-specific knowledge filter based on an attention mechanism to refine relevant elements according to the visual context. Additionally, we leverage causal graph theory insights into the prompt learning process to further enhance the generalization ability of our method. During training, a spurious correlation elimination paradigm is employed, which removes category-irrelevant local image patches using guidance from knowledge-based text features, fostering the learning of generalized causal prompts that align with category-relevant local patches. Experimental results demonstrate that the proposed method achieves state-of-the-art intra- and cross-scenario detection performance on benchmark datasets.

Problem

• Detect 3D mask attacks on face recognition systems robustly
• Overcome high costs and limited generalization in existing methods
• Explore vision-language models for attack detection via knowledge-based prompts

Innovation

• Knowledge-based prompt learning for 3D mask detection
• Visual-specific knowledge filter using attention mechanism
• Spurious correlation elimination for generalized causal prompts