Spoofing-aware Prompt Learning for Unified Physical-Digital Facial Attack Detection

2025-12-06

AI Summary

Real-world face recognition systems face dual threats from physical presentation attacks (PA) and digital forgery attacks (DF); existing methods struggle to jointly optimize detection for both attack types within a unified prompt space due to conflicting optimization directions. This paper proposes a deception-aware prompt learning framework: (1) a learnable dual-parallel prompt branch architecture that decouples PA- and DF-specific optimization pathways; (2) an adaptive contextual prompt generation mechanism to enhance semantic discriminability; and (3) cue-aware hard-sample mining to improve generalization. Built upon the CLIP architecture, the framework is trained and evaluated on the large-scale UniAttackDataPlus benchmark. Experiments demonstrate significant improvements over state-of-the-art methods, with strong robustness against unseen attack types. To our knowledge, this is the first approach enabling unified, cooperative, and disentangled detection of both physical and digital face attacks.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Adversarial Learning & Robustness
• Natural Language Processing: Prompt Engineering / Prompting

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Real-world face recognition systems are vulnerable to both physical presentation attacks (PAs) and digital forgery attacks (DFs). We aim to achieve comprehensive protection of biometric data by implementing a unified physical-digital defense framework with advanced detection. Existing approaches primarily employ CLIP with regularization constraints to enhance model generalization across both tasks. However, these methods suffer from conflicting optimization directions between physical and digital attack detection under same category prompt spaces. To overcome this limitation, we propose a Spoofing-aware Prompt Learning for Unified Attack Detection (SPL-UAD) framework, which decouples optimization branches for physical and digital attacks in the prompt space. Specifically, we construct a learnable parallel prompt branch enhanced with adaptive Spoofing Context Prompt Generation, enabling independent control of optimization for each attack type. Furthermore, we design a Cues-awareness Augmentation that leverages the dual-prompt mechanism to generate challenging sample mining tasks on data, significantly enhancing the model's robustness against unseen attack types. Extensive experiments on the large-scale UniAttackDataPlus dataset demonstrate that the proposed method achieves significant performance improvements in unified attack detection tasks.

Problem

• Unified detection of physical and digital facial attacks
• Decoupling optimization for conflicting attack detection tasks
• Enhancing robustness against unseen attack types

Innovation

• Decouples optimization branches for physical and digital attacks
• Uses adaptive Spoofing Context Prompt Generation for independent control
• Implements Cues-awareness Augmentation to enhance robustness