Steering Vision-Language Pre-trained Models for Incremental Face Presentation Attack Detection

TL;DR

This paper introduces SVLP-IL, a novel framework leveraging vision-language pre-trained models with multi-aspect prompting and selective weight consolidation to improve incremental face presentation attack detection without retaining past data.

Contribution

It proposes a rehearsal-free incremental learning method using VLP models, combining multi-aspect prompting and selective weight consolidation for robust, privacy-compliant face PAD.

Findings

Reduces catastrophic forgetting significantly.

Improves detection performance on unseen domains.

Outperforms existing incremental learning methods.

Abstract

Face Presentation Attack Detection (PAD) demands incremental learning (IL) to combat evolving spoofing tactics and domains. Privacy regulations, however, forbid retaining past data, necessitating rehearsal-free IL (RF-IL). Vision-Language Pre-trained (VLP) models, with their prompt-tunable cross-modal representations, enable efficient adaptation to new spoofing styles and domains. Capitalizing on this strength, we propose \textbf{SVLP-IL}, a VLP-based RF-IL framework that balances stability and plasticity via \textit{Multi-Aspect Prompting} (MAP) and \textit{Selective Elastic Weight Consolidation} (SEWC). MAP isolates domain dependencies, enhances distribution-shift sensitivity, and mitigates forgetting by jointly exploiting universal and domain-specific cues. SEWC selectively preserves critical weights from previous tasks, retaining essential knowledge while allowing flexibility for new adaptations. Comprehensive experiments across multiple PAD benchmarks show that SVLP-IL significantly reduces catastrophic forgetting and enhances performance on unseen domains. SVLP-IL offers a privacy-compliant, practical solution for robust lifelong PAD deployment in RF-IL settings.