The field of face anti-spoofing and security is moving towards more robust and generalizable solutions. Researchers are exploring new approaches to detect and prevent face spoofing attacks, including the use of content-aware composite prompts, unified frameworks for physical and digital attack detection, and self-supervised learning methods. These innovations aim to improve the security and privacy of face recognition systems. Noteworthy papers include: Domain Generalization for Face Anti-spoofing via Content-aware Composite Prompt Engineering, which proposes a novel prompt engineering method for improved generalization. SelfMAD: Enhancing Generalization and Robustness in Morphing Attack Detection via Self-Supervised Learning, which presents a self-supervised approach for detecting morphing attacks.
Advances in Face Anti-Spoofing and Security
Sources
Domain Generalization for Face Anti-spoofing via Content-aware Composite Prompt Engineering
From Specificity to Generality: Revisiting Generalizable Artifacts in Detecting Face Deepfakes
SUEDE:Shared Unified Experts for Physical-Digital Face Attack Detection Enhancement
SelfMAD: Enhancing Generalization and Robustness in Morphing Attack Detection via Self-Supervised Learning
Security Analysis of Thumbnail-Preserving Image Encryption and a New Framework
FaceCloak: Learning to Protect Face Templates
PEEL the Layers and Find Yourself: Revisiting Inference-time Data Leakage for Residual Neural Networks