New framework tackles disguise makeup attacks on facial recognition systems

Researchers have developed a novel framework to detect disguise makeup presentation attacks, which are particularly challenging for facial recognition systems. The proposed method uses a two-phase approach: first, a style-invariant full-face model extracts attention scores, and second, a patch-based analysis performs localized discrimination. This framework was tested on a newly constructed dataset and demonstrated strong generalization capabilities, outperforming previous methods. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Improves the robustness of facial recognition systems against sophisticated cosmetic-based spoofing attacks.

RANK_REASON This is a research paper detailing a new framework for a specific AI safety problem.

Read on arXiv cs.CV →

• arXiv
• Grad-CAM
• SIW-Mv2

• paper
• safety