Open access
Jul 2026
CLIP-Guided Progressive Body-Part Semantic Alignment for Visible-Infrared Person Re-Identification
A CLIP-Guided Progressive Body-Part Semantic Alignment Network is proposed, termed PBSA-Net, which introduces CLIP-derived textual semantics as modality-agnostic guidance for both global representation learning and local body-part feature extraction.
Hongjin Huang, Xia Geng
· Algorithms · 0 citations