Visible-guided multigranularity prompt learning for visible-infrared person re-identification

Key Points

• Person re-identification accuracy improves through novel multigranularity prompt learning methods, boosting recognition in diverse lighting conditions.
• Top metric indicates a 25% increase in matching performance compared to traditional methods, highlighting the effectiveness of visible-infrared approaches.
• Assessment using advanced prompt learning techniques evaluates data from 1,000 visible and infrared image pairs for robust analysis.
• This work supports developing techniques in AI to enhance recognition systems under varying visibility conditions.