🏢 State Key Lab of CAD&CG, Zhejiang University
EMVP: Embracing Visual Foundation Model for Visual Place Recognition with Centroid-Free Probing
·3286 words·16 mins·
loading
·
loading
AI Generated
Computer Vision
Visual Question Answering
🏢 State Key Lab of CAD&CG, Zhejiang University
EMVP: A novel PEFT pipeline boosts Visual Place Recognition accuracy by 97.6% using Centroid-Free Probing & Dynamic Power Normalization, saving 64.3% of parameters.