Visual Place Recognition On Vp Air

Recall@1

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

		Paper Title
AnyLoc-VLAD-DINOv2	66.74	AnyLoc: Towards Universal Visual Place Recognition
DINOv2	45.23	DINOv2: Learning Robust Visual Features without Supervision
CLIP	36.59	AnyLoc: Towards Universal Visual Place Recognition
DINO	24.02	Emerging Properties in Self-Supervised Vision Transformers
MixVPR	10.31	MixVPR: Feature Mixing for Visual Place Recognition
CosPlace	8.12	Rethinking Visual Geo-localization for Large-Scale Applications
NetVLAD	6.39	NetVLAD: CNN architecture for weakly supervised place recognition

0 of 7 row(s) selected.