HyperAI超神経

プラットフォーム

Visual Place Recognition On St Lucia

評価指標

Recall@1

Recall@5

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
EffoVPR	100.0	100.0	EffoVPR: Effective Foundation Model Utilization for Visual Place Recognition
BoQ (DINOv2)	100.0	100	BoQ: A Place is Worth a Bag of Learnable Queries
SelaVPR	99.8	-	Towards Seamless Adaptation of Pre-trained Models for Visual Place Recognition
ProGEO	99.7	99.9	ProGEO: Generating Prompts through Image-Text Contrastive Learning for Visual Geo-localization
MixVPR	99.66	-	MixVPR: Feature Mixing for Visual Place Recognition
CosPlace	99.59	99.9	Rethinking Visual Geo-localization for Large-Scale Applications
AnyLoc-VLAD-DINOv2	96.17	-	AnyLoc: Towards Universal Visual Place Recognition
DINOv2	78.62	-	DINOv2: Learning Robust Visual Features without Supervision
CLIP	62.7	-	AnyLoc: Towards Universal Visual Place Recognition
NetVLAD	57.92	-	NetVLAD: CNN architecture for weakly supervised place recognition
DINO	45.22	-	Emerging Properties in Self-Supervised Vision Transformers
BoQ	-	100	BoQ: A Place is Worth a Bag of Learnable Queries

0 of 12 row(s) selected.

Visual Place Recognition On St Lucia | SOTA | HyperAI超神経