Image Classification on Im2GPS3k

Metric: City level (25 km) (higher is better)

LeaderboardDataset

Loading chart...

Results

Sort:

#	Model↕	City level (25 km)▼	Extra Data	Paper	Date↕	Code
1	GeoRanker	45.05	No	GeoRanker: Distance-Aware Ranking for Worldwide ...	2025-05-19	-
2	G3	40.94	No	G3: An Effective and Adaptive Framework for Worl...	2024-05-23	Code
3	PIGEOTTO	36.7	No	PIGEON: Predicting Image Geolocations	2023-07-11	Code
4	GeoCLIP	34.5	No	GeoCLIP: Clip-Inspired Alignment between Locatio...	2023-09-27	Code
5	GeoDecoder	33.5	No	Where We Are and What We're Looking At: Query Ba...	2023-03-07	-
6	Translocator	31.1	No	Where in the World is this Image? Transformer-ba...	2022-04-29	Code
7	ISNs (M, f*, S3)	28	No	-	-	Code
8	base (M, f*)	27	No	-	-	Code
9	CPlaNet (1-5, PlaNet)	26.5	No	CPlaNet: Enhancing Image Geolocalization by Comb...	2018-08-06	-
10	base (L, m)	24.9	No	-	-	Code
11	StreetCLIP (Zero-Shot)	22.4	No	Learning Generalized Zero-Shot Learners for Open...	2023-02-01	Code
12	Im2GPS (kNN, sigma = 4)	19.4	No	Revisiting IM2GPS in the Deep Learning Era	2017-05-13	-
13	Im2GPS ([L] 7011C)	14.8	No	Revisiting IM2GPS in the Deep Learning Era	2017-05-13	-
14	Im2GPS ([M] 7011C)	14.2	No	Revisiting IM2GPS in the Deep Learning Era	2017-05-13	-

#1GeoRankerSOTA
45.05
City level (25 km)· 2025-05-19
GeoRanker: Distance-Aware Ranking for Worldwide Image Geolocalization
#2G3SOTA
40.94
City level (25 km)· 2024-05-23
G3: An Effective and Adaptive Framework for Worldwide Geolocalization Using Large Multi-Modality Models Code
#3PIGEOTTOSOTA
36.7
City level (25 km)· 2023-07-11
PIGEON: Predicting Image Geolocations Code
#4GeoCLIP
34.5
City level (25 km)· 2023-09-27
GeoCLIP: Clip-Inspired Alignment between Locations and Images for Effective Worldwide Geo-localization Code
#5GeoDecoderSOTA
33.5
City level (25 km)· 2023-03-07
Where We Are and What We're Looking At: Query Based Worldwide Image Geo-localization Using Hierarchies and Scenes
#6TranslocatorSOTA
31.1
City level (25 km)· 2022-04-29
Where in the World is this Image? Transformer-based Geo-localization in the Wild Code
#7ISNs (M, f*, S3)
28
City level (25 km)
No paperCode
#8base (M, f*)
27
City level (25 km)
No paperCode
#9CPlaNet (1-5, PlaNet)SOTA
26.5
City level (25 km)· 2018-08-06
CPlaNet: Enhancing Image Geolocalization by Combinatorial Partitioning of Maps
#10base (L, m)
24.9
City level (25 km)
No paperCode
#11StreetCLIP (Zero-Shot)
22.4
City level (25 km)· 2023-02-01
Learning Generalized Zero-Shot Learners for Open-Domain Image Geolocalization Code
#12Im2GPS (kNN, sigma = 4)SOTA
19.4
City level (25 km)· 2017-05-13
Revisiting IM2GPS in the Deep Learning Era
#13Im2GPS ([L] 7011C)
14.8
City level (25 km)· 2017-05-13
Revisiting IM2GPS in the Deep Learning Era
#14Im2GPS ([M] 7011C)
14.2
City level (25 km)· 2017-05-13
Revisiting IM2GPS in the Deep Learning Era