Photo Geolocation Estimation On Im2Gps3K

评估指标

City level (25 km)
Continent level (2500 km)
Country level (750 km)
Region level (200 km)
Street level (1 km)
Training Images

评测结果

各个模型在此基准测试上的表现结果

Paper TitleRepository
PIGEOTTO36.785.372.453.811.34.5MPIGEON: Predicting Image Geolocations
GeoCLIP34.583.869.750.714.14.7MGeoCLIP: Clip-Inspired Alignment between Locations and Images for Effective Worldwide Geo-localization
GeoDecoder33.576.161.045.912.84.7MWhere We Are and What We're Looking At: Query Based Worldwide Image Geo-localization Using Hierarchies and Scenes-
Translocator31.180.158.946.711.84.7MWhere in the World is this Image? Transformer-based Geo-localization in the Wild
ISNs (M, f*, S3)28.066.049.736.610.54.7MGeolocation Estimation of Photos using a Hierarchical Model and Scene Classification-
base (M, f*)27.066.049.235.69.74.7MGeolocation Estimation of Photos using a Hierarchical Model and Scene Classification-
CPlaNet (1-5, PlaNet)26.564.448.634.610.230.3MCPlaNet: Enhancing Image Geolocalization by Combinatorial Partitioning of Maps-
base (L, m)24.965.848.834.08.34.7MGeolocation Estimation of Photos using a Hierarchical Model and Scene Classification-
StreetCLIP (Zero-Shot) 22.4 80.461.3 37.4-1.1MLearning Generalized Zero-Shot Learners for Open-Domain Image Geolocalization
Im2GPS (kNN, sigma = 4)19.455.938.926.97.26MRevisiting IM2GPS in the Deep Learning Era-
Im2GPS ([L] 7011C)14.852.432.621.44.06MRevisiting IM2GPS in the Deep Learning Era-
Im2GPS ([M] 7011C)14.252.733.521.33.76MRevisiting IM2GPS in the Deep Learning Era-
0 of 12 row(s) selected.
Photo Geolocation Estimation On Im2Gps3K | SOTA | HyperAI超神经