| Paper | Code | Median Error (km) | Street level (1 km) | City level (25 km) | Region level (200 km) | Country level (750 km) | Continent level (2500 km) | Training images | Reference images | ModelName | ReleaseDate |
|---|---|---|---|---|---|---|---|---|---|---|---|
| PIGEON: Predicting Image Geolocations | ✓ Link | 70.5 | 14.8 | 40.9 | 63.3 | 82.3 | 91.1 | 4.5M | 4.5M | PIGEOTTO | 2023-07-11 |
| Geolocation Estimation of Photos using a Hierarchical Model and Scene Classification | ✓ Link | 16.9 | 43.0 | 51.9 | 66.7 | 80.2 | 4.7M | 0 | ISNs (M, f*, S3) | 2018-09-01 | |
| CPlaNet: Enhancing Image Geolocalization by Combinatorial Partitioning of Maps | 16.5 | 37.1 | 46.6 | 62.0 | 78.5 | 30.3M | 0 | CPlaNet (1-5, PlaNet) | 2018-08-06 | ||
| Geolocation Estimation of Photos using a Hierarchical Model and Scene Classification | ✓ Link | 15.2 | 40.9 | 51.5 | 65.4 | 78.5 | 4.7M | 0 | base (M, f*) | 2018-09-01 | |
| Revisiting IM2GPS in the Deep Learning Era | 14.4 | 33.3 | 47.7 | 61.6 | 73.4 | 6M | 28M | Im2GPS (... 28m database) | 2017-05-13 | ||
| Geolocation Estimation of Photos using a Hierarchical Model and Scene Classification | ✓ Link | 13.5 | 35.0 | 49.8 | 64.1 | 79.7 | 4.7M | 0 | base (L, m) | 2018-09-01 | |
| Revisiting IM2GPS in the Deep Learning Era | 12.2 | 33.3 | 44.3 | 57.4 | 71.3 | 6M | 0 | Im2GPS ([L] KNN, sigma=4) | 2017-05-13 | ||
| PlaNet - Photo Geolocation with Convolutional Neural Networks | ✓ Link | 8.4 | 24.5 | 37.6 | 53.6 | 71.3 | 91M | 0 | PlaNet (91M) | 2016-02-17 | |
| Revisiting IM2GPS in the Deep Learning Era | 6.8 | 21.9 | 34.6 | 49.4 | 63.7 | 6M | 0 | Im2GPS ([L] 7011C) | 2017-05-13 | ||
| PlaNet - Photo Geolocation with Convolutional Neural Networks | ✓ Link | 6.3 | 18.1 | 30.0 | 45.6 | 65.8 | 6.2M | 0 | PlaNet (6.2M) | 2016-02-17 | |
| Learning Generalized Zero-Shot Learners for Open-Domain Image Geolocalization | ✓ Link | 28.3 | 45.1 | 74.7 | 88.2 | 1.1M | 0 | StreetCLIP (Zero-Shot) | 2023-02-01 |