Saved in:
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2404.18873 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913334503145472 |
|---|---|
| author | Astruc, Guillaume Dufour, Nicolas Siglidis, Ioannis Aronssohn, Constantin Bouia, Nacim Fu, Stephanie Loiseau, Romain Nguyen, Van Nguyen Raude, Charles Vincent, Elliot XU, Lintao Zhou, Hongyu Landrieu, Loic |
| author_facet | Astruc, Guillaume Dufour, Nicolas Siglidis, Ioannis Aronssohn, Constantin Bouia, Nacim Fu, Stephanie Loiseau, Romain Nguyen, Van Nguyen Raude, Charles Vincent, Elliot XU, Lintao Zhou, Hongyu Landrieu, Loic |
| contents | Determining the location of an image anywhere on Earth is a complex visual task, which makes it particularly relevant for evaluating computer vision algorithms. Yet, the absence of standard, large-scale, open-access datasets with reliably localizable images has limited its potential. To address this issue, we introduce OpenStreetView-5M, a large-scale, open-access dataset comprising over 5.1 million geo-referenced street view images, covering 225 countries and territories. In contrast to existing benchmarks, we enforce a strict train/test separation, allowing us to evaluate the relevance of learned geographical features beyond mere memorization. To demonstrate the utility of our dataset, we conduct an extensive benchmark of various state-of-the-art image encoders, spatial representations, and training strategies. All associated codes and models can be found at https://github.com/gastruc/osv5m. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_18873 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | OpenStreetView-5M: The Many Roads to Global Visual Geolocation Astruc, Guillaume Dufour, Nicolas Siglidis, Ioannis Aronssohn, Constantin Bouia, Nacim Fu, Stephanie Loiseau, Romain Nguyen, Van Nguyen Raude, Charles Vincent, Elliot XU, Lintao Zhou, Hongyu Landrieu, Loic Computer Vision and Pattern Recognition Artificial Intelligence Determining the location of an image anywhere on Earth is a complex visual task, which makes it particularly relevant for evaluating computer vision algorithms. Yet, the absence of standard, large-scale, open-access datasets with reliably localizable images has limited its potential. To address this issue, we introduce OpenStreetView-5M, a large-scale, open-access dataset comprising over 5.1 million geo-referenced street view images, covering 225 countries and territories. In contrast to existing benchmarks, we enforce a strict train/test separation, allowing us to evaluate the relevance of learned geographical features beyond mere memorization. To demonstrate the utility of our dataset, we conduct an extensive benchmark of various state-of-the-art image encoders, spatial representations, and training strategies. All associated codes and models can be found at https://github.com/gastruc/osv5m. |
| title | OpenStreetView-5M: The Many Roads to Global Visual Geolocation |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2404.18873 |