Skip to yearly menu bar Skip to main content


Poster

OpenStreetView-5M: The Many Roads to Global Visual Geolocation

Guillaume Astruc · Nicolas Dufour · Ioannis Siglidis · Constantin Aronssohn · Nacim Bouia · Stephanie Fu · Romain Loiseau · Van Nguyen Nguyen · Charles Raude · Elliot Vincent · Lintao XU · Hongyu Zhou · Loic Landrieu


Abstract:

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.

Live content is unavailable. Log in and register to view live content