Request for Coarse-Grained Geographic Distribution Information of the Mapillary Vistas Dataset

[ywher]

Hello and thank you for maintaining this dataset.

I am currently conducting research on cross-region robustness and domain generalization in semantic segmentation. The Mapillary Vistas Dataset is particularly valuable for this purpose due to its diverse street-level imagery collected from around the world.

I understand that GPS and EXIF metadata were removed from the public release to comply with privacy and GDPR regulations, and I am not requesting any precise location data.

Instead, I would like to ask whether the dataset creators could provide coarse-grained, anonymized geographic information, such as:

1. Aggregated image distribution per country or region, or

2. A mapping from image identifiers to country-/city-level labels, if such labels exist internally in a privacy-safe form.

This information would allow us to analyze the dataset’s geographical diversity and better understand how regional visual differences affect semantic segmentation performance — without attempting any form of re-identification.

Our research is strictly academic, and the data would be used solely for distribution analysis and robustness evaluation, not for any commercial or identity-related purpose.

If such aggregated metadata can be shared (even under research data usage terms), it would be extremely helpful to the research community.

Thank you very much for your time, and for your continued work on the Mapillary Vistas dataset.

Best regards,