Summary: The presentation is about introducing “MapTCHA,” a new CAPTCHA system that combines computer vision with human verification to improve mapping efforts in OpenStreetMap. The system identifies objects in images and relies on users to confirm or deny their presence, ultimately suggesting locations for mapping in OSM. The project uses fAIr, an AI-assisted mapping system, to train machine learning models for building footprints. Future plans include expanding to more objects, refining AI models, and integrating MapTCHA into login systems. The session will cover the development of the solution, its potential impact on mapping projects, and challenges of working with AI data.
https://fosdem.org/2025/schedule/event/fosdem-2025-5879-maptcha-the-open-source-captcha-that-improves-openstreetmap/