AI-powered Game Recognition: A Collaborative Dataset for Traditional Games
DIAL (Catholic University of Leuven)
Abstract
We introduce an expert-annotated image dataset of traditional board games and benchmark supervised visual retrieval against the Ludii corpus. Given a photo, our models will rank likely matches with top-k metrics, and we will provide a lightweight web tool for archaeologists to get interpretable suggestions on site. The resource is designed to enable reproducible research and broaden access to cultural-heritage game data.
Authors 0
- Author list not loaded yet.