Curated datasets, software, and competitions for the Document Image Analysis and Recognition community.
A community-curated index of datasets, software & models, and competitions for the Document Image Analysis and Recognition (DIAR) community. Each entry is reviewed by a TC10/TC11 curator and links to artifacts hosted on Zenodo, Hugging Face Hub, GitHub, and institutional servers.
The ICDAR 2021 Competition on Historical Map Segmentation (MapSeg) ran from November 2020 to April 2021. Twelve teams from academia and industry submitted results across the three tasks listed above, evaluated on a held-out test set of historical Paris atlas sheets (1894-1937).
A reference U-Net baseline released alongside the ICDAR 2021 MapSeg competition. The model is trained to perform two of the three competition tasks (building-block detection and map-content segmentation) on 1894-1937 atlas sheets of the City of Paris.
Dataset of the ICDAR 2021 Competition on Historical Map Segmentation (“MapSeg”): segmented map sheets with building blocks, map-content masks, and graticule points.
Evaluation tools used to score submissions to the ICDAR 2021 MapSeg competition.
This package provides training and test data from the competitions CROHME 2011, 2012, 2013, 2014, 2016 and 2019.