ScaDS.AI German Full-Page Handwriting Dataset

Description This dataset contains 77 complete pages of German handwriting, with each page representing an excerpt from a Wikipedia article. For each handwritten page (in JPG data format), a transcription in plain UTF-8 text is provided (can be used as ground truth). Structure - images/: one jpg file per handwritten page (a total of 77 items) - ground_truth/: one text file per handwritten page (a total of 77 items) - metadata/scads_topic_list.csv: dataset description - metadata/scads_topic_list.xml: TEI-Lite description - metadata/Uni_Leipzig_01-export-251205.csv: contains URLs to the corresponding Wikipedia articles - README.md - license.txt The usable dataset includes 77 images with German handwritten Wikipedia excerpts and corresponding ground truth text files. Version * 1.0 (January 17, 2026) Authors * Thomas Burghardt (0009-0003-7874-9134) * Till Nestler (0009-0001-2922-525X) Licence Creative Commons Attribution 4.0 International Use cases - Handwritten Text Recognition (HTR) - OCR training - Computer vision / handwriting analysis Acknowledgements We acknowledge the financial support by the Federal Ministry of Education and Research of Germany and by Sächsische Staatsministerium für Wissenschaft, Kultur und Tourismus in the programme Center of Excellence for AI-research „Center for Scalable Data Analytics and Artificial Intelligence Dresden/Leipzig“, project identification number: ScaDS.AI