Nuremberg Letterbooks: A Multi-Transcriptional Dataset of Early 15th Century Manuscripts for Document Analysis

This dataset contains the images and labels of the Nuremberg Letterbooks dataset. It consists of four books (books 2 - 5) with line-wise transcriptions. Three kinds of transcriptions are reported: basic, regularized, and diplomatic, with additional expanded abbreviations. Code templates for text verification and writer verification are available at: https://github.com/M4rt1nM4yr/letterbooks_text_verification https://github.com/M4rt1nM4yr/letterbooks_writer_verification When using this dataset, please cite: M. Mayr, J. Krenz, K. Neumeier, A. Bub, S. Bürcky, N. Brolich, K. Herbers, M. Habermann, P. Fleischmann, A. Maier, and V. Christlein. Nuremberg Letterbooks: A Multi-Transcriptional Dataset of Early 15th Century Manuscripts for Document Analysis. Sci Data 12, 811 (2025).https://doi.org/10.1038/s41597-025-05144-z