Nuremberg Letterbooks: A Multi-Transcriptional Dataset of Early 15th Century Manuscripts for Document Analysis
This dataset addresses a gap for humanities researchers by providing multi-transcriptional historical documents, though it is incremental as it builds on existing document analysis methods.
The authors tackled the lack of datasets with diverse transcriptions for humanities research by introducing the Nuremberg Letterbooks dataset, which includes 4 books with 1711 labeled pages and multiple transcription types, establishing baselines for tasks like handwritten text recognition and writer identification.
Most datasets in the field of document analysis utilize highly standardized labels, which, while simplifying specific tasks, often produce outputs that are not directly applicable to humanities research. In contrast, the Nuremberg Letterbooks dataset, which comprises historical documents from the early 15th century, addresses this gap by providing multiple types of transcriptions and accompanying metadata. This approach allows for developing methods that are more closely aligned with the needs of the humanities. The dataset includes 4 books containing 1711 labeled pages written by 10 scribes. Three types of transcriptions are provided for handwritten text recognition: Basic, diplomatic, and regularized. For the latter two, versions with and without expanded abbreviations are also available. A combination of letter ID and writer ID supports writer identification due to changing writers within pages. In the technical validation, we established baselines for various tasks, demonstrating data consistency and providing benchmarks for future research to build upon.