6000 ground truth of VOC and notarial deeds 3.000.000 HTR of VOC, WIC and notarial deeds
The National Archives of the Netherlands and Noord-Hollands Archief conducted a project using the Transkribus HTR (Handwritten Text Recognition) platform. The aim was to semi automatically transcribe 2 million pages of old Dutch texts. The transcribed archives are 17th and 18th century documents from the Dutch East-Asia Company (VOC). And 19th century notarial deeds from Noord-Hollands Archief and other archives in the provinces. In order to train the HTR software a team produced transcriptions of approximately 6000 scans. The scans are randomly selected from the dataset. With the transcriptio... Mehr ...
Verfasser: | |
---|---|
Dokumenttyp: | other |
Erscheinungsdatum: | 2020 |
Schlagwörter: | Transciptions / Verenigde Oost-Indische Compagnie / West-Indische Compagnie / Notarial deeds / Nationaal Archief / Noord-Hollands Archief / Transkribus |
Sprache: | unknown |
Permalink: | https://search.fid-benelux.de/Record/base-26728466 |
Datenquelle: | BASE; Originalkatalog |
Powered By: | BASE |
Link(s) : | https://zenodo.org/record/4159268 |
The National Archives of the Netherlands and Noord-Hollands Archief conducted a project using the Transkribus HTR (Handwritten Text Recognition) platform. The aim was to semi automatically transcribe 2 million pages of old Dutch texts. The transcribed archives are 17th and 18th century documents from the Dutch East-Asia Company (VOC). And 19th century notarial deeds from Noord-Hollands Archief and other archives in the provinces. In order to train the HTR software a team produced transcriptions of approximately 6000 scans. The scans are randomly selected from the dataset. With the transcriptions a model is trained that can recognize more than 90% of the characters correctly. Transkribus transcribed the 2 million scans automatically using the trained model. The following Transkribus HTR+ model has been trained for the text recognition: "IJsberg". More information about the model can be found here. See the chapter "Dutch Handwriting". However, the Transkribus team retrained the model with PyLaia technology, which improved the HTR+ model. This PyLaia model is not publicly available. Later on, 1 million extra scans concerning the West India Company (WIC) were transcribed automatically without adding extra ground truth or training. These archives are from the 17th and 18th century. The datasets published in Zenodo contain the ground truth (scans in JPG, transcription in PAGE XML) and the HTR results (in PAGE XML and TXT). See the overview below. Scroll to the bottom of the page to download the actual files. Disclaimer: due to the languages (English, French and German) used in the archive of "1.05.21, Dutch series Guyana" and the bad state of the archive itself (waterdamage, discolouration and ink damage) HTR results are often of poor quality and not usable for research. --- Dataset HTR Dataset, name archive, number archive, inventory numbers, link to inventory) HTR results 1.05.01.01, Oude WIC, 1.05.01.01, 1-87, EAD HTR results 1.05.01.02, Tweede WIC, 1.05.01.02, 1-1382, EAD HTR results 1.05.02, Raad der Koloniën, ...