Published June 7, 2021 | Version 1
Dataset Open

Recognition results for the Handwritten Text Recognition Test Set: Minutes of the Swiss Federal Council (1848-1903)

  • 1. Digital Humanities/Walter Benjamin Kolleg/Universität Bern

Description

This data set reports the results of different handwritten text recognition engines on the test set "Minutes of the Swiss Federal Council (1848-1903)". The test set with correct transcriptions is available under the following DOI: 10.5281/zenodo.4746341.

The following models have been applied (available for download as zipped folders) and report their result per line in page XML:

- German Kurrent M2, engine "HTR+", URL: https://readcoop.eu/model/german-kurrent-19th-century/

- German Kurrent M2, engine "Pylaia", URL: https://readcoop.eu/model/german-kurrent/

- Transkribus German Kurrent M2, engine "HTR+", URL: https://readcoop.eu/model/german-kurrent-and-sutterlin-17th-20th-century/

- RRB, engine "HTR+", no URL available.

All models are available within the text recognition software Transkribus (https://readcoop.eu/transkribus/).

The images are also part of the data set. Images and page XML are connected by an identical filename.

Files

german_kurrent-m2_htr+.zip

Files (237.7 MB)

Name Size Download all
md5:6222c84849a2a2ef711ce9ba4ef68b3f
5.6 MB Preview Download
md5:4347ace94a87e417fac9db7360fba241
5.0 MB Preview Download
md5:1a4afbd2b1e2f3ac1c3ef66814ac64ed
216.9 MB Preview Download
md5:cca0463a8bfb190ad1dec87447cb2b11
5.1 MB Preview Download
md5:94396a00fe5b4a21e2c9b41537ddd676
5.2 MB Preview Download

Additional details

Related works

References
Dataset: 10.5281/zenodo.4746341 (DOI)