Segmentation model for two-column printed reports trained from Reichstagsprotokolle (1920–1932)

Read full text Summarise with VIRAT
Authors: Jan Kamlah, Benjamin Rohr
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Historical Era and source material: The training material is derived from the official parliamentary protocols of the German Reichstag (the national parliament), spanning the entire period of the Weimar Republic (1920–1932). The protocols were obtained from https://www.reichstagsprotokolle.de/. Layout : Historical two-column printed reports.From scratch training (16 Epochs): A training set was created consisting of 60 pages of manually corrected ground truth data drawn from the corpus.These models were used to create a new corpus of all speeches delivered in the German Reichstag during the Weimar Republic (1920-1932).

Fields of Study & Subjects

Handwritten Text Recognition Techniques Digital Humanities and Scholarship Topic Modeling

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
25 March 2026
Language
EN