Bullinger Project

Bullinger Project Automatic Handwriting Recognition for 16th Century Letters
  • HES-SO
  • Dr. Anna Scius-Bertrand
  • Prof. Andreas Fischer
  • Document Analysis
  • Historical Documents
  • Handwriting Recognition
  • Document Analysis
  • Writer Adaptation
Our skills
Document Analysis, Deep Learning, Machine Learning, Computer Vision
Digital edition of the letters available to the general public
Our Partners:

  • Prof. Martin Volk, University of Zürich
  • Prof. Tobias Hodel, University of Bern
This project is funded by the Hasler Foundation
Heinrich Bullinger (1504-1575) was an influential Swiss Reformer, whose comprehensive letter collection is being made available via an electronic online edition to historians and the general public alike. Automatic handwriting recognition is needed to make contents of the letters searchable and browsable. In this project, we develop novel methods for handwriting recognition in historical documents to improve recognition accuracy, especially for the challenging case of non-frequent writers with only few letters in the dataset.