The primary overall goal of the HUGIN-MUNIN project is to develop a new deep neural network technology with general, robust models that will enable HTR to be used on a large scale as part of the standard digitisation process of Norwegian LAM collections without the need for massive and ongoing manual annotation and model training.
See the Hugin-Munin official website
As part of this project, we maintain and develop PyLaia, one of the most widely used open-source libraries for handwritten text recognition: https://gitlab.teklia.com/atr/pylaia
We have published NordHand, the largest dataset of manually transcribed handwritten documents in Norwegian :
- version 1 : https://zenodo.org/records/6542056