metadata

library_name: PyLaia
license: mit
tags:
  - PyLaia
  - PyTorch
  - atr
  - htr
  - ocr
  - historical
  - handwritten
metrics:
  - CER
  - WER
language:
  - 'no'
pipeline_tag: image-to-text

PyLaia - NorHand v1 (post-processed)

This model performs Handwritten Text Recognition in Norwegian. It was developed during the HUGIN-MUNIN project.

Model description

The model has been trained using the PyLaia library on the NorHand document images. Line bounding boxes were improved using a post-processing step.

Training images were resized with a fixed height of 128 pixels, keeping the original aspect ratio.

Evaluation results

The model achieves the following results:

set	CER (%)	WER (%)
train	2.33	5.62
val	8.20	24.75
test	7.81	23.3

Results improve on validation and test sets when PyLaia is combined with a 6-gram language model. The language model is trained on this text corpus published by the National Library of Norway.

set	CER (%)	WER (%)
train	2.62	6.13
val	7.01	19.75
test	6.75	18.22

How to use

Please refer to the documentation.

Cite us!

@inproceedings{10.1007/978-3-031-06555-2_27,
author = {Maarand, Martin and Beyer, Yngvil and K\r{a}sen, Andre and Fosseide, Knut T. and Kermorvant, Christopher},
title = {A Comprehensive Comparison of Open-Source Libraries for Handwritten Text Recognition in Norwegian},
year = {2022},
isbn = {978-3-031-06554-5},
publisher = {Springer-Verlag},
address = {Berlin, Heidelberg},
url = {https://doi.org/10.1007/978-3-031-06555-2_27},
doi = {10.1007/978-3-031-06555-2_27},
booktitle = {Document Analysis Systems: 15th IAPR International Workshop, DAS 2022, La Rochelle, France, May 22–25, 2022, Proceedings},
pages = {399–413},
numpages = {15},
keywords = {Norwegian language, Open-source, Handwriting recognition},
location = {La Rochelle, France}
}