# On the Accuracy of CRNNs for Line-Based OCR: A Multi-Parameter Evaluation

arXiv: https://arxiv.org/abs/2008.02777 · HTML: https://mlsanity.com/paper/2008.02777/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 6 Aug 2020.

## Built on by (most cited first, 2 of 2)

- [CHURRO: Making History Readable with an Open-Weight Large Vision-Language Model for High-Accuracy, Low-Cost Historical Text Recognition](https://mlsanity.com/paper/2509.19768/) (2025)
- [An End-to-End Khmer Optical Character Recognition using Sequence-to-Sequence with Attention](https://mlsanity.com/paper/2106.10875/) (2021)

## Abstract (arXiv)

We investigate how to train a high quality optical character recognition (OCR) model for difficult historical typefaces on degraded paper. Through extensive grid searches, we obtain a neural network architecture and a set of optimal data augmentation settings. We discuss the influence of factors such as binarization, input line height, network width, network depth, and other network training parameters such as dropout. Implementing these findings into a practical model, we are able to obtain a 0.44% character error rate (CER) model from only 10,000 lines of training data, outperforming currently available pretrained models that were trained on more than 20 times the amount of data. We show ablations for all components of our training pipeline, which relies on the open source framework Calamari.