# Handwriting Recognition of Historical Documents with few labeled data

arXiv: https://arxiv.org/abs/1811.07768 · HTML: https://mlsanity.com/paper/1811.07768/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 10 Nov 2018.

## Built on by (most cited first, 10 of 10)

- [How to Choose Pretrained Handwriting Recognition Models for Single Writer Fine-Tuning](https://mlsanity.com/paper/2305.02593/) (2023)
- [Benchmarking Large Language Models for Handwritten Text Recognition](https://mlsanity.com/paper/2503.15195/) (2025)
- [Fine-tuning Handwriting Recognition systems with Temporal Dropout](https://mlsanity.com/paper/2102.00511/) (2021)
- [StackMix and Blot Augmentations for Handwritten Text Recognition](https://mlsanity.com/paper/2108.11667/) (2021)
- [Quo Vadis Handwritten Text Generation for Handwritten Text Recognition?](https://mlsanity.com/paper/2508.09936/) (2025)
- [Digitizing Nepal's Written Heritage: A Comprehensive HTR Pipeline for Old Nepali Manuscripts](https://mlsanity.com/paper/2512.17111/) (2025)
- [Handwritten text generation and strikethrough characters augmentation](https://mlsanity.com/paper/2112.07395/) (2021)
- [Handwritten Text Recognition of Historical Manuscripts Using Transformer-Based Models](https://mlsanity.com/paper/2508.11499/) (2025)
- [Automated sign detection across the Electronic Babylonian Library: A large-scale dataset and end-to-end cuneiform OCR pipeline](https://mlsanity.com/paper/2606.22608/) (2026)
- [Cross-codex Learning for Reliable Scribe Identification in Medieval Manuscripts](https://mlsanity.com/paper/2312.04296/) (2023)

## Abstract (arXiv)

Historical documents present many challenges for offline handwriting recognition systems, among them, the segmentation and labeling steps. Carefully annotated textlines are needed to train an HTR system. In some scenarios, transcripts are only available at the paragraph level with no text-line information. In this work, we demonstrate how to train an HTR system with few labeled data. Specifically, we train a deep convolutional recurrent neural network (CRNN) system on only 10% of manually labeled text-line data from a dataset and propose an incremental training procedure that covers the rest of the data. Performance is further increased by augmenting the training set with specially crafted multiscale data. We also propose a model-based normalization scheme which considers the variability in the writing scale at the recognition phase. We apply this approach to the publicly available READ dataset. Our system achieved the second best result during the ICDAR2017 competition.