# Self-Training of Handwritten Word Recognition for Synthetic-to-Real Adaptation

arXiv: https://arxiv.org/abs/2206.03149 · HTML: https://mlsanity.com/paper/2206.03149/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 41 references that are in the map.
- First posted on arXiv 7 Jun 2022.

## Builds on (most cited first, 18 of 18)

- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Domain-Adversarial Training of Neural Networks](https://mlsanity.com/paper/1505.07818/) (2015)
- [FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence](https://mlsanity.com/paper/2001.07685/) (2020)
- [MixMatch: A Holistic Approach to Semi-Supervised Learning](https://mlsanity.com/paper/1905.02249/) (2019)
- [Adversarial Discriminative Domain Adaptation](https://mlsanity.com/paper/1702.05464/) (2017)
- [What Is Wrong With Scene Text Recognition Model Comparisons? Dataset and Model Analysis](https://mlsanity.com/paper/1904.01906/) (2019)
- [GANwriting: Content-Conditioned Generation of Styled Handwritten Word Images](https://mlsanity.com/paper/2003.02567/) (2020)
- [Scan, Attend and Read: End-to-End Handwritten Paragraph Recognition with MDLSTM Attention](https://mlsanity.com/paper/1604.03286/) (2016)
- [Adversarial Generation of Handwritten Text Images Conditioned on Sequences](https://mlsanity.com/paper/1903.00277/) (2019)
- [Zero-Shot Deep Domain Adaptation](https://mlsanity.com/paper/1707.01922/) (2017)
- [SmartPatch: Improving Handwritten Word Imitation with Patch Discriminators](https://mlsanity.com/paper/2105.10528/) (2021)
- [Asking questions on handwritten document collections](https://mlsanity.com/paper/2110.00711/) (2021)
- [HWNet v2: An Efficient Word Image Representation for Handwritten Documents](https://mlsanity.com/paper/1802.06194/) (2018)
- [AT-ST: Self-Training Adaptation Strategy for OCR in Domains with Limited Transcriptions](https://mlsanity.com/paper/2104.13037/) (2021)
- [Implicit Language Model in LSTM for OCR](https://mlsanity.com/paper/1805.09441/) (2018)
- [Learning Deep Representations for Word Spotting Under Weak Supervision](https://mlsanity.com/paper/1712.00250/) (2017)
- [Annotation-free Learning of Deep Representations for Word Spotting using Synthetic Data and Self Labeling](https://mlsanity.com/paper/2003.01989/) (2020)
- [Language Model Supervision for Handwriting Recognition Model Adaptation](https://mlsanity.com/paper/1808.01423/) (2018)

## Built on by (most cited first, 5 of 5)

- [Improving Handwritten OCR with Training Samples Generated by Glyph Conditional Denoising Diffusion Probabilistic Model](https://mlsanity.com/paper/2305.19543/) (2023)
- [Fine-tuning Is a Surprisingly Effective Domain Adaptation Baseline in Handwriting Recognition](https://mlsanity.com/paper/2302.06308/) (2023)
- [SoftCTC -- Semi-Supervised Learning for Text Recognition using Soft Pseudo-Labels](https://mlsanity.com/paper/2212.02135/) (2022)
- [Data Generation for Post-OCR correction of Cyrillic handwriting](https://mlsanity.com/paper/2311.15896/) (2023)
- [Optimal Transport for Handwritten Text Recognition in a Low-Resource Regime](https://mlsanity.com/paper/2509.16977/) (2025)

## Abstract (arXiv)

Performances of Handwritten Text Recognition (HTR) models are largely determined by the availability of labeled and representative training samples. However, in many application scenarios labeled samples are scarce or costly to obtain. In this work, we propose a self-training approach to train a HTR model solely on synthetic samples and unlabeled data. The proposed training scheme uses an initial model trained on synthetic data to make predictions for the unlabeled target dataset. Starting from this initial model with rather poor performance, we show that a considerable adaptation is possible by training against the predicted pseudo-labels. Moreover, the investigated self-training strategy does not require any manually annotated training samples. We evaluate the proposed method on four widely used benchmark datasets and show its effectiveness on closing the gap to a model trained in a fully-supervised manner.