# TextCaps : Handwritten Character Recognition with Very Small Datasets

arXiv: https://arxiv.org/abs/1904.08095 · HTML: https://mlsanity.com/paper/1904.08095/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 17 Apr 2019.

## Built on by (most cited first, 9 of 9)

- [Boosting Modern and Historical Handwritten Text Recognition with Deformable Convolutions](https://mlsanity.com/paper/2208.08109/) (2022)
- [Learning from One and Only One Shot](https://mlsanity.com/paper/2201.08815/) (2022)
- [Handwritten Arabic Character Recognition for Children Writ-ing Using Convolutional Neural Network and Stroke Identification](https://mlsanity.com/paper/2211.02119/) (2022)
- [Kernel function impact on convolutional neural networks](https://mlsanity.com/paper/2302.10266/) (2023)
- [Signature-Graph Networks](https://mlsanity.com/paper/2110.11551/) (2021)
- [Intelligent Character Recognition of Handwritten Forms with Deep Neural Networks](https://mlsanity.com/paper/2606.08858/) (2026)
- [An Improved Nearest Neighbour Classifier](https://mlsanity.com/paper/2204.13141/) (2022)
- [Local Approximations, Real Interpolation and Machine Learning](https://mlsanity.com/paper/2207.07720/) (2022)
- [Rethinking Handwritten Character Recognition](https://mlsanity.com/paper/2609.10572/) (2026)

## Abstract (arXiv)

Many localized languages struggle to reap the benefits of recent advancements in character recognition systems due to the lack of substantial amount of labeled training data. This is due to the difficulty in generating large amounts of labeled data for such languages and inability of deep learning techniques to properly learn from small number of training samples. We solve this problem by introducing a technique of generating new training samples from the existing samples, with realistic augmentations which reflect actual variations that are present in human hand writing, by adding random controlled noise to their corresponding instantiation parameters. Our results with a mere 200 training samples per class surpass existing character recognition results in the EMNIST-letter dataset while achieving the existing results in the three datasets: EMNIST-balanced, EMNIST-digits, and MNIST. We also develop a strategy to effectively use a combination of loss functions to improve reconstructions. Our system is useful in character recognition for localized languages that lack much labeled training data and even in other related more general contexts such as object recognition.