One-shot Compositional Data Generation for Low Resource Handwritten Text Recognition

TL;DR

This paper introduces a Bayesian Program Learning-based data generation method that creates synthetic handwritten symbols from just one sample per symbol, significantly aiding low-resource handwritten text recognition.

Contribution

The paper presents a novel one-shot data generation approach for handwritten symbols using Bayesian Program Learning, improving training data scarcity issues in low-resource HTR tasks.

Findings

Generated synthetic data improves HTR model performance.

Method effectively creates human-like handwriting from minimal samples.

Quantitative and qualitative analyses validate the approach.

Abstract

Low resource Handwritten Text Recognition (HTR) is a hard problem due to the scarce annotated data and the very limited linguistic information (dictionaries and language models). For example, in the case of historical ciphered manuscripts, which are usually written with invented alphabets to hide the message contents. Thus, in this paper we address this problem through a data generation technique based on Bayesian Program Learning (BPL). Contrary to traditional generation approaches, which require a huge amount of annotated images, our method is able to generate human-like handwriting using only one sample of each symbol in the alphabet. After generating symbols, we create synthetic lines to train state-of-the-art HTR architectures in a segmentation free fashion. Quantitative and qualitative analyses were carried out and confirm the effectiveness of the proposed method.