One-shot Compositional Data Generation for Low Resource Handwritten Text Recognition
This addresses the challenge of scarce annotated data for HTR, particularly in domains like historical ciphered manuscripts, though it is incremental as it builds on existing BPL methods.
The paper tackles the problem of low resource Handwritten Text Recognition (HTR) by proposing a data generation technique based on Bayesian Program Learning (BPL) that uses only one sample per symbol to create synthetic training data, achieving effective results as confirmed by quantitative and qualitative analyses.
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.