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

Mohamed Ali Souibgui, Ali Furkan Biten, Sounak Dey, Alicia Fornés, Yousri Kessentini, Lluís Gómez, Dimosthenis Karatzas, Josep Lladós; Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2022, pp. 935-943

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.

Related Material


[pdf] [supp] [arXiv]
[bibtex]
@InProceedings{Souibgui_2022_WACV, author = {Souibgui, Mohamed Ali and Biten, Ali Furkan and Dey, Sounak and Forn\'es, Alicia and Kessentini, Yousri and G\'omez, Llu{\'\i}s and Karatzas, Dimosthenis and Llad\'os, Josep}, title = {One-Shot Compositional Data Generation for Low Resource Handwritten Text Recognition}, booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)}, month = {January}, year = {2022}, pages = {935-943} }