Fonts-2-Handwriting: A Seed-Augment-Train framework for universal digit classification

TL;DR

This paper introduces a novel framework that generates synthetic training data from open font datasets to train neural networks for universal handwritten digit classification across multiple Indic scripts, demonstrating effective transfer learning.

Contribution

The paper presents a Seed-Augment-Train/Transfer framework that creates synthetic datasets from font files for training models on handwritten digits in diverse scripts, enabling universal digit recognition.

Findings

CNN trained on synthetic data performs well on real handwritten digits

GAN generates realistic digit images in five Indic scripts

Framework bridges font datasets and transfer learning for digit classification

Abstract

In this paper, we propose a Seed-Augment-Train/Transfer (SAT) framework that contains a synthetic seed image dataset generation procedure for languages with different numeral systems using freely available open font file datasets. This seed dataset of images is then augmented to create a purely synthetic training dataset, which is in turn used to train a deep neural network and test on held-out real world handwritten digits dataset spanning five Indic scripts, Kannada, Tamil, Gujarati, Malayalam, and Devanagari. We showcase the efficacy of this approach both qualitatively, by training a Boundary-seeking GAN (BGAN) that generates realistic digit images in the five languages, and also quantitatively by testing a CNN trained on the synthetic data on the real-world datasets. This establishes not only an interesting nexus between the font-datasets-world and transfer learning but also provides a recipe for universal-digit classification in any script.