SynthTIGER: Synthetic Text Image GEneratoR Towards Better Text Recognition Models

TL;DR

SynthTIGER is a new synthetic text image generator that improves scene text recognition models by producing more diverse and balanced training data, outperforming existing synthetic datasets.

Contribution

It introduces SynthTIGER, a comprehensive synthetic text image generator that integrates multiple synthesis techniques and addresses data imbalance issues.

Findings

SynthTIGER outperforms combined existing synthetic datasets in STR tasks.

Ablation studies confirm the effectiveness of SynthTIGER's components.

Guidelines for synthetic image generation improve STR model training.

Abstract

For successful scene text recognition (STR) models, synthetic text image generators have alleviated the lack of annotated text images from the real world. Specifically, they generate multiple text images with diverse backgrounds, font styles, and text shapes and enable STR models to learn visual patterns that might not be accessible from manually annotated data. In this paper, we introduce a new synthetic text image generator, SynthTIGER, by analyzing techniques used for text image synthesis and integrating effective ones under a single algorithm. Moreover, we propose two techniques that alleviate the long-tail problem in length and character distributions of training data. In our experiments, SynthTIGER achieves better STR performance than the combination of synthetic datasets, MJSynth (MJ) and SynthText (ST). Our ablation study demonstrates the benefits of using sub-components of SynthTIGER and the guideline on generating synthetic text images for STR models. Our implementation is publicly available at https://github.com/clovaai/synthtiger.