FastTextSpotter: A High-Efficiency Transformer for Multilingual Scene Text Spotting

TL;DR

FastTextSpotter introduces a high-efficiency transformer-based framework for multilingual scene text spotting, achieving superior accuracy and speed across diverse datasets while maintaining robustness in recognizing complex text shapes.

Contribution

It presents a novel, faster self-attention unit SAC2 integrated into a transformer architecture, significantly improving efficiency without sacrificing accuracy.

Findings

Achieves state-of-the-art accuracy on multiple scene text datasets.

Demonstrates improved processing speed over existing models.

Effectively recognizes multilingual and arbitrarily-shaped texts.

Abstract

The proliferation of scene text in both structured and unstructured environments presents significant challenges in optical character recognition (OCR), necessitating more efficient and robust text spotting solutions. This paper presents FastTextSpotter, a framework that integrates a Swin Transformer visual backbone with a Transformer Encoder-Decoder architecture, enhanced by a novel, faster self-attention unit, SAC2, to improve processing speeds while maintaining accuracy. FastTextSpotter has been validated across multiple datasets, including ICDAR2015 for regular texts and CTW1500 and TotalText for arbitrary-shaped texts, benchmarking against current state-of-the-art models. Our results indicate that FastTextSpotter not only achieves superior accuracy in detecting and recognizing multilingual scene text (English and Vietnamese) but also improves model efficiency, thereby setting new benchmarks in the field. This study underscores the potential of advanced transformer architectures in improving the adaptability and speed of text spotting applications in diverse real-world settings. The dataset, code, and pre-trained models have been released in our Github.