All You Need Is Boundary: Toward Arbitrary-Shaped Text Spotting

TL;DR

This paper introduces a boundary-based method for end-to-end text spotting that effectively detects and recognizes arbitrarily shaped text in cluttered images, outperforming previous approaches.

Contribution

It proposes a novel boundary point localization scheme for text detection, enabling accurate recognition of arbitrary-shaped text in a unified framework.

Findings

Outperforms state-of-the-art on ICDAR2015, TotalText, COCO-Text datasets.

Effective for detecting and recognizing arbitrarily shaped text.

Demonstrates robustness in cluttered scene images.

Abstract

Recently, end-to-end text spotting that aims to detect and recognize text from cluttered images simultaneously has received particularly growing interest in computer vision. Different from the existing approaches that formulate text detection as bounding box extraction or instance segmentation, we localize a set of points on the boundary of each text instance. With the representation of such boundary points, we establish a simple yet effective scheme for end-to-end text spotting, which can read the text of arbitrary shapes. Experiments on three challenging datasets, including ICDAR2015, TotalText and COCO-Text demonstrate that the proposed method consistently surpasses the state-of-the-art in both scene text detection and end-to-end text recognition tasks.