TextBoxes: A Fast Text Detector with a Single Deep Neural Network

TL;DR

TextBoxes introduces a fast, end-to-end deep neural network for scene text detection that achieves high accuracy and efficiency, outperforming existing methods in speed and localization precision, and enhancing end-to-end recognition tasks.

Contribution

The paper proposes a novel single-network architecture for scene text detection that is both fast and accurate, with minimal post-processing.

Findings

Achieves 0.09s per image detection speed

Outperforms state-of-the-art in text localization accuracy

Significantly improves end-to-end text recognition results

Abstract

This paper presents an end-to-end trainable fast scene text detector, named TextBoxes, which detects scene text with both high accuracy and efficiency in a single network forward pass, involving no post-process except for a standard non-maximum suppression. TextBoxes outperforms competing methods in terms of text localization accuracy and is much faster, taking only 0.09s per image in a fast implementation. Furthermore, combined with a text recognizer, TextBoxes significantly outperforms state-of-the-art approaches on word spotting and end-to-end text recognition tasks.