Multi-Oriented Text Detection with Fully Convolutional Networks

TL;DR

This paper introduces a fully convolutional network-based method for detecting multi-oriented text in natural images, leveraging local and global cues for accurate localization across various languages and fonts.

Contribution

It presents a novel multi-stage FCN framework that combines salient maps, character centroid prediction, and hypothesis filtering for robust multi-oriented text detection.

Findings

Achieves state-of-the-art results on MSRA-TD500, ICDAR2015, and ICDAR2013 datasets.

Effectively handles multiple orientations, languages, and fonts.

Demonstrates high accuracy and robustness in natural scene text detection.

Abstract

In this paper, we propose a novel approach for text detec- tion in natural images. Both local and global cues are taken into account for localizing text lines in a coarse-to-fine pro- cedure. First, a Fully Convolutional Network (FCN) model is trained to predict the salient map of text regions in a holistic manner. Then, text line hypotheses are estimated by combining the salient map and character components. Fi- nally, another FCN classifier is used to predict the centroid of each character, in order to remove the false hypotheses. The framework is general for handling text in multiple ori- entations, languages and fonts. The proposed method con- sistently achieves the state-of-the-art performance on three text detection benchmarks: MSRA-TD500, ICDAR2015 and ICDAR2013.