Scene Text Detection with Selected Anchor

TL;DR

This paper introduces an anchor selection-based region proposal network (AS-RPN) that improves scene text detection by reducing computational costs while maintaining high recall and accuracy, outperforming dense anchor methods.

Contribution

The paper proposes learnable, effective selected anchors in AS-RPN, replacing dense anchors, leading to high recall and fewer anchors, integrated into Faster RCNN for improved scene text detection.

Findings

Achieves high recall with fewer anchors

Comparable performance to state-of-the-art on multiple benchmarks

Reduces computational costs in text detection pipeline

Abstract

Object proposal technique with dense anchoring scheme for scene text detection were applied frequently to achieve high recall. It results in the significant improvement in accuracy but waste of computational searching, regression and classification. In this paper, we propose an anchor selection-based region proposal network (AS-RPN) using effective selected anchors instead of dense anchors to extract text proposals. The center, scales, aspect ratios and orientations of anchors are learnable instead of fixing, which leads to high recall and greatly reduced numbers of anchors. By replacing the anchor-based RPN in Faster RCNN, the AS-RPN-based Faster RCNN can achieve comparable performance with previous state-of-the-art text detecting approaches on standard benchmarks, including COCO-Text, ICDAR2013, ICDAR2015 and MSRA-TD500 when using single-scale and single model (ResNet50) testing only.