Progressive Scene Text Erasing with Self-Supervision

TL;DR

This paper introduces a self-supervised, progressive scene text erasing method that enhances generalization and achieves state-of-the-art results by progressively removing text using intermediate results.

Contribution

It proposes a novel self-supervised pretext task and a progressive erasing network that improves scene text removal on real-world images, surpassing previous methods.

Findings

Significant improvement in generalization to real-world data

Achieves state-of-the-art performance on public benchmarks

Effective progressive text erasing using intermediate results

Abstract

Scene text erasing seeks to erase text contents from scene images and current state-of-the-art text erasing models are trained on large-scale synthetic data. Although data synthetic engines can provide vast amounts of annotated training samples, there are differences between synthetic and real-world data. In this paper, we employ self-supervision for feature representation on unlabeled real-world scene text images. A novel pretext task is designed to keep consistent among text stroke masks of image variants. We design the Progressive Erasing Network in order to remove residual texts. The scene text is erased progressively by leveraging the intermediate generated results which provide the foundation for subsequent higher quality results. Experiments show that our method significantly improves the generalization of the text erasing task and achieves state-of-the-art performance on public benchmarks.