# Self-Training for Domain Adaptive Scene Text Detection

arXiv: https://arxiv.org/abs/2005.11487 · HTML: https://mlsanity.com/paper/2005.11487/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 May 2020.

## Built on by (most cited first, 11 of 11)

- [PIMNet: A Parallel, Iterative and Mimicking Network for Scene Text Recognition](https://mlsanity.com/paper/2109.04145/) (2021)
- [Dense Semantic Contrast for Self-Supervised Visual Representation Learning](https://mlsanity.com/paper/2109.07756/) (2021)
- [Mask is All You Need: Rethinking Mask R-CNN for Dense and Arbitrary-Shaped Scene Text Detection](https://mlsanity.com/paper/2109.03426/) (2021)
- [Bridging Synthetic and Real Worlds for Pre-training Scene Text Detectors](https://mlsanity.com/paper/2312.05286/) (2023)
- [Linguistics-aware Masked Image Modeling for Self-supervised Scene Text Recognition](https://mlsanity.com/paper/2503.18746/) (2025)
- [MMF: Multi-Task Multi-Structure Fusion for Hierarchical Image Classification](https://mlsanity.com/paper/2107.00808/) (2021)
- [Which and Where to Focus: A Simple yet Accurate Framework for Arbitrary-Shaped Nearby Text Detection in Scene Images](https://mlsanity.com/paper/2109.03451/) (2021)
- [UNITS: Unsupervised Intermediate Training Stage for Scene Text Detection](https://mlsanity.com/paper/2205.04683/) (2022)
- [Domain Adaptive Scene Text Detection via Subcategorization](https://mlsanity.com/paper/2212.00377/) (2022)
- [Class-Agnostic Region-of-Interest Matching in Document Images](https://mlsanity.com/paper/2506.21055/) (2025)
- [VimTS: A Unified Video and Image Text Spotter for Enhancing the Cross-domain Generalization](https://mlsanity.com/paper/2404.19652/) (2024)

## Abstract (arXiv)

Though deep learning based scene text detection has achieved great progress, well-trained detectors suffer from severe performance degradation for different domains. In general, a tremendous amount of data is indispensable to train the detector in the target domain. However, data collection and annotation are expensive and time-consuming. To address this problem, we propose a self-training framework to automatically mine hard examples with pseudo-labels from unannotated videos or images. To reduce the noise of hard examples, a novel text mining module is implemented based on the fusion of detection and tracking results. Then, an image-to-video generation method is designed for the tasks that videos are unavailable and only images can be used. Experimental results on standard benchmarks, including ICDAR2015, MSRA-TD500, ICDAR2017 MLT, demonstrate the effectiveness of our self-training method. The simple Mask R-CNN adapted with self-training and fine-tuned on real data can achieve comparable or even superior results with the state-of-the-art methods.