STELA: A Real-Time Scene Text Detector with Learned Anchor
arXiv:1909.07549
Abstract
To achieve high coverage of target boxes, a normal strategy of conventional one-stage anchor-based detectors is to utilize multiple priors at each spatial position, especially in scene text detection tasks. In this work, we present a simple and intuitive method for multi-oriented text detection where each location of feature maps only associates with one reference box. The idea is inspired from the twostage R-CNN framework that can estimate the location of objects with any shape by using learned proposals. The aim of our method is to integrate this mechanism into a onestage detector and employ the learned anchor which is obtained through a regression operation to replace the original one into the final predictions. Based on RetinaNet, our method achieves competitive performances on several public benchmarks with a totally real-time efficiency (26:5fps at 800p), which surpasses all of anchor-based scene text detectors. In addition, with less attention on anchor design, we believe our method is easy to be applied on other analogous detection tasks. The code will publicly available at https://github.com/xhzdeng/stela.
References in corpus (9)
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
- R2CNN: Rotational Region CNN for Orientation Robust Scene Text Detection
- Synthetic Data for Text Localisation in Natural Images
- EAST: An Efficient and Accurate Scene Text Detector
- DeepText: A Unified Framework for Text Proposal Generation and Text Detection in Natural Images
- Deep Matching Prior Network: Toward Tighter Multi-oriented Text Detection
- Single Shot Text Detector with Regional Attention
- Multi-Oriented Scene Text Detection via Corner Localization and Region Segmentation
- Consistent Optimization for Single-Shot Object Detection