STELA: A Real-Time Scene Text Detector With Learned Anchor

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 two-stage 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 one-stage 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 (<inline-formula> <tex-math notation="LaTeX">$26.5fps$ </tex-math></inline-formula> at <inline-formula> <tex-math notation="LaTeX">$800p$ </tex-math></inline-formula>), 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 is publicly available at <uri>https://github.com/xhzdeng/stela</uri>.