IOS-Net: An Inside-to-outside Supervision Network for Scale Robust Text Detection in the wild

•We propose an inside-to-outside supervision network to detect text in natural scenes.•Our method can comprehensively tackle the R-scale and S-scale problems.•Our method achieves a state-of-the-art performance on three public benchmarks.•Our method is the best text detector with respect to the speed and F-score trade-off.