MTRNet++: One-stage Mask-based Scene Text Eraser
arXiv:1912.07183 · doi:10.1016/j.cviu.2020.103066
Abstract
A precise, controllable, interpretable and easily trainable text removal approach is necessary for both user-specific and large-scale text removal applications. To achieve this, we propose a one-stage mask-based text inpainting network, MTRNet++. It has a novel architecture that includes mask-refine, coarse-inpainting and fine-inpainting branches, and attention blocks. With this architecture, MTRNet++ can remove text either with or without an external mask. It achieves state-of-the-art results on both the Oxford and SCUT datasets without using external ground-truth masks. The results of ablation studies demonstrate that the proposed multi-branch architecture with attention blocks is effective and essential. It also demonstrates controllability and interpretability.
arXiv commentThis paper is under CVIU review (after major revision)
References in corpus (7)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Spectral Normalization for Generative Adversarial Networks
- EdgeConnect: Generative Image Inpainting with Adversarial Edge Learning
- Image Restoration Using Convolutional Auto-encoders with Symmetric Skip Connections
- Generative Image Inpainting with Contextual Attention
- MTRNet: A Generic Scene Text Eraser
- EnsNet: Ensconce Text in the Wild