# EnsNet: Ensconce Text in the Wild

arXiv: https://arxiv.org/abs/1812.00723 · HTML: https://mlsanity.com/paper/1812.00723/

## Key facts

- Cited by 27 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 3 Dec 2018.

## Built on by (most cited first, 27 of 27)

- [De-rendering Stylized Texts](https://mlsanity.com/paper/2110.01890/) (2021)
- [Stroke-Based Scene Text Erasing Using Synthetic Data for Training](https://mlsanity.com/paper/2104.11493/) (2021)
- [Visual Text Processing: A Comprehensive Review and Unified Evaluation](https://mlsanity.com/paper/2504.21682/) (2025)
- [Don't Forget Me: Accurate Background Recovery for Text Removal via Modeling Local-Global Context](https://mlsanity.com/paper/2207.10273/) (2022)
- [PERT: A Progressively Region-based Network for Scene Text Removal](https://mlsanity.com/paper/2106.13029/) (2021)
- [The Surprisingly Straightforward Scene Text Removal Method With Gated Attention and Region of Interest Generation: A Comprehensive Prominent Model Analysis](https://mlsanity.com/paper/2210.07489/) (2022)
- [FETNet: Feature Erasing and Transferring Network for Scene Text Removal](https://mlsanity.com/paper/2306.09593/) (2023)
- [ViTEraser: Harnessing the Power of Vision Transformers for Scene Text Removal with SegMIM Pretraining](https://mlsanity.com/paper/2306.12106/) (2023)
- [UPOCR: Towards Unified Pixel-Level OCR Interface](https://mlsanity.com/paper/2312.02694/) (2023)
- [Self-Supervised Text Erasing with Controllable Image Synthesis](https://mlsanity.com/paper/2204.12743/) (2022)
- [Progressive Scene Text Erasing with Self-Supervision](https://mlsanity.com/paper/2207.11469/) (2022)
- [PSSTRNet: Progressive Segmentation-guided Scene Text Removal Network](https://mlsanity.com/paper/2306.07842/) (2023)
- [Visual Text Meets Low-level Vision: A Comprehensive Survey on Visual Text Processing](https://mlsanity.com/paper/2402.03082/) (2024)
- [On Manipulating Scene Text in the Wild with Diffusion Models](https://mlsanity.com/paper/2311.00734/) (2023)
- [DeepEraser: Deep Iterative Context Mining for Generic Text Eraser](https://mlsanity.com/paper/2402.19108/) (2024)
- [DiffSTR: Controlled Diffusion Models for Scene Text Removal](https://mlsanity.com/paper/2410.21721/) (2024)
- [ConText: Driving In-context Learning for Text Removal and Segmentation](https://mlsanity.com/paper/2506.03799/) (2025)
- [TPFNet: A Novel Text In-painting Transformer for Text Removal](https://mlsanity.com/paper/2210.14461/) (2022)
- [Selective Scene Text Removal](https://mlsanity.com/paper/2309.00410/) (2023)
- [StrokeNet: Stroke Assisted and Hierarchical Graph Reasoning Networks](https://mlsanity.com/paper/2111.11718/) (2021)
- [Leveraging Text Localization for Scene Text Removal via Text-aware Masked Image Modeling](https://mlsanity.com/paper/2409.13431/) (2024)
- [MSLKANet: A Multi-Scale Large Kernel Attention Network for Scene Text Removal](https://mlsanity.com/paper/2211.06565/) (2022)
- [Text Image Generation for Low-Resource Languages with Dual Translation Learning](https://mlsanity.com/paper/2409.17747/) (2024)
- [TextDestroyer: A Training- and Annotation-Free Diffusion Method for Destroying Anomal Text from Images](https://mlsanity.com/paper/2411.00355/) (2024)
- [Inverse Scene Text Removal](https://mlsanity.com/paper/2506.21002/) (2025)
- [OTR: Synthesizing Overlay Text Dataset for Text Removal](https://mlsanity.com/paper/2510.02787/) (2025)
- [What Shape Is Optimal for Masks in Text Removal?](https://mlsanity.com/paper/2511.22499/) (2025)

## Abstract (arXiv)

A new method is proposed for removing text from natural images. The challenge is to first accurately localize text on the stroke-level and then replace it with a visually plausible background. Unlike previous methods that require image patches to erase scene text, our method, namely ensconce network (EnsNet), can operate end-to-end on a single image without any prior knowledge. The overall structure is an end-to-end trainable FCN-ResNet-18 network with a conditional generative adversarial network (cGAN). The feature of the former is first enhanced by a novel lateral connection structure and then refined by four carefully designed losses: multiscale regression loss and content loss, which capture the global discrepancy of different level features; texture loss and total variation loss, which primarily target filling the text region and preserving the reality of the background. The latter is a novel local-sensitive GAN, which attentively assesses the local consistency of the text erased regions. Both qualitative and quantitative sensitivity experiments on synthetic images and the ICDAR 2013 dataset demonstrate that each component of the EnsNet is essential to achieve a good performance. Moreover, our EnsNet can significantly outperform previous state-of-the-art methods in terms of all metrics. In addition, a qualitative experiment conducted on the SMBNet dataset further demonstrates that the proposed method can also preform well on general object (such as pedestrians) removal tasks. EnsNet is extremely fast, which can preform at 333 fps on an i5-8600 CPU device.