Update TextEraseBench dataset card

README.md

@@ -1,51 +1,54 @@
 ---
-license: mit
+license: cc-by-4.0
 pretty_name: TextEraseBench
 task_categories:
 - image-to-image
 tags:
 - text-removal
-- object-removal
+- scene-text
 - image-inpainting
-- mask-conditioned-editing
-- osor
+- benchmark
 ---
 
 # TextEraseBench
 
-TextEraseBench is a text-removal benchmark for evaluating object removal and inpainting methods on text overlays and scene text-like objects.
+[Paper](https://arxiv.org/abs/2606.28094) | [Code](https://github.com/Zhouqm-Git/osor)
 
-Dataset page: https://huggingface.co/datasets/QinmingZhou/TextEraseBench
+TextEraseBench is a paired benchmark for scene-text removal. It contains **185** samples drawn from a mixture of synthetic scenes and real-world photographs captured by the authors.
 
-## Dataset Summary
+## Construction
 
-TextEraseBench contains 185 paired samples. Each sample includes an input image containing target text, a paired clean background, and a removal mask. The benchmark is designed to test whether methods can remove text while preserving the surrounding background structure.
+1. Fine-grained text regions were manually annotated in each source image.
+2. Nano Banana 2 was used to remove the selected text and generate the paired text-free background.
+3. The resulting pairs were checked again to remove failed edits and visible artifacts.
 
-## Dataset Structure
+## Data Structure
 
 ```text
 TextEraseBench/
-├── shot/   # text-present input images
-├── bg/     # paired clean background images
-└── mask/   # target text/effect masks
+├── shot/    # Source images containing text
+├── bg/      # Paired text-free backgrounds
+└── mask/    # Fine-grained text masks
 ```
 
-For each sample id `<stem>`:
+Files with the same relative name form one evaluation sample.
 
-```text
-shot/<stem>.png
-bg/<stem>.png
-mask/<stem>.png
-```
+## Evaluation
 
-## Usage
+Use `shot/` and `mask/` as model inputs and `bg/` as the reconstruction target. Report results over the complete 185-sample set and keep preprocessing settings consistent across compared methods.
 
-Use `shot` as the input image, `mask` as the removal condition, and `bg` as the paired target for evaluation.
+## License
 
-## Dataset Creation
+TextEraseBench is released under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) to the extent that the dataset authors hold rights in the images and annotations. Some paired backgrounds were produced using the Gemini/Nano Banana service; users remain responsible for complying with any applicable service terms and for assessing their intended use.
 
-TextEraseBench is constructed through a manual-to-automated pipeline. Target text regions are manually annotated with fine-grained bounding boxes, then removed to create paired clean backgrounds. Samples undergo secondary verification to filter artifacts and semantic inconsistencies.
+## Citation
 
-## Intended Use
-
-TextEraseBench is intended for evaluating text removal, object removal, and image inpainting models on paired-background examples.
+```bibtex
+@inproceedings{zhou2026osor,
+  title     = {OSOR: One-Step Diffusion Inpainting for Effect-Aware Object Removal},
+  author    = {Zhou, Qinming and Sun, Chenxi and Kong, Deyang and He, Junhao and Tang, Xiangheng and Yu, Peike and Wu, Haotian and Cao, Leilei and Zhang, Linfeng},
+  booktitle = {European Conference on Computer Vision (ECCV)},
+  year      = {2026},
+  url       = {https://arxiv.org/abs/2606.28094}
+}
+```