Shilin-LU/W-Bench
[ICLR 2025] Robust Watermarking Using Generative Priors Against Image Editing: From Benchmarking to Advances What is it? W-Bench is the first benchmark to evaluate watermarking robustness across four types of image editing techniques, including regeneration, global editing, local editing, and image-to-video generation. Eleven representative watermarking methods are evaluated on the W-Bench. The W-Bench contains 10,000 instances sourced from datasets such as COCO… See the full description on the dataset page: https://huggingface.co/datasets/Shilin-LU/W-Bench.
| Author / organization | Shilin-LU |
|---|---|
| License (as declared) | mit |
| Downloads | 1.6K (as reported by Hugging Face) |
| Likes | 2 |
| Created | 22 February 2025 |
| Last updated at source | 2 March 2025 |
Popularity, last 3 days
1,615 now (+45 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Recent changes
- 16 Oct 2026 Description: [ICLR 2025] Robust Watermarking Using Generative Priors Against Image Editing: From Benchmarking to Advances What is it? W-Bench is the first benchmark to evaluate watermarking robustness across four types of image editing techniques, inclu → [ICLR 2025] Robust Watermarking Using Generative Priors Against Image Editing: From Benchmarking to Advances What is it? W-Bench is the first benchmark to evaluate watermarking robustness across four types of image editing techniques, inclu
Changes detected by Silytics between refreshes of the source.
More from Shilin-LU
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.