# Image Quality Assessment on Diffusion-based Watermarked Images

Source: [Image Quality Assessment on Diffusion-based Watermarked Images](<https://www.sota2.com/research/sota/image-quality-assessment-on-diffusion-based-watermarked-images>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-01T08:38:22.943Z

Evaluation of image quality assessment performance using metrics like CLIP score and FID on a dataset composed of watermarked images generated by diffusion models.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Image Quality Assessment |
| Dataset (including any stated split) | Diffusion-based Watermarked Images |
| Metric | CLIP Score |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [CLIP Score](<https://www.sota2.com/research/sota/image-quality-assessment-on-diffusion-based-watermarked-images.md?metric=CLIP%20Score>): Higher is better; unit: SCALAR.

- [FID](<https://www.sota2.com/research/sota/image-quality-assessment-on-diffusion-based-watermarked-images.md?metric=FID>): Lower is better; unit: SCALAR.

## Recorded results

| Method | CLIP Score | Result date | Report paper |
| --- | --- | --- | --- |
| SHIFT | 32.194 | Not recorded | 69cca6d20636e5bf4afcbee2 |
| Black-box attack | 31.877 | Not recorded | 69cca6d20636e5bf4afcbee2 |
| Forgery attack | 31.356 | Not recorded | 69cca6d20636e5bf4afcbee2 |

### SHIFT

CLIP Score: 32.194. Result ID: 69ccd96e47f34ce924b6924c.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6d20636e5bf4afcbee2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cca6d20636e5bf4afcbee2.md>)

Recorded evaluation configuration: {"Method Group":"Our attack"}. This record does not assert independent reproduction.

### Black-box attack

CLIP Score: 31.877. Result ID: 69ccd96e47f34ce924b6924a.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6d20636e5bf4afcbee2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79301e209131cea7c83b.md>)

Recorded evaluation configuration: {"Method Group":"Black-box attack"}. This record does not assert independent reproduction.

### Forgery attack

CLIP Score: 31.356. Result ID: 69ccd96e47f34ce924b6924b.

[Reporting paper](<https://www.sota2.com/research/paper/69cca6d20636e5bf4afcbee2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78421e209131cea7775f.md>)

Recorded evaluation configuration: {"Method Group":"Forgery attack","Attack Type":"Latent-noise"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
