# Diffusion Model Watermarking on MS-COCO and SDP Clean (test)

Source: [Diffusion Model Watermarking on MS-COCO and SDP Clean (test)](<https://www.sota2.com/research/sota/diffusion-model-watermarking-on-ms-coco-and-sdp-clean-test>)

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

Data updated: 2026-03-11T08:49:51.692Z

Evaluation of Diffusion Model Watermarking techniques using the MS-COCO and SDP Clean test set, assessing both watermark detectability (TPR, Bit Accuracy) and image quality (FID, CLIP Score).

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 | Diffusion Model Watermarking |
| Dataset (including any stated split) | MS-COCO and SDP Clean (test) |
| Metric | CLIP Score |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 10 of 10 |
| 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/diffusion-model-watermarking-on-ms-coco-and-sdp-clean-test.md?metric=CLIP%20Score>): Higher is better; unit: SCALAR.

- [TPR @ FPR=1e-6](<https://www.sota2.com/research/sota/diffusion-model-watermarking-on-ms-coco-and-sdp-clean-test.md?metric=TPR%20%40%20FPR%3D1e-6>): Higher is better; unit: PERCENT.

- [Bit Accuracy](<https://www.sota2.com/research/sota/diffusion-model-watermarking-on-ms-coco-and-sdp-clean-test.md?metric=Bit%20Accuracy>): Higher is better; unit: PERCENT.

- [Diversity](<https://www.sota2.com/research/sota/diffusion-model-watermarking-on-ms-coco-and-sdp-clean-test.md?metric=Diversity>): Higher is better; unit: SCALAR.

- [FID](<https://www.sota2.com/research/sota/diffusion-model-watermarking-on-ms-coco-and-sdp-clean-test.md?metric=FID>): Lower is better; unit: SCALAR.

## Recorded results

| Method | CLIP Score | Result date | Report paper |
| --- | --- | --- | --- |
| GS | 0.308 | Not recorded | 69b0f72b2d603e6de3e45f53 |
| StableSig | 0.3078 | Not recorded | 69b0f72b2d603e6de3e45f53 |
| T2SMark | 0.3077 | Not recorded | 69b0f72b2d603e6de3e45f53 |
| SD v2.1 | 0.3076 | Not recorded | 69b0f72b2d603e6de3e45f53 |
| PRCW | 0.3075 | Not recorded | 69b0f72b2d603e6de3e45f53 |
| dwtDctSvd | 0.3073 | Not recorded | 69b0f72b2d603e6de3e45f53 |
| RivaGAN | 0.3073 | Not recorded | 69b0f72b2d603e6de3e45f53 |
| TRW | 0.3073 | Not recorded | 69b0f72b2d603e6de3e45f53 |
| ShapeMark | 0.3073 | Not recorded | 69b0f72b2d603e6de3e45f53 |
| dwtDct | 0.3072 | Not recorded | 69b0f72b2d603e6de3e45f53 |

### GS

CLIP Score: 0.308. Result ID: 69b12caa6dbb7075b95babd8.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f72b2d603e6de3e45f53.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73131e209131cea5a4a6.md>)

Recorded evaluation configuration: {"Backbone":"SD v2.1"}. This record does not assert independent reproduction.

### StableSig

CLIP Score: 0.3078. Result ID: 69b12caa6dbb7075b95babd6.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f72b2d603e6de3e45f53.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70271e209131cea496f1.md>)

Recorded evaluation configuration: {"Backbone":"SD v2.1"}. This record does not assert independent reproduction.

### T2SMark

CLIP Score: 0.3077. Result ID: 69b12caa6dbb7075b95babda.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f72b2d603e6de3e45f53.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69004de33ef7ccd74e5e28b8.md>)

Recorded evaluation configuration: {"Backbone":"SD v2.1"}. This record does not assert independent reproduction.

### SD v2.1

CLIP Score: 0.3076. Result ID: 69b12caa6dbb7075b95babd2.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f72b2d603e6de3e45f53.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c391e209131cea32e4a.md>)

[Implementation](<https://github.com/CompVis/latent-diffusion>)

Recorded evaluation configuration: {"Backbone":"SD v2.1"}. This record does not assert independent reproduction.

### PRCW

CLIP Score: 0.3075. Result ID: 69b12caa6dbb7075b95babd9.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f72b2d603e6de3e45f53.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78351e209131cea77372.md>)

Recorded evaluation configuration: {"Backbone":"SD v2.1"}. This record does not assert independent reproduction.

### dwtDctSvd

CLIP Score: 0.3073. Result ID: 69b12caa6dbb7075b95babd4.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f72b2d603e6de3e45f53.md>)

Recorded evaluation configuration: {"Backbone":"SD v2.1"}. This record does not assert independent reproduction.

### RivaGAN

CLIP Score: 0.3073. Result ID: 69b12caa6dbb7075b95babd5.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f72b2d603e6de3e45f53.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66581e209131cea0faaa.md>)

Recorded evaluation configuration: {"Backbone":"SD v2.1"}. This record does not assert independent reproduction.

### TRW

CLIP Score: 0.3073. Result ID: 69b12caa6dbb7075b95babd7.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f72b2d603e6de3e45f53.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70001e209131cea4872a.md>)

Recorded evaluation configuration: {"Backbone":"SD v2.1"}. This record does not assert independent reproduction.

### ShapeMark

CLIP Score: 0.3073. Result ID: 69b12caa6dbb7075b95babdb.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f72b2d603e6de3e45f53.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b0f72b2d603e6de3e45f53.md>)

Recorded evaluation configuration: {"Backbone":"SD v2.1"}. This record does not assert independent reproduction.

### dwtDct

CLIP Score: 0.3072. Result ID: 69b12caa6dbb7075b95babd3.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f72b2d603e6de3e45f53.md>)

Recorded evaluation configuration: {"Backbone":"SD v2.1"}. This record does not assert independent reproduction.

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