# Watermark Removal on MS-COCO watermarked with RivaGAN (32 bits) (test)

Source: [Watermark Removal on MS-COCO watermarked with RivaGAN (32 bits) (test)](<https://www.sota2.com/research/sota/watermark-removal-on-ms-coco-watermarked-with-rivagan-32-bits-test>)

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

Data updated: 2026-03-04T09:14:19.153Z

Evaluation of algorithms designed to remove watermarks embedded using the RivaGAN method (32-bit version) from images in the MS-COCO test set, assessed using quality (PSNR, SSIM, FID) and accuracy metrics (BA, ACC, T1%F).

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 | Watermark Removal |
| Dataset (including any stated split) | MS-COCO watermarked with RivaGAN (32 bits) (test) |
| Metric | BA |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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

- [BA](<https://www.sota2.com/research/sota/watermark-removal-on-ms-coco-watermarked-with-rivagan-32-bits-test.md?metric=BA>): Higher is better; unit: PERCENT.

- [ACC](<https://www.sota2.com/research/sota/watermark-removal-on-ms-coco-watermarked-with-rivagan-32-bits-test.md?metric=ACC>): Higher is better; unit: PERCENT.

- [Top-1 F-score](<https://www.sota2.com/research/sota/watermark-removal-on-ms-coco-watermarked-with-rivagan-32-bits-test.md?metric=Top-1%20F-score>): Higher is better; unit: PERCENT.

- [PSNR](<https://www.sota2.com/research/sota/watermark-removal-on-ms-coco-watermarked-with-rivagan-32-bits-test.md?metric=PSNR>): Higher is better; unit: SCALAR.

- [SSIM](<https://www.sota2.com/research/sota/watermark-removal-on-ms-coco-watermarked-with-rivagan-32-bits-test.md?metric=SSIM>): Higher is better; unit: SCALAR.

- [FID](<https://www.sota2.com/research/sota/watermark-removal-on-ms-coco-watermarked-with-rivagan-32-bits-test.md?metric=FID>): Lower is better; unit: SCALAR.

## Recorded results

| Method | BA | Result date | Report paper |
| --- | --- | --- | --- |
| Watermarked | 100 | Not recorded | 69a66bde2d603e6de3e3cccd |
| Watermarked | 100 | Not recorded | 69a66bde2d603e6de3e3cccd |
| Regen | 60 | Not recorded | 69a66bde2d603e6de3e3cccd |
| Regen | 59 | Not recorded | 69a66bde2d603e6de3e3cccd |
| Ours | 56 | Not recorded | 69a66bde2d603e6de3e3cccd |
| Rinse | 53 | Not recorded | 69a66bde2d603e6de3e3cccd |
| Rinse | 52 | Not recorded | 69a66bde2d603e6de3e3cccd |
| Ours | 52 | Not recorded | 69a66bde2d603e6de3e3cccd |

### Watermarked

BA: 100. Result ID: 69a7f68b6dbb7075b94cff2a.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bde2d603e6de3e3cccd.md>)

Recorded evaluation configuration: {"Attack Setting":"Late start, noisy latent"}. This record does not assert independent reproduction.

### Watermarked

BA: 100. Result ID: 69a7f68b6dbb7075b94cff2e.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bde2d603e6de3e3cccd.md>)

Recorded evaluation configuration: {"Attack Setting":"Early start, inverse latent"}. This record does not assert independent reproduction.

### Regen

BA: 60. Result ID: 69a7f68b6dbb7075b94cff2b.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bde2d603e6de3e3cccd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa757a1e209131cea68378.md>)

[Implementation](<https://github.com/XuandongZhao/WatermarkAttacker>)

Recorded evaluation configuration: {"Attack Setting":"Late start, noisy latent"}. This record does not assert independent reproduction.

### Regen

BA: 59. Result ID: 69a7f68b6dbb7075b94cff2f.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bde2d603e6de3e3cccd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa757a1e209131cea68378.md>)

[Implementation](<https://github.com/XuandongZhao/WatermarkAttacker>)

Recorded evaluation configuration: {"Attack Setting":"Early start, inverse latent"}. This record does not assert independent reproduction.

### Ours

BA: 56. Result ID: 69a7f68b6dbb7075b94cff2d.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bde2d603e6de3e3cccd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66bde2d603e6de3e3cccd.md>)

Recorded evaluation configuration: {"Attack Setting":"Late start, noisy latent"}. This record does not assert independent reproduction.

### Rinse

BA: 53. Result ID: 69a7f68b6dbb7075b94cff30.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bde2d603e6de3e3cccd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738f1e209131cea5d37d.md>)

Recorded evaluation configuration: {"Attack Setting":"Early start, inverse latent"}. This record does not assert independent reproduction.

### Rinse

BA: 52. Result ID: 69a7f68b6dbb7075b94cff2c.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bde2d603e6de3e3cccd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738f1e209131cea5d37d.md>)

Recorded evaluation configuration: {"Attack Setting":"Late start, noisy latent"}. This record does not assert independent reproduction.

### Ours

BA: 52. Result ID: 69a7f68b6dbb7075b94cff31.

[Reporting paper](<https://www.sota2.com/research/paper/69a66bde2d603e6de3e3cccd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66bde2d603e6de3e3cccd.md>)

Recorded evaluation configuration: {"Attack Setting":"Early start, inverse latent"}. This record does not assert independent reproduction.

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