# Imperceptibility Evaluation on CelebA-HQ

Source: [Imperceptibility Evaluation on CelebA-HQ](<https://www.sota2.com/research/sota/imperceptibility-evaluation-on-celeba-hq>)

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

Data updated: 2026-02-26T06:03:25.318Z

Evaluation of image imperceptibility on the CelebA-HQ dataset using structural similarity (SSIM), peak signal-to-noise ratio (PSNR), and learned perceptual image patch similarity (LPIPS).

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 | Imperceptibility Evaluation |
| Dataset (including any stated split) | CelebA-HQ |
| Metric | SSIM |
| 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

- [SSIM](<https://www.sota2.com/research/sota/imperceptibility-evaluation-on-celeba-hq.md?metric=SSIM>): Higher is better; unit: SCALAR.

- [PSNR](<https://www.sota2.com/research/sota/imperceptibility-evaluation-on-celeba-hq.md?metric=PSNR>): Higher is better; unit: SCALAR.

- [LPIPS](<https://www.sota2.com/research/sota/imperceptibility-evaluation-on-celeba-hq.md?metric=LPIPS>): Lower is better; unit: SCALAR.

## Recorded results

| Method | SSIM | Result date | Report paper |
| --- | --- | --- | --- |
| FaceDefense | 0.96 | Not recorded | 69802f72a8b01e76cc8e153b |
| Myface | 0.924 | Not recorded | 69802f72a8b01e76cc8e153b |
| SDST | 0.73 | Not recorded | 69802f72a8b01e76cc8e153b |
| SDS+ | 0.662 | Not recorded | 69802f72a8b01e76cc8e153b |
| AdvDM- | 0.653 | Not recorded | 69802f72a8b01e76cc8e153b |
| Mist | 0.589 | Not recorded | 69802f72a8b01e76cc8e153b |
| DiffusionGuard | 0.554 | Not recorded | 69802f72a8b01e76cc8e153b |
| SDS- | 0.553 | Not recorded | 69802f72a8b01e76cc8e153b |
| PhotoGuard | 0.509 | Not recorded | 69802f72a8b01e76cc8e153b |
| AdvDM+ | 0.507 | Not recorded | 69802f72a8b01e76cc8e153b |

### FaceDefense

SSIM: 0.96. Result ID: 6997718d5698665ec4c58d10.

[Reporting paper](<https://www.sota2.com/research/paper/69802f72a8b01e76cc8e153b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69802f72a8b01e76cc8e153b.md>)

Recorded evaluation configuration: {"epsilon":"75/255"}. This record does not assert independent reproduction.

### Myface

SSIM: 0.924. Result ID: 6997718d5698665ec4c58d0f.

[Reporting paper](<https://www.sota2.com/research/paper/69802f72a8b01e76cc8e153b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78b71e209131cea79fb0.md>)

Recorded evaluation configuration: {"epsilon":"75/255"}. This record does not assert independent reproduction.

### SDST

SSIM: 0.73. Result ID: 6997718d5698665ec4c58d0e.

[Reporting paper](<https://www.sota2.com/research/paper/69802f72a8b01e76cc8e153b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa721c1e209131cea548c5.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### SDS+

SSIM: 0.662. Result ID: 6997718d5698665ec4c58d0c.

[Reporting paper](<https://www.sota2.com/research/paper/69802f72a8b01e76cc8e153b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa721c1e209131cea548c5.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### AdvDM-

SSIM: 0.653. Result ID: 6997718d5698665ec4c58d08.

[Reporting paper](<https://www.sota2.com/research/paper/69802f72a8b01e76cc8e153b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa721c1e209131cea548c5.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Mist

SSIM: 0.589. Result ID: 6997718d5698665ec4c58d0a.

[Reporting paper](<https://www.sota2.com/research/paper/69802f72a8b01e76cc8e153b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f641e209131cea45464.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### DiffusionGuard

SSIM: 0.554. Result ID: 6997718d5698665ec4c58d09.

[Reporting paper](<https://www.sota2.com/research/paper/69802f72a8b01e76cc8e153b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ba01e209131cea896a1.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### SDS-

SSIM: 0.553. Result ID: 6997718d5698665ec4c58d0d.

[Reporting paper](<https://www.sota2.com/research/paper/69802f72a8b01e76cc8e153b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa721c1e209131cea548c5.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### PhotoGuard

SSIM: 0.509. Result ID: 6997718d5698665ec4c58d0b.

[Reporting paper](<https://www.sota2.com/research/paper/69802f72a8b01e76cc8e153b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e921e209131cea40c44.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### AdvDM+

SSIM: 0.507. Result ID: 6997718d5698665ec4c58d07.

[Reporting paper](<https://www.sota2.com/research/paper/69802f72a8b01e76cc8e153b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fa11e209131cea46cbd.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

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