# Puppeteering attack detection on CREMA-D and RAVDESS Cross-domain

Source: [Puppeteering attack detection on CREMA-D and RAVDESS Cross-domain](<https://www.sota2.com/research/sota/puppeteering-attack-detection-on-crema-d-and-ravdess-cross-domain>)

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

Data updated: 2026-04-13T09:40:14.966Z

Evaluation of model performance in detecting puppeteering attacks using a cross-domain setup combining the CREMA-D and RAVDESS datasets, reporting AUC scores across various attack methods (3DFaceShop, MCNET, EMOPortraits, SDFR, LivePortrait).

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 | Puppeteering attack detection |
| Dataset (including any stated split) | CREMA-D and RAVDESS Cross-domain |
| Metric | 3DFaceShop AUC |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 9 of 9 |
| 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

- [3DFaceShop AUC](<https://www.sota2.com/research/sota/puppeteering-attack-detection-on-crema-d-and-ravdess-cross-domain.md?metric=3DFaceShop%20AUC>): Higher is better; unit: PERCENT.

- [MCNET AUC](<https://www.sota2.com/research/sota/puppeteering-attack-detection-on-crema-d-and-ravdess-cross-domain.md?metric=MCNET%20AUC>): Higher is better; unit: PERCENT.

- [EMOPortraits AUC](<https://www.sota2.com/research/sota/puppeteering-attack-detection-on-crema-d-and-ravdess-cross-domain.md?metric=EMOPortraits%20AUC>): Higher is better; unit: PERCENT.

- [SDFR AUC](<https://www.sota2.com/research/sota/puppeteering-attack-detection-on-crema-d-and-ravdess-cross-domain.md?metric=SDFR%20AUC>): Higher is better; unit: PERCENT.

- [LivePortrait AUC](<https://www.sota2.com/research/sota/puppeteering-attack-detection-on-crema-d-and-ravdess-cross-domain.md?metric=LivePortrait%20AUC>): Higher is better; unit: PERCENT.

- [Average AUC](<https://www.sota2.com/research/sota/puppeteering-attack-detection-on-crema-d-and-ravdess-cross-domain.md?metric=Average%20AUC>): Higher is better; unit: PERCENT.

## Recorded results

| Method | 3DFaceShop AUC | Result date | Report paper |
| --- | --- | --- | --- |
| PC-LMCL | 0.914 | Not recorded | 68fa7be91e209131cea8b44b |
| Vahdati et al. | 0.895 | Not recorded | 68fa7be91e209131cea8b44b |
| SupCon | 0.659 | Not recorded | 68fa7be91e209131cea8b44b |
| TALL | 0.639 | Not recorded | 68fa7be91e209131cea8b44b |
| CNN Ensemble | 0.629 | Not recorded | 68fa7be91e209131cea8b44b |
| CLR NET | 0.627 | Not recorded | 68fa7be91e209131cea8b44b |
| CCE ViT | 0.598 | Not recorded | 68fa7be91e209131cea8b44b |
| Efficient ViT | 0.58 | Not recorded | 68fa7be91e209131cea8b44b |
| LAANet | 0.558 | Not recorded | 68fa7be91e209131cea8b44b |

### PC-LMCL

3DFaceShop AUC: 0.914. Result ID: 69dcb9b647f34ce924c40995.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7be91e209131cea8b44b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7be91e209131cea8b44b.md>)

Recorded evaluation configuration: {"Domain":"Cross-domain"}. This record does not assert independent reproduction.

### Vahdati et al.

3DFaceShop AUC: 0.895. Result ID: 69dcb9b647f34ce924c40994.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7be91e209131cea8b44b.md>)

Recorded evaluation configuration: {"Domain":"Cross-domain"}. This record does not assert independent reproduction.

### SupCon

3DFaceShop AUC: 0.659. Result ID: 69dcb9b647f34ce924c40991.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7be91e209131cea8b44b.md>)

Recorded evaluation configuration: {"Domain":"Cross-domain"}. This record does not assert independent reproduction.

### TALL

3DFaceShop AUC: 0.639. Result ID: 69dcb9b647f34ce924c40990.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7be91e209131cea8b44b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72381e209131cea55071.md>)

Recorded evaluation configuration: {"Domain":"Cross-domain"}. This record does not assert independent reproduction.

### CNN Ensemble

3DFaceShop AUC: 0.629. Result ID: 69dcb9b647f34ce924c4098f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7be91e209131cea8b44b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa675f1e209131cea15f8d.md>)

[Implementation](<https://github.com/polimi-ispl/icpr2020dfdc>)

Recorded evaluation configuration: {"Domain":"Cross-domain"}. This record does not assert independent reproduction.

### CLR NET

3DFaceShop AUC: 0.627. Result ID: 69dcb9b647f34ce924c40992.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7be91e209131cea8b44b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa684e1e209131cea1b4e4.md>)

Recorded evaluation configuration: {"Domain":"Cross-domain"}. This record does not assert independent reproduction.

### CCE ViT

3DFaceShop AUC: 0.598. Result ID: 69dcb9b647f34ce924c4098e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7be91e209131cea8b44b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ce31e209131cea3689a.md>)

[Implementation](<https://github.com/davide-coccomini/Combining-EfficientNet-and-Vision-Transformers-for-Video-Deepfake-Detection>)

Recorded evaluation configuration: {"Domain":"Cross-domain"}. This record does not assert independent reproduction.

### Efficient ViT

3DFaceShop AUC: 0.58. Result ID: 69dcb9b647f34ce924c4098d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7be91e209131cea8b44b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ce31e209131cea3689a.md>)

[Implementation](<https://github.com/davide-coccomini/Combining-EfficientNet-and-Vision-Transformers-for-Video-Deepfake-Detection>)

Recorded evaluation configuration: {"Domain":"Cross-domain"}. This record does not assert independent reproduction.

### LAANet

3DFaceShop AUC: 0.558. Result ID: 69dcb9b647f34ce924c40993.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7be91e209131cea8b44b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73561e209131cea5becb.md>)

Recorded evaluation configuration: {"Domain":"Cross-domain"}. This record does not assert independent reproduction.

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