Deepfake Detection on UniFace
98.1AUCSePL
Evaluation results
| Method | Links | |
|---|---|---|
| SePLTraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 98.1 | |
| HAADTraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation2026.05 | 97.1 | |
| EffortTraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 96.2 | |
| ForAdaTraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 95.4 | |
| EffortTraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation, Re-implemented=true2026.05 | 95 | |
| FIA-USATraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation, Re-implemented=true2026.05 | 91.8 | |
| ProDetTraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 90.8 | |
| DAIDTraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation, Re-implemented=true2026.05 | 90.2 | |
| RECCETraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 89.8 | |
| RECCETraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation2026.05 | 89.8 | |
| SLADDTraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 87.8 | |
| VLFFDTraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation, Re-implemented=true2026.05 | 87.5 | |
| X^2-DFDTraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation, Re-implemented=true2026.05 | 87.4 | |
| LSDATraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 87.2 | |
| LSDATraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation2026.05 | 87.2 | |
| CORETraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 87.1 | |
| CORETraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation2026.05 | 87.1 | |
| IIDTraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 83.9 | |
| IIDTraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation2026.05 | 83.9 | |
| UCFTraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 83.1 | |
| UCFTraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation2026.05 | 83.1 | |
| F3NetTraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 80.9 | |
| F3NetTraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation2026.05 | 80.9 | |
| CDFATraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 76.2 | |
| SRMTraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 74.9 | |
| SRMTraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation2026.05 | 74.9 | |
| SPSLTraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 74.7 | |
| SPSLTraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation2026.05 | 74.7 | |
| SBITraining Dataset=FaceForensics++ (c23), Evaluation Protocol=Cross-method2026.04 | 72.4 | |
| SBITraining Dataset=FF++, Evaluation Protocol=Cross-method Evaluation2026.05 | 72.4 |