Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

FF++

Benchmarks

Task NameDataset NameSOTA ResultTrend
Deepfake DetectionFF++ video-level 8 (test)
Accuracy98.01
40
Deepfake DetectionFF++ (test)
AUC99.96
39
Deepfake DetectionFF++
AUC99.9
34
Face Forgery DetectionFF++ (HQ)
AUC DF100
27
Deepfake DetectionFF++ HQ (test)
AUC (DF)0.999
24
Deepfake DetectionFF++ Intra-dataset c23
AUC100
24
Deepfake AttackFF++ spatial-based (test)
ResNet Accuracy98.2
24
Adversarial Attack on Deepfake DetectorsFF++ (test)
ASR (DCTA)87.5
18
Facial manipulation detectionFF++
AUC0.9444
16
Deepfake DetectionFF++ Intra-dataset (c40)
Accuracy100
15
Deepfake DetectionFF++ HQ
AUC0.9931
12
Frame-level Face Forgery DetectionFF++ HQ
AUC99.46
12
DeepFake DetectionFF++ Average across scenarios
Accuracy95.7
10
DeepFake DetectionFF++ Adversarial Attack
Accuracy91
10
DeepFake DetectionFF++ Partial Noisy
Accuracy100
10
DeepFake DetectionFF++ Partial Blurring
Accuracy96
10
DeepFake DetectionFF++ Sunglass Masking
Accuracy96
10
Face SwappingFF++
ID Similarity0.636
10
Face Forgery ClassificationFF++ (test)
AUC99.9
9
Face swap detectionFF++ c40 compression (test)
Accuracy93.83
7
Face swap detectionFF++ c23 compression (test)
Accuracy99.11
7
LipSync Manipulation DetectionFF++ (test)
ACC95.1
7
Deepfake DetectionFF++ Protocol-2 (c23)
UniFace Score96
5
Face Forgery DetectionFF++ leave-one-out manipulation HQ (test)
Avg Video-level AUC99.7
4
Deepfake DetectionFF++
Accuracy97
4
Showing 25 of 26 rows