Video Anomaly Detection on ShanghaiTech Mixture [01,02] scene-dependent
91Micro AUCHSC w/ MA-+
Evaluation Results
| Method | Links | |
|---|---|---|
| HSC w/ MA-+Input=Appearance (OA) + Skeleton (OS), Motion Augmentation=Normal and abnormal samples2023.03 | 91 | |
| HSC w/ MA-Input=Appearance (OA) + Skeleton (OS), Motion Augmentation=Normal samples only2023.03 | 90.1 | |
| HSCInput=Appearance (OA) + Skeleton (OS), Motion Augmentation=None2023.03 | 87.3 | |
| HSC w/ MA-+Input=Appearance (OA) + Skeleton (OS), Motion Augmentation=Normal and abnormal samples2023.03 | 86.9 | |
| HSC w/ MA-Input=Appearance (OA) + Skeleton (OS), Motion Augmentation=Normal samples only2023.03 | 85.7 | |
| HSCInput=Appearance (OA) + Skeleton (OS), Motion Augmentation=None2023.03 | 82.8 | |
| MPNInput=Frame-level (F)2023.03 | 78.4 | |
| MNADInput=Frame-level (F)2023.03 | 77.8 | |
| Mem-AEInput=Frame-level (F)2023.03 | 77.7 | |
| HF2-VADInput=Appearance (OA) + Optical Flow (OF)2023.03 | 74.8 | |
| HF2-VADInput=Appearance (OA) + Optical Flow (OF)2023.03 | 66.8 | |
| Mem-AEInput=Frame-level (F)2023.03 | 50.2 | |
| MNADInput=Frame-level (F)2023.03 | 50 | |
| MPNInput=Frame-level (F)2023.03 | 45.3 |