Action Recognition on Drive&Act front-top NIR camera (test)
71.52Macro AccuracyPO-GUISE+
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PO-GUISE+Backbone=InternVideo2-B/14, GFlops=3992024.07 | 71.52 | 84.03 | |
| PO-GUISEBackbone=InternVideo2-B/14, GFlops=3992024.07 | 70.86 | 84.47 | |
| PO-GUISE+Backbone=VideoMAEv2-base, GFlops=2512024.07 | 70.35 | 84.83 | |
| PO-GUISEBackbone=VideoMAEv2-base, GFlops=2512024.07 | 69.47 | 83.6 | |
| InternVideo2-B/14GFlops=5092024.07 | 69.26 | 84.29 | |
| VideoMAEv2-baseGFlops=3602024.07 | 68.27 | 83.5 | |
| DRVMon-VMGFlops=3602024.07 | 62.64 | 77.27 | |
| TransdarcReproduction source=original code and weight parameters, GFlops=2812024.07 | 55.3 | 66.92 | |
| OA-SAR2024.07 | 54.07 | — | |
| SMOMS2024.07 | 51.39 | — | |
| I3DReproduction source=MMAction2, GFlops=332024.07 | 48.87 | 71.5 | |
| 3-stream2024.07 | 46.95 | — | |
| 2-stream2024.07 | 45.39 | — | |
| Pose2024.07 | 44.36 | — | |
| Interior2024.07 | 40.3 | — | |
| st-MLPReproduction source=original code and weight parameters2024.07 | 33.51 | — | |
| 3D-studentNetGFlops=372024.07 | — | 65.69 | |
| CTA-NET2024.07 | — | 65.25 |