Action Recognition In Videos On Something 2
Metrics
Top 1 Accuracy
Results
Performance results of various models on this benchmark
| Paper Title | ||
|---|---|---|
| STM (16 frames, ImageNet pretraining) | 50.7 | STM: SpatioTemporal and Motion Encoding for Action Recognition |
| Motion Feature Net | 43.9 | Motion Feature Network: Fixed Motion Filter for Action Recognition |
| 2-Stream TRN | 42.01 | Temporal Relational Reasoning in Videos |
0 of 3 row(s) selected.