简体中文 | English
We implemented action recgonition model and action localization model in this repo.
| Action recognition method | ||||
| PP-TSM (PP series) | PP-TSN (PP series) | PP-TimeSformer (PP series) | TSN (2D’) | TSM (2D') |
| SlowFast (3D’) | TimeSformer (Transformer') | VideoSwin (Transformer’) | TokenShift (3D’) | AttentionLSTM (RNN‘) |
| MoViNet (Lite‘) | ||||
| Skeleton based action recognition | ||||
| ST-GCN (Custom’) | AGCN (Adaptive') | 2s-AGCN (Adaptive') | CTR-GCN (GCN‘) | |
| Sequence action detection method | ||||
| BMN (One-stage') | ||||
| temporal segment | ||||
| MS-TCN | ASRF | |||
| Spatio-temporal motion detection method | ||||
| SlowFast+Fast R-CNN | ||||
| Multimodal | ||||
| ActBERT (Learning') | T2VLAD (Retrieval') | |||
| Video target segmentation | ||||
| CFBI (Semi') | MA-Net (Supervised') | |||
| Monocular depth estimation | ||||
| ADDS (Unsupervised‘) | ||||