Long Video Activity Recognition On Breakfast
评估指标
mAP
评测结果
各个模型在此基准测试上的表现结果
| Paper Title | Repository | ||
|---|---|---|---|
| AdaFocus (MViT-Breakfast-Pretrain-feature, GHRM) | 79.5 | Towards Weakly Supervised End-to-end Learning for Long-video Action Recognition | - |
| AdaFocus (MViT-Breakfast-Pretrain-feature, Timeception) | 79.2 | Towards Weakly Supervised End-to-end Learning for Long-video Action Recognition | - |
| AdaFocus (I3D-Breakfast-Pretrain-feature, Timeception) | 70.4 | Towards Weakly Supervised End-to-end Learning for Long-video Action Recognition | - |
| AdaFocus (I3D-Breakfast-Pretrain-feature, GHRM) | 69.6 | Towards Weakly Supervised End-to-end Learning for Long-video Action Recognition | - |
| GHRM (I3D-K400-Pretrain-feature) | 65.86 | Graph-Based High-Order Relation Modeling for Long-Term Action Recognition | - |
| VideoGraph (I3D-K400-Pretrain-feature) | 63.14 | VideoGraph: Recognizing Minutes-Long Human Activities in Videos | - |
| Timeception (I3D-K400-Pretrain-feature) | 61.82 | Timeception for Complex Action Recognition | |
| ActionVlad (I3D-K400-Pretrain-feature) | 60.20 | ActionVLAD: Learning spatio-temporal aggregation for action classification | - |
0 of 8 row(s) selected.