Action Detection On J Hmdb
評価指標
Frame-mAP 0.5
Video-mAP 0.2
Video-mAP 0.5
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
比較表
| モデル名 | Frame-mAP 0.5 | Video-mAP 0.2 | Video-mAP 0.5 |
|---|---|---|---|
| actions-as-moving-points | 74 | 80.7 | 80.5 |
| tube-convolutional-neural-network-t-cnn-for | 61.3 | 78.4 | 76.9 |
| actionness-estimation-using-hybrid-fully | - | - | 56.4 |
| actionness-estimation-using-hybrid-fully | 39.9 | - | - |
| dance-with-flow-two-in-one-stream-action | - | - | 57.96 |
| tacnet-transition-aware-context-network-for-1 | 65.5 | 74.1 | 73.4 |
| you-only-watch-once-a-unified-cnn | 74.4 | 87.8 | 85.7 |
| holistic-interaction-transformer-network-for | 83.8 | 89.7 | 88.1 |
| hierarchical-self-attention-network-for | 76.72 | 85.97 | 84.02 |
| you-only-watch-once-a-unified-cnn | 75.7 | 88.3 | 85.9 |
| finding-action-tubes-with-a-sparse-to-dense | - | 76.1 | 74.3 |
| hierarchical-self-attention-network-for | - | 87.59 | 86.49 |
| multi-region-two-stream-r-cnn-for-action | 58.5 | 74.3 | 73.09 |
| multi-region-two-stream-r-cnn-for-action | 56.9 | 71.1 | 70.6 |
| finding-action-tubes | 36.2 | - | 53.3 |
| ava-a-video-dataset-of-spatio-temporally | 73.3 | - | 78.6 |
| dance-with-flow-two-in-one-stream-action | - | - | 74.74 |