Weakly Supervised Action Localization On 5
評価指標
avg-mAP (0.1-0.5)
avg-mAP (0.1:0.7)
avg-mAP (0.3-0.7)
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 | avg-mAP (0.1-0.5) | avg-mAP (0.1:0.7) | avg-mAP (0.3-0.7) |
|---|---|---|---|
| adaptive-mutual-supervision-for-weakly | 52.0 | 42.3 | 32.4 |
| proposal-based-multiple-instance-learning-for-1 | 57.4 | 47.0 | 38.0 |
| hr-pro-point-supervised-temporal-action | 71.6 | 60.3 | 51.1 |
| learning-action-completeness-from-points-for | 62.7 | 52.8 | 44.5 |
| weakly-supervised-action-localization-with-2 | 44.9 | 37.7 | 30.4 |
| two-stream-consensus-network-for-weakly-1 | 47.0 | 37.8 | 28.8 |
| acm-net-action-context-modeling-network-for | 53.2 | 42.6 | 33.4 |
| background-modeling-via-uncertainty | 51.6 | 41.9 | 32.9 |
| cola-weakly-supervised-temporal-action | 50.3 | 40.9 | 32.1 |
| point-level-temporal-action-localization | 55.6 | 44.8 | 35.4 |
| action-unit-memory-network-for-weakly | 52.1 | 41.5 | 32.4 |
| adversarial-background-aware-loss-for-weakly | 46.9 | 37.8 | 30.6 |