Actions as Moving Points

14 Jan 2020Yixuan LiZixu WangLimin WangGangshan Wu

The existing action tubelet detectors mainly depend on heuristic anchor design and placement, which might be computationally expensive and sub-optimal for precise localization of action instances. In this paper, we present a conceptually simple, computationally efficient, and more precise action tubelet detection framework, termed as MovingCenter Detector (MOC-detector), by treating an action instance as a trajectory of moving points... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Action Detection UCF101-24 Ours (MOC) mAP 27.7 # 1
Video-mAP 0.5 53.9 # 1
Video-mAP 0.75 28.5 # 1
Action Detection UCF101-24 MOC Video-mAP 0.2 81.8 # 1

Methods used in the Paper


METHOD TYPE
🤖 No Methods Found Help the community by adding them if they're not listed; e.g. Deep Residual Learning for Image Recognition uses ResNet