Link: Spatio-Temporal Action Detection with Occlusion.

STADO can be decomposed into one module and three branches:

1. Mask-Guided Attention Module

   Produces a spatial attention mask to modulate features generated by the backbone to focus on non-occlusion patterns.

2. Multi-Task Branches

   (2.1) Center Branch for center localization and action recognition.

   (2.2) Movement Branch for movement estimation at adjacent frames to form moving point trajectories.

   (2.3) Box Branch for spatial extent detection by directly regressing bounding box size at the estimated center point of each frame.

1. Installation
2. Dataset
3. Train
4. Evaluation

• Baseline model codes from MOC and its LICENSE.
• Data augmentation codes and evaluation codes from ACT and its LICENSE.
• DLA-34 backbone codes from CenterNet and its LICENSE.
• See more in NOTICE.