Kinematic-aware random forest for static and dynamic action recognition from depth sequences.
Seungryul BaekZhiyuan ShiMasato KawadeTae-Kyun KimPublished in: CoRR (2016)
Keyphrases
- random forest
- action recognition
- depth sensors
- random forests
- depth cameras
- decision trees
- human actions
- bag of words
- activity recognition
- motion history images
- feature set
- computer vision
- action classification
- action recognition in videos
- body parts
- recognizing human actions
- multi label
- depth images
- depth information
- human activities
- recognizing actions
- bag of features
- human pose
- depth data
- active learning
- machine learning
- object detection
- image classification
- k nearest neighbor
- depth map
- ensemble methods