Trimmed Action Recognition, Dense-Captioning Events in Videos, and Spatio-temporal Action Localization with Focus on ActivityNet Challenge 2019.
Zhaofan QiuDong LiYehao LiQi CaiYingwei PanTing YaoPublished in: CoRR (2019)
Keyphrases
- human actions
- action recognition
- spatio temporal
- human activities
- action classification
- view invariant
- spatio temporal interest points
- recognition of human actions
- video dataset
- recognizing human actions
- recognizing actions
- activity recognition
- spatial temporal
- action recognition in videos
- video database
- space time interest points
- computer vision
- bag of words
- atomic actions
- human detection
- space time
- body parts
- bag of features
- action detection
- motion history images
- motion features
- human motion
- motion recognition
- video sequences
- ucf sports
- event detection
- event recognition
- static images
- human pose
- action primitives
- video surveillance
- human object interactions
- global features
- human computer interaction
- machine learning