Stacked Temporal Attention: Improving First-person Action Recognition by Emphasizing Discriminative Clips.
Lijin YangYifei HuangYusuke SuganoYoichi SatoPublished in: CoRR (2021)
Keyphrases
- action recognition
- action classification
- spatial temporal
- video database
- human actions
- mid level
- motion history images
- human detection
- bag of words
- recognizing actions
- activity recognition
- computer vision
- max margin
- spatio temporal
- body parts
- deformable part models
- human activities
- spatial and temporal
- recognizing human actions
- static images
- view invariant
- depth sensors
- recognition of human actions
- feature extraction
- bag of features
- temporal relations
- human pose
- video clips
- temporal information
- space time
- object recognition
- mid level features
- action detection
- human motion
- image classification
- action recognition in videos
- view invariant action recognition