Learnable Irrelevant Modality Dropout for Multimodal Action Recognition on Modality-Specific Annotated Videos.
Saghir AlfaslyJian LuChen XuYuru ZouPublished in: CoRR (2022)
Keyphrases
- action recognition
- human actions
- multi modal
- action classification
- video dataset
- recognition of human actions
- recognizing human actions
- recognizing actions
- view invariant
- human activities
- action detection
- static images
- spatio temporal interest points
- human detection
- computer vision
- ucf sports
- space time interest points
- motion features
- activity recognition
- bag of words
- spatial temporal
- spatio temporal
- action recognition in videos
- motion history images
- body parts
- mid level features
- human object interactions
- video search
- motion recognition
- depth sensors
- human activity recognition
- video surveillance
- video sequences
- learning algorithm
- text classification
- view invariant action recognition
- high level
- human pose
- positive examples