ST-Adapter: Disentangled-and-Deformable Spatio-Temporal Adapter for Few-shot Action Recognition.
Wenjie PeiQizhong TanGuangming LuJiandong TianPublished in: CoRR (2023)
Keyphrases
- action recognition
- human actions
- spatio temporal
- spatial temporal
- action recognition in videos
- view invariant
- spatio temporal interest points
- recognition of human actions
- deformable part models
- computer vision
- human detection
- bag of words
- activity recognition
- video shots
- action classification
- video sequences
- recognizing human actions
- body parts
- depth sensors
- human pose
- visual features
- mid level
- action primitives
- human motion
- human activities
- independent subspace analysis
- recognizing actions
- keywords
- image sequences
- bag of features
- pose estimation
- static images
- video dataset
- single view
- moving objects
- space time
- spatial and temporal
- machine learning