AR-Net: Adaptive Frame Resolution for Efficient Action Recognition.
Yue MengChung-Ching LinRameswar PandaPrasanna SattigeriLeonid KarlinskyAude OlivaKate SaenkoRogério FerisPublished in: CoRR (2020)
Keyphrases
- action recognition
- human actions
- computer vision
- human detection
- action classification
- bag of words
- static images
- recognizing human actions
- body parts
- activity recognition
- spatial temporal
- depth sensors
- view invariant
- video dataset
- human pose
- human activities
- recognizing actions
- view invariant action recognition
- bag of features
- video frames
- low resolution
- three dimensional