Do less and achieve more: Training CNNs for action recognition utilizing action images from the Web.
Shugao MaSarah Adel BargalJianming ZhangLeonid SigalStan SclaroffPublished in: Pattern Recognit. (2017)
Keyphrases
- action recognition
- human actions
- action classification
- recognizing human actions
- static images
- recognizing actions
- bag of features
- computer vision
- mid level
- view invariant
- recognition of human actions
- action detection
- action primitives
- bag of words
- video dataset
- image features
- web images
- image retrieval
- input image
- object recognition
- human detection
- image classification
- activity recognition
- three dimensional
- depth sensors
- space time interest points
- body parts
- feature points
- visual similarity
- human pose
- salient features
- human motion
- human activities
- test images
- video sequences
- search engine