VDARN: Video Disentangling Attentive Relation Network for Few-Shot and Zero-Shot Action Recognition.
Yong SuMeng XingSimin AnWeilong PengZhiyong FengPublished in: Ad Hoc Networks (2021)
Keyphrases
- action recognition
- human actions
- action classification
- spatial temporal
- video database
- video dataset
- video sequences
- action detection
- recognizing human actions
- video data
- recognition of human actions
- video content
- human activities
- video shots
- activity recognition
- bag of words
- motion features
- human detection
- space time interest points
- key frames
- static images
- body parts
- computer vision
- mid level
- view invariant
- bag of features
- space time
- recognizing actions
- video indexing
- video streams
- news video
- motion history images
- video analysis
- multimedia
- human pose
- machine learning
- video retrieval
- video surveillance
- visual features
- image features
- spatio temporal
- object recognition
- view invariant action recognition