Unsupervised approximate-semantic vocabulary learning for human action and video classification.
Qiong ZhaoHorace Ho-Shing IpPublished in: Pattern Recognit. Lett. (2013)
Keyphrases
- video classification
- human actions
- vocabulary learning
- video shots
- action recognition
- spatio temporal
- human motion
- space time
- human activities
- language learning
- video clips
- activity recognition
- visual features
- video database
- video sequences
- computer assisted
- semantic information
- video content
- human pose
- semi supervised
- high level
- semantic concepts
- dynamic textures
- low level features
- video indexing
- body parts
- information extraction
- video retrieval
- active learning
- motion trajectories
- foreign language
- mobile devices
- key frames