Localizing Actions from Video Labels and Pseudo-Annotations.
Pascal MettesCees G. M. SnoekShih-Fu ChangPublished in: CoRR (2017)
Keyphrases
- image sequences
- video sequences
- semantic labels
- human actions
- action detection
- video data
- human motion
- video material
- weakly labeled
- natural language descriptions
- video content
- video streams
- video clips
- body motions
- training data
- real time
- metadata
- video analysis
- video database
- temporal structure
- situation calculus
- event detection
- human activities
- space time interest points
- plan recognition
- video frames
- multi label
- action recognition
- keywords
- multimedia
- digital video
- action selection
- user generated
- ordering constraints
- video surveillance
- semantic annotation