Localizing Actions from Video Labels and Pseudo-Annotations.
Pascal MettesCees SnoekShih-Fu ChangPublished in: BMVC (2017)
Keyphrases
- semantic labels
- action detection
- video sequences
- weakly labeled
- human actions
- natural language descriptions
- video data
- action recognition
- recognizing human actions
- video content
- multimedia
- situation calculus
- video streams
- human activities
- video database
- plan recognition
- video analysis
- real time
- video frames
- space time interest points
- online video
- temporal structure
- body motions
- metadata
- reasoning about actions
- digital video
- multimedia data
- spatial and temporal
- visual concepts
- visual content
- video retrieval
- video surveillance
- annotation tool
- moving objects
- pairwise
- training set
- keywords