Object-centric Video Prediction without Annotation.
Karl SchmeckpeperGeorgios GeorgakisKostas DaniilidisPublished in: ICRA (2021)
Keyphrases
- weakly labeled
- video annotation
- video sequences
- video data
- object detection and tracking
- d objects
- prediction accuracy
- video streams
- successive frames
- video frames
- multimedia
- complex objects
- moving objects
- automatic annotation
- real time
- semantic annotation
- video material
- video content
- active learning
- object motion
- prediction model
- object model
- prediction error
- image retrieval
- video objects
- motion trajectories
- stationary camera
- key frames
- multimedia data
- semantic labels
- object recognition
- combining information from multiple
- video analysis
- bounding box
- object tracking
- target object
- video surveillance
- data objects
- event detection
- spatial and temporal