Expected Exponential Loss for Gaze-Based Video and Volume Ground Truth Annotation.
Laurent LejeuneMario ChristoudiasRaphael SznitmanPublished in: CVII-STENT/LABELS@MICCAI (2017)
Keyphrases
- ground truth
- video annotation
- weakly labeled
- video data
- video material
- video content
- eye tracking
- video database
- eye tracking data
- video analysis
- video streams
- multimedia
- video sequences
- eye movements
- metadata
- real time
- video images
- video clips
- human observers
- ground truth data
- worst case bounds
- high quality
- video surveillance
- semantic annotation
- video frames
- digital photos
- active learning
- computer vision
- gold standard
- highest resolution
- image retrieval
- semantic labels
- head movements
- human computer interaction
- eye typing
- key frames
- automatic annotation
- digital video