COBE: Contextualized Object Embeddings from Narrated Instructional Video.
Gedas BertasiusLorenzo TorresaniPublished in: CoRR (2020)
Keyphrases
- multimedia
- complex objects
- video data
- video content
- object detection and tracking
- video streams
- real time
- object model
- instructional design
- learning outcomes
- video clips
- video sequences
- spatial and temporal
- video retrieval
- object motion
- successive frames
- vector space
- object tracking
- learning process
- temporal continuity
- spatial relationships
- bounding box
- video database
- video objects
- video analysis
- multiple objects
- object segmentation
- data objects
- key frames
- video frames
- d objects
- moving objects