Zero-shot Object Prediction using Semantic Scene Knowledge.
Rene GrzeszickGernot A. FinkPublished in: VISIGRAPP (5: VISAPP) (2017)
Keyphrases
- complex scenes
- domain knowledge
- multiple objects
- domain ontology
- moving objects
- semantic knowledge
- semantic information
- real world scenes
- semantically rich
- semantic models
- semantic network
- target object
- knowledge base
- object models
- d scene
- ground plane
- image sequences
- uncalibrated images
- related concepts
- three dimensional
- relative position
- physical objects
- prediction accuracy
- knowledge management
- expert systems
- natural language
- video sequences
- spatial relations
- object detectors
- image regions
- semantic web
- real objects
- individual objects
- d objects
- knowledge representation
- visual context