VISITRON: Visual Semantics-Aligned Interactively Trained Object-Navigator.
Ayush ShrivastavaKarthik GopalakrishnanYang LiuRobinson PiramuthuGökhan TürDevi ParikhDilek Hakkani-TürPublished in: CoRR (2021)
Keyphrases
- visual objects
- visual appearance
- spatial relations
- visual properties
- visual representations
- visual features
- visual information
- formal semantics
- visual input
- object description
- semantically relevant
- semantic information
- complex objects
- d objects
- high level
- low level
- contextual cues
- logic programming
- object model
- multiple objects
- object models
- object identity
- spatial relationships
- object classes
- conceptual model
- object tracking
- keypoints
- feature vectors
- moving objects
- object recognition