SAViR-T: Spatially Attentive Visual Reasoning with Transformers.
Pritish SahuKalliopi BasiotiVladimir PavlovicPublished in: CoRR (2022)
Keyphrases
- visual motion
- visual information
- knowledge base
- pre attentive
- spatial reasoning
- visual features
- meta level
- analogical reasoning
- reasoning process
- visual perception
- knowledge representation
- low level
- visual processing
- neural network
- semantic web
- visual attention
- visual data
- real time
- human vision
- reasoning tasks
- video sequences
- visual analysis
- legal reasoning
- reasoning systems
- high level
- data sets