ViSaRL: Visual Reinforcement Learning Guided by Human Saliency.
Anthony LiangJesse ThomasonErdem BiyikPublished in: CoRR (2024)
Keyphrases
- visual saliency
- reinforcement learning
- scan path
- human visual
- eye fixations
- visual attention
- function approximation
- state space
- saliency map
- natural images
- saliency model
- human behavior
- visual attributes
- optimal control
- high level
- information content
- spatial relations
- eye tracking
- visual features
- natural scenes
- human users
- visual search
- visual cues
- model free
- visual processing
- eye movements
- color perception
- low level
- visual search tasks