Multi-Agent Embodied Visual Semantic Navigation With Scene Prior Knowledge.
Xinzhu LiuDi GuoHuaping LiuFuchun SunPublished in: IEEE Robotics Autom. Lett. (2022)
Keyphrases
- prior knowledge
- multi agent
- semantic context
- observed scene
- scene categorization
- scene categories
- visual information
- visual context
- visual data
- high level
- d scene
- intelligent agents
- visual scene
- video sequences
- semantic information
- scene recognition
- spatial relations
- single image
- visual environment
- multi agent systems
- low level
- scene understanding
- three dimensional
- semantically relevant
- semantic space
- scene images
- semantic content
- visual concepts
- visual navigation
- complex scenes
- reinforcement learning
- moving objects
- low level features
- domain knowledge
- image sequences
- multiple images
- real scenes
- semantic web
- multiagent systems
- scene interpretation
- image regions
- image representation
- indoor environments
- spatial layout
- visual appearance
- visual input
- prior information
- cognitive science
- sports video
- input image
- scene classification
- natural language
- semantic similarity
- cognitive processes