ESceme: Vision-and-Language Navigation with Episodic Scene Memory.
Qi ZhengDaqing LiuChaoyue WangJing ZhangDadong WangDacheng TaoPublished in: CoRR (2023)
Keyphrases
- d scene
- episodic memory
- multiple images
- computer vision
- real scenes
- programming language
- video sequences
- vision system
- single image
- memory requirements
- three dimensional
- real time
- natural language
- image processing
- long term memory
- language understanding
- memory space
- scene classification
- multiple objects
- language learning
- image sequences
- scene analysis
- memory usage
- scene understanding
- human vision
- obstacle avoidance
- service robots
- input image