GPT4SGG: Synthesizing Scene Graphs from Holistic and Region-specific Narratives.
Zuyao ChenJinlin WuZhen LeiZhaoxiang ZhangChangwen ChenPublished in: CoRR (2023)
Keyphrases
- input image
- d scene
- image regions
- three dimensional
- scene analysis
- graph representation
- real scenes
- dynamic scenes
- region of interest
- high level
- viewing angle
- single image
- higher level
- image sequences
- directed graph
- graph structure
- multiple images
- weighted graph
- graph model
- moving objects
- video sequences
- spatial location