VLPrompt: Vision-Language Prompting for Panoptic Scene Graph Generation.
Zijian ZhouMiaojing ShiHolger CaesarPublished in: CoRR (2023)
Keyphrases
- d scene
- video sequences
- computer vision
- three dimensional
- multiple images
- graph theory
- visual scene
- input image
- vision system
- graph model
- scene analysis
- real time
- image processing
- graph theoretic
- scene classification
- generation process
- graph representation
- scene understanding
- spanning tree
- rewriting rules
- complex scenes
- directed acyclic graph
- connected components
- structured data
- natural language
- weighted graph
- graph structure
- single image
- moving objects
- image sequences