Supplementing Missing Visions via Dialog for Scene Graph Generations.
Ye ZhuXiaoguang ZhuYuzhang ShangZhenghao ZhaoYan YanPublished in: CoRR (2022)
Keyphrases
- d scene
- video sequences
- ambient intelligence
- graph structure
- graph theory
- graph matching
- three dimensional
- natural language
- scene understanding
- multiple images
- missing values
- bipartite graph
- mixed initiative
- graph theoretic
- weighted graph
- image sequences
- complex scenes
- scene classification
- scene analysis
- graph representation
- real scenes
- ground plane
- incomplete data
- connected components
- directed graph
- geometric constraints
- directed acyclic graph
- dynamic scenes
- input image
- single image
- multiple objects
- missing data
- viewpoint