Improving Weakly Supervised Scene Graph Parsing through Object Grounding.
Yizhou ZhangZhaoheng ZhengRam NevatiaYan LiuPublished in: ICPR (2022)
Keyphrases
- weakly supervised
- object detectors
- multiple objects
- superpixels
- object detection
- object class
- moving objects
- target object
- relation extraction
- object recognition
- object segmentation
- d objects
- graph structure
- spatial relations
- single image
- topic models
- object model
- object classes
- input image
- bounding box
- video sequences
- image sequences
- image regions
- natural language processing
- viewpoint
- partial occlusion
- object categories
- multiple images
- semi supervised
- natural language
- directed graph
- named entities
- natural images
- information extraction
- multiscale
- three dimensional