Hi-SIGIR: Hierachical Semantic-Guided Image-to-image Retrieval via Scene Graph.
Yulu WangPengwen DaiXiaojun JiaZhitao ZengRui LiXiaochun CaoPublished in: ACM Multimedia (2023)
Keyphrases
- image retrieval
- image content
- input image
- single image
- image representation
- image collections
- image data
- low level features
- image features
- image matching
- image regions
- scene understanding
- image classification
- information retrieval
- complex scenes
- visual context
- multiscale
- image database
- reference images
- low level visual features
- image similarity
- d scene
- high level
- spatial relations
- scene geometry
- scene classification
- scene images
- image segmentation
- image set
- semantic information
- scene matching
- object recognition
- feature space
- high resolution
- feature points
- segmentation method
- spatial information
- semantic gap
- ground plane
- relevance feedback
- low level
- semantic concepts
- geometric constraints
- piecewise planar
- visual content