Aligning where to see and what to tell: image caption with region-based attention and scene factorization.
Junqi JinKun FuRunpeng CuiFei ShaChangshui ZhangPublished in: CoRR (2015)
Keyphrases
- single image
- input image
- image segmentation
- multiscale
- image features
- image regions
- image retrieval
- image data
- image classification
- imaging process
- image content
- complex scenes
- scene matching
- scene images
- multiple images
- d scene
- geometric constraints
- piecewise planar
- video sequences
- reference images
- ground plane
- scene classification
- image representation
- low level
- high resolution
- segmentation method
- scene understanding
- image segments
- region based image
- multiple views
- uncalibrated images
- outdoor scenes
- vanishing points
- image registration
- moving objects
- object recognition
- image sequences