High-Level Bottom-Up Cues for Top-Down Parsing of Facade Images.
David OkMateusz KozinskiRenaud MarletNikos ParagiosPublished in: 3DIMPVT (2012)
Keyphrases
- high level
- image data
- high level information
- mid level
- input image
- low level
- ground truth
- image database
- image retrieval
- image analysis
- test images
- image classification
- figure ground segmentation
- image collections
- three dimensional
- feature points
- figure ground
- visual patterns
- rigid body
- edge detection
- object recognition
- satellite images
- depth perception
- visual cues
- image set
- image features
- image regions
- segmentation method
- natural language
- partially occluded
- segmentation algorithm
- depth cues
- image sequences
- natural images
- image registration