EdgeNet: Semantic Scene Completion from a Single RGB- D Image.
Aloisio DouradoTeófilo E. de CamposHansung KimAdrian HiltonPublished in: ICPR (2020)
Keyphrases
- input image
- single image
- image features
- image data
- image regions
- scene understanding
- image segmentation
- scene images
- indoor scenes
- image representation
- image segments
- image classification
- scene classification
- complex scenes
- multiscale
- multiple objects
- imaging process
- multiple images
- high resolution
- uncalibrated cameras
- vanishing points
- d scene
- visual context
- image set
- spatial relations
- image sequences
- geometric constraints
- segmentation method
- scene categorization
- semantic information
- reference images
- segmentation algorithm
- feature points
- image retrieval
- scene matching
- piecewise planar
- camera tracking
- scene geometry
- ground plane
- image collections
- image matching
- color images
- moving objects
- three dimensional