Annotation-Free and One-Shot Learning for Instance Segmentation of Homogeneous Object Clusters.
Zheng WuRuiheng ChangJiaxu MaCewu LuChi-Keung TangPublished in: CoRR (2018)
Keyphrases
- object segmentation
- multiple objects
- image segments
- graph cut segmentation
- data objects
- image segmentation
- bounding box
- global shape
- segmentation method
- clustering algorithm
- signed distance
- individual objects
- foreground and background
- medical images
- active learning
- d objects
- fully automatic
- level set
- segmentation algorithm
- shape prior
- unsupervised clustering
- multiscale
- image regions
- normalized cut
- object categories
- moving objects
- semantic annotation
- similar objects
- pixel level
- metadata
- object model
- region growing
- image annotation
- background subtraction
- intensity images
- edge detection
- object shape
- spatial relations
- fuzzy clustering
- hierarchical clustering
- segmented regions
- document clustering
- instance level
- shape model