The Fusion of Supervised and Unsupervised Techniques for Segmentation of Abnormal Regions.
Ali HadjarianJerzy W. BalaSrinivas GuttaShirin TrachiotisPeter PachowiczPublished in: Digital Mammography / IWDM (1998)
Keyphrases
- fully unsupervised
- segmented images
- unsupervised methods
- unsupervised learning
- semi supervised
- supervised learning
- image regions
- accurate segmentation
- image segmentation
- segmentation algorithm
- homogeneous regions
- supervised classification
- region segmentation
- motion based segmentation
- watershed algorithm
- textured regions
- region merging
- supervised training
- segmentation scheme
- segmenting images
- region matching
- weakly supervised
- watershed segmentation
- text segmentation
- unsupervised feature selection
- segmented regions
- figure ground
- segmentation method
- bayesian image segmentation
- medical images
- image segmentations
- automatically segmented
- anomaly detection
- piecewise constant
- normalized cut
- outdoor scenes
- multiscale
- learning algorithm
- connected regions
- supervised methods
- region boundaries
- image pixels
- complex background
- intensity images
- image segments
- discriminant projection
- histogram analysis
- data fusion
- level set
- foreground and background
- background pixels
- image features
- initial segmentation
- object segmentation
- image fusion
- energy function
- spatial coherence
- region growing