Generic User-guided Interaction Paradigm for Precise Post-slice-wise Processing of Tomographic Deep Learning Segmentations Utilizing Graph Cut and Graph Segmentation.
Gerald A. ZwettlerWerner BackfriederRonald A. KarwoskiDavid R. Holmes IIIPublished in: VISIGRAPP (4: VISAPP) (2021)
Keyphrases
- graph cuts
- deep learning
- object segmentation
- min cut
- image segmentation
- shape prior
- segmentation algorithm
- energy function
- figure ground
- energy minimization
- graph cut segmentation
- graph based image segmentation
- interactive segmentation
- weakly supervised
- max flow
- markov random field
- superpixels
- unsupervised learning
- segmentation result
- pairwise
- belief propagation
- mean shift
- multi label
- user interaction
- domain specific
- graph partitioning
- foreground background segmentation
- segmentation method
- level set
- machine learning
- medical images
- labeling problems
- mental models
- bounding box
- image processing
- multiscale
- directed graph
- deformable models
- natural images
- loopy belief propagation
- shape model
- long range
- active contours
- higher order
- ground truth
- computer vision