SoftCut: A Fully Differentiable Relaxed Graph Cut Approach for Deep Learning Image Segmentation.
Alessio BonfiglioMarco CanniciMatteo MatteucciPublished in: LOD (1) (2023)
Keyphrases
- graph cuts
- deep learning
- image segmentation
- energy function
- energy minimization
- markov random field
- shape prior
- unsupervised learning
- segmentation algorithm
- interactive image segmentation
- belief propagation
- graph based image segmentation
- object segmentation
- machine learning
- mean shift
- multi label
- weakly supervised
- interactive segmentation
- multiscale
- segmentation method
- min cut
- foreground background segmentation
- labeling problems
- computer vision
- level set
- pairwise
- mental models
- image processing
- probabilistic relaxation
- graph cut segmentation