A Topological Loss Function for Deep-Learning based Image Segmentation using Persistent Homology.
James R. CloughIlkay ÖksüzNicholas ByrneVeronika A. ZimmerJulia A. SchnabelAndrew P. KingPublished in: CoRR (2019)
Keyphrases
- loss function
- persistent homology
- deep learning
- image segmentation
- topological features
- morse theory
- computational geometry
- pairwise
- unsupervised learning
- support vector
- markov random field
- machine learning
- segmentation algorithm
- mental models
- graph cuts
- image processing
- multiscale
- segmentation method
- weakly supervised
- region growing
- multi class
- shape prior
- object segmentation
- computer vision
- higher order
- natural images
- data mining