A Persistent Homology-Based Topological Loss for CNN-Based Multiclass Segmentation of CMR.
Nick ByrneJames R. CloughIsrael ValverdeGiovanni MontanaAndrew P. KingPublished in: IEEE Trans. Medical Imaging (2023)
Keyphrases
- graph cuts
- multi class
- persistent homology
- segmentation algorithm
- shape prior
- image segmentation
- topological features
- morse theory
- support vector machine
- multiclass classification
- computational geometry
- binary classification
- multiclass problems
- pairwise
- region growing
- object detection
- cost sensitive
- level set
- multi class classification
- binary classifiers
- segmentation method
- error correcting output codes
- hinge loss
- multiple classes
- perceptron algorithm
- multiclass learning
- multiclass support vector machines
- feature selection
- multiscale
- binary classification problems
- base classifiers
- link prediction
- knn