On the influence of Dice loss function in multi-class organ segmentation of abdominal CT using 3D fully convolutional networks.
Chen ShenHolger R. RothHirohisa OdaMasahiro OdaYuichiro HayashiKazunari MisawaKensaku MoriPublished in: CoRR (2018)
Keyphrases
- multi class
- loss function
- liver segmentation
- organ segmentation
- ct data
- ct images
- pairwise
- computed tomography
- medical image processing
- abdominal organs
- fully automatic
- medical image segmentation
- ct scans
- boosting algorithms
- automatic segmentation
- support vector
- medical images
- patient specific
- medical imaging
- support vector machine
- multi class classification
- intraoperative
- binary classification
- logistic regression
- volume rendering
- statistical shape model
- cost sensitive
- multi class boosting
- regularization term
- image reconstruction
- base classifiers
- error correcting output codes
- region of interest
- three dimensional
- machine learning
- semi supervised
- similarity measure
- image processing
- feature selection