VAFO-Loss: VAscular Feature Optimised Loss Function for Retinal Artery/Vein Segmentation.
Yukun ZhouMoucheng XuYipeng HuStefano B. BlumbergAn ZhaoSiegfried K. WagnerPearse A. KeaneDaniel C. AlexanderPublished in: CoRR (2022)
Keyphrases
- loss function
- vessel segmentation
- hinge loss
- blood vessels
- optical coherence tomography
- retinal images
- carotid artery
- tubular structures
- oct images
- retinal fundus images
- fundus images
- convex loss functions
- optic disc
- learning to rank
- logistic regression
- pairwise
- coronary artery
- low contrast
- square loss
- image segmentation
- support vector
- level set
- infrared images
- image features
- blood flow
- regularization term
- empirical risk
- risk minimization
- reproducing kernel hilbert space
- boosting framework
- segmentation algorithm
- region growing
- segmentation method
- potential functions
- convex optimization
- magnetic resonance angiography
- weight vector
- segmented regions
- boosting algorithms
- ultrasound images
- medical images
- feature vectors
- training data