Utilizing Transfer Learning and a Customized Loss Function for Optic Disc Segmentation from Retinal Images.
Abdullah SarhanAli Al-KhazÁlyAdam GornerAndrew SwiftJon G. RokneReda AlhajjAndrew CrichtonPublished in: CoRR (2020)
Keyphrases
- optic disc
- loss function
- transfer learning
- retinal images
- fundus images
- optic disk
- learning tasks
- vessel segmentation
- blood vessels
- support vector
- pairwise
- circular hough transform
- cross domain
- active learning
- retinal fundus images
- reinforcement learning
- labeled data
- diabetic retinopathy
- text classification
- machine learning
- logistic regression
- semi supervised learning
- domain adaptation
- learning algorithm
- multi task
- text mining
- region of interest
- multiscale
- image processing
- text categorization
- unlabeled data
- feature extraction
- feature selection
- computer vision