Domain adaptation for retinal vessel segmentation using asymmetrical maximum classifier discrepancy.
Jingyu ZhuangZiliang ChenJunyi ZhangDongyu ZhangZhaoquan CaiPublished in: ACM TUR-C (2019)
Keyphrases
- domain adaptation
- vessel segmentation
- retinal images
- blood vessels
- training data
- co training
- labeled data
- test data
- target domain
- multiple sources
- vessel enhancement
- semi supervised
- low contrast
- region growing
- cross domain
- training samples
- semi supervised learning
- decision trees
- optic disc
- sentiment classification
- learning algorithm
- unlabeled data
- classification algorithm
- transfer learning
- feature space
- active learning
- feature set
- training set
- support vector machine
- multiscale
- image segmentation
- class labels
- feature selection
- data sets
- edge detection