Test-Time Augmentation-Based Active Learning and Self-training for Label-Efficient Segmentation.
Bella Specktor-FadidaAnna LevchakovDana SchonbergerLiat Ben-SiraDafna Ben-BashatLeo JoskowiczPublished in: MILLanD@MICCAI (2023)
Keyphrases
- active learning
- uncertainty sampling
- experimental design
- level set
- image labeling
- image segmentation
- medical images
- cost sensitive
- segmentation algorithm
- semi supervised learning
- random sampling
- label noise
- multiscale
- selective sampling
- co training
- active learner
- training examples
- segmentation method
- learning strategies
- graph cuts
- ground truth
- image analysis
- training set
- training data