Test-time augmentation-based active learning and self-training for label-efficient segmentation

Open in new window