Deeply Supervised Active Learning for Finger Bones Segmentation
This work addresses the problem of reducing annotation effort for medical image segmentation, but it appears incremental as it builds on existing active learning and deep supervision techniques.
The paper tackles the challenge of finger bones segmentation in medical images by introducing a deeply supervised active learning approach, which achieves competitive segmentation results using fewer labeled samples compared to full annotation.
Segmentation is a prerequisite yet challenging task for medical image analysis. In this paper, we introduce a novel deeply supervised active learning approach for finger bones segmentation. The proposed architecture is fine-tuned in an iterative and incremental learning manner. In each step, the deep supervision mechanism guides the learning process of hidden layers and selects samples to be labeled. Extensive experiments demonstrated that our method achieves competitive segmentation results using less labeled samples as compared with full annotation.