Study Group Learning: Improving Retinal Vessel Segmentation Trained with Noisy Labels
This work addresses data insufficiency and label noise in medical imaging for clinicians, but it is incremental as it builds on existing deep learning methods.
The paper tackles the problem of retinal vessel segmentation with noisy or incomplete labels by proposing a Study Group Learning scheme, which improves segmentation performance on DRIVE and CHASE_DB1 datasets, especially under noisy training conditions.
Retinal vessel segmentation from retinal images is an essential task for developing the computer-aided diagnosis system for retinal diseases. Efforts have been made on high-performance deep learning-based approaches to segment the retinal images in an end-to-end manner. However, the acquisition of retinal vessel images and segmentation labels requires onerous work from professional clinicians, which results in smaller training dataset with incomplete labels. As known, data-driven methods suffer from data insufficiency, and the models will easily over-fit the small-scale training data. Such a situation becomes more severe when the training vessel labels are incomplete or incorrect. In this paper, we propose a Study Group Learning (SGL) scheme to improve the robustness of the model trained on noisy labels. Besides, a learned enhancement map provides better visualization than conventional methods as an auxiliary tool for clinicians. Experiments demonstrate that the proposed method further improves the vessel segmentation performance in DRIVE and CHASE$\_$DB1 datasets, especially when the training labels are noisy.