Hard Exudate Detection with Supervised Contrastive Learning
A novel supervised contrastive learning framework to optimize hard exudate detection is developed because early detection of hard exudates plays a crucial role in identifying DR, which aids in treating diabetes and preventing vision loss.
The unique characteristics of hard exudates, ranging from their inconsistent shapes to indistinct boundaries, pose significant challenges to existing segmentation techniques.
The technique exhibits its effectiveness and shows potential for computer-assisted hard exudate detection.
The supervised contrastive learning addresses these challenges effectively by utilizing label information to distinctly separate different classes, therefore managing varying lesion densities and unclear boundaries.
