Soft exudates segmentation
Computer Vision, Segmentation, 2024
This project utilizes a dataset from Kaggle containing original fundus images and their corresponding ground truth annotations.
The dataset is trained on a pretrained U-Net model over 20 epochs for medical image segmentation. The project demonstrates deep learning applied to medical imaging, specifically for segmenting key features in retinal images. Explore the code and images in this GitHub repository.
Link to the gihub repository:
