fundus-images

Official website of our paper: Applications of Deep Learning in Fundus Images: A Review. Newly-released datasets and recently-published papers will be updated regularly.

Official code release for the EMNLP 2026 paper "Med-Banana: Learning Quality-Controlled Medical Image Editing from Success-and-Failure Trajectories" — Med-Banana-80K dataset + edit–verify–refine system

AUTOMATED TYPE CLASSIFICATION OF GLAUCOMA DETECTION USING DEEP LEARNING

Joint Vessel Segmentation and Deformable Registration on Multi-Modal Retinal Images based on Style Transfer

Glaucoma detection automation project. Trained a binary image classifier using CNNs and deployed as a streamlit web app. It takes eye (retinal scan) image as input and outputs whether the person is affected by glaucoma or not.

Classification of Fundus Images into 5 stages of Diabetic Retinopathy, and segmentation of blood vessels in fundus images

Code for the paper "OTRE: Where Optimal Transport Guided Unpaired Image-to-Image Translation Meets Regularization by Enhancing"

Learning Self-Supervised Representations for Label Efficient Cross-Domain Knowledge Transfer on Diabetic Retinopathy Fundus Images (IJCNN 2023)

Deep learning pipeline for classification of Cataract, Diabetic Retinopathy, Glaucoma and Normal using fundus images

Ultra: Multi-Granularity Topological Reasoning for Anatomically Consistent Vasculature Parsing

Blood vessels and Exudates extraction for the detection of Diabetic Retinopathy

AI-powered diabetic retinopathy detection web app using EfficientNet with Grad-CAM visual explanations for interpretable medical diagnosis, developed for research and educational purposes.

Open-source glaucoma detection AI for mobile/low-resource clinics using synthetic training data

Fair, demographically-controllable synthetic medical image generation (chest X-ray & fundus) via hierarchical compositional diffusion.

Reproducible optic disc/cup segmentation and public-to-clinical transfer study for AI-enhanced ophthalmoscopy.

Diabetic Retinopathy using Patch Networks.

Binary classification of Diabetic Retinopathy using SVM in MATLAB with fundus image features.

Domain-adaptive retinal vessel segmentation. A weakly supervised framework transferring vessel segmentation from DRIVE, CHASE_DB1 and FIVES to the unannotated RFMiD dataset, using confidence-masked pseudo-label supervision and perturbation consistency. Includes training, adaptation, evaluation and pseudo-label generation code.

Comparison of Classical and Deep Learning-Based Feature Representations for Age-Related Macular Degeneration

Research about glaucoma detection using CDR value. Trying to create a new and cheaper method of detecting glaucoma