EyeCare is a desktop application designed to assist ophthalmologists in diagnosing and analyzing retinal diseases using image processing and machine learning techniques. This project provides tools for disease classification, optic disk and optic cup segmentation, and retinal vessel segmentation.
EyeCare is an advanced desktop application designed to aid ophthalmologists in diagnosing and analyzing retinal diseases. It leverages state-of-the-art image processing and machine learning techniques to provide tools for classifying some of retinal diseases and segmenting the optic disk, optic cup and retinal vessels. The intuitive interface simplifies the analysis of retinal fundus images. Ophthalmic conditions such as diabetic retinopathy, glaucoma, cataract, and age-related macular degeneration are major causes of global blindness. Early diagnosis is crucial but challenging due to the labor-intensive nature of manual evaluations and the expertise required. Artificial intelligence (AI) offers a solution by enabling automated, precise identification of multiple fundus pathologies. However, challenges remain due to the presence of multiple concurrent diseases, scarcity of high-quality images, and image noise. The application features disease classification, optic disk optic cup segmentation, and vessel segmentation tools, generating diagnostic reports that aid in clinical decision-making. Developed with rigorous research and planning, EyeCare promises contributes to public health by improving access to advanced diagnostic tools
/docs/: Contains project documentation (abstract, chapters, figures)./notebooks/: Contains Kaggle notebooks for data processing, classification, and segmentation./data/: Holds aREADME.mdfile linking to the datasets used in Kaggle./results/: Experimentation results and logs./GUI/: Desktop application that use the trained Models .
This project is designed to be run in Kaggle Notebooks. You can find the Kaggle Notebooks for each part of the project:
- Data Processing and Augmentation: Pre-process and augment the fundus images from the ODIR-5K dataset.
- Classification: Train and evaluate the classification models for various retinal diseases.
- Segmentation: Perform optic disk, optic cup, and vessel segmentation.
To use the notebooks, you need to load the datasets from Kaggle
The datasets are not included in the GitHub repository. Make sure to download or link the datasets in your Kaggle Notebook environment.
- Import the Notebooks on Kaggle: Go to Kaggle and Import the provided notebooks.
- Link Datasets: In each notebook, you can link the Kaggle datasets required for training and testing the models.
- Run the Cells: Execute the cells to preprocess data, train models, and evaluate results.