Diabetic Retinopathy Blindness Detection
Developed a computer vision system to identify and categorize Diabetic Retinopathy stages, aiding ophthalmologists in preventing potential blindness. This system utilized standardized retina images and applied advanced preprocessing techniques, including Gaussian blur.
Experiments with ResNet models achieved up to 97% accuracy. Future enhancements will explore higher-resolution images and larger models to further improve accuracy and address data imbalances.
-
Machine Learning and Deep Learning: PyTorch, Scikit-learn
-
Quantum Machine Learning: Pennylane
-
Data Visualization: Matplotlib, Seaborn
-
Model: ResNet-152