A collection of projects I've built and contributed to.
academic-projectActive
Diabetic Retinopathy Detection
Deep learning pipeline that classifies diabetic retinopathy severity (5 stages) from retinal fundus images, achieving a 0.94 Quadratic Weighted Kappa using a fine-tuned EfficientNet-B5 model.
PythonPyTorchOpenCVEfficientNet
academic-projectActive
Gait Recognition System
Behavioral-biometrics system that identifies individuals from walking-pattern data across a 10,000+ record dataset, using an SVM classifier that reached 91% accuracy.