Retinal disease screening in low-resource settings is limited by a simple infrastructure gap: fundus cameras cost tens of thousands of dollars and require trained operators, making population-scale screening impractical where the burden is highest. Clarifeye was built to close that gap — a low-cost, smartphone-based fundus imaging system targeting scalable diabetic retinopathy and glaucoma screening in LMICs. As co-creator and technical lead, I shaped both the concept — including the smartphone-mount imaging approach and end-user workflow — and owned project execution end-to-end. The system was designed around cost, portability, and reproducibility rather than lab-grade performance, with early market research informing the specific gaps we were targeting.
My core technical contribution was a modular medical image preprocessing toolkit built to support reproducible experimentation across medical imaging datasets. It abstracts common steps — dark border cropping, circular fundus cropping, CLAHE contrast enhancement, resizing — into plug-and-play transforms with consistent output structure and a Jupyter-based visual preview workflow. In medical imaging, preprocessing decisions often drive downstream model performance, and this toolkit makes those decisions explicit, inspectable, and reusable beyond Clarifeye itself. The project is early-stage and ongoing, with the infrastructure now in place for dataset expansion and model development. GitHub