•Led the project’s feature/color recognition & photo categorization efforts
•Researched and implemented computer vision techniques in Python using libraries like PyTorch, OpenCV, NumPy, and SciPy
•Significantly improved image processing by devising a segmentation algorithm with a U^2-Net for salient object detection, flexible threshold adjustments for masking, and conditional image cropping
•Engineered a deterministic color extraction technique to advance from k-means
•Developed a complete Python program for classification while adhering to a Scrum framework with 2-week sprints
This work was done in support of a project aiming to create non-invasive digital health tools to enhance infant nutrition practices. The mobile health tool will be utilized as part of an IRB-approved study in the lab of Dr. Elizabeth Johnson in the Division of Nutritional Sciences at Cornell in collaboration with the start-up, SimpliFed.
