Towards Developing a Mobile-Based Care (KidneyCare) for Patients with Kidney Diseases Using Ten-second Fingertip Video and PPG with Machine Learning

Abstract

In this manuscript, we describe the architecture and development of a non-invasive prototype- KidneyCare that captures the ten-second fingertip video using the smartphone camera and NIR LED. Then this fingertip video will be analyzed to estimate the creatinine, GFR, and CrCl levels by using ML models. KidneyCare acts as a regular, remote, self-monitoring, and noninvasive point-of-care by providing information about the user’s kidney condition. KidneyCare provides data visualization features and conducts as a data management platform for practitioners which promotes early diagnosis and initiation of an early treatment plan. With all the above-mentioned features, KidneyCare will be a powerful mHealth-based system for monitoring kidney conditions by estimating Creatinine, GFR, and CrCl levels non-invasively.