AccuRate: iPhone Heart Rate Monitor

Monitoring one’s heart rate has numerous applications in the world of health and fitness. A person’s heart rate is, somewhat surprisingly, subtly visible when a light is shined through a thin body part, such as a fingertip. This subtle change in skin transparency can be detected by the cameras of modern smartphones utilizing the camera’s flash to illuminate the flesh of the finger. In recent years useful and viable mobile applications have emerged using this technique to estimate a user’s heart rate. These applications boast moderate to high accuracy but suffer from issues regarding noise and intermittency, especially when trying to estimate the user’s heart rate from less than thirty seconds of unbroken data. A recent paper published by Fan and Wang at the University of Pittsburgh presents BayesHeart, a new probabilistic model for accurately determining heart rate from camera data. BayesHeart uses a hidden Markov model and retains accuracy even with data intermittency and noise. Our goal is to replicate the BayesHeart model and pair it with a simple iOS mobile interface to create a superior heart rate monitor app.