Iris Tracking and Heart Beat Monitor

Affiliation: Project performed under Wolfram New Kind of Science Summer School 2013.

Project Title: Assaying Patient Vital Signs through Facial Image Analysis

Description: The development of non-intrusive methods for assaying subjects' vital signs is of interest for scenarios where there is too large of a population to deploy traditional instruments or where there is a stigma against traditional instruments. We investigate the use of computer vision algorithms for determining a person's eye movement and heart rate from webcam-quality video feeds. Our analysis includes properties traditionally associated with the level of activity, such as pigmentation and trajectories of facial features. We envision that these techniques will serve as a potential alternative to medical techniques such as pulse oximetry and sphygmomanometers, paving the way for crowd-sourced, large-scale public health studies and real-time population medical analytics.

Iris Tracking

We locate the position of the subject's pupils and two suitable reference points on the side of the head (e.g. the ears). By tracking all four of these points, we can determine the movement of the subject's irises relative to the head in real-time.

Heart Beat Monitor

We locate the position of the subject's right cheek (from the subject's perspective) and measure its pigmentation over a period of 20 seconds. We then analyze the frequency domain of the pigmentation levels to determine the frequency of the subject's pulse in order to calculate heart rate.

The project notebook can be downloaded here. To view the Mathematica notebook file, you will need Mathematica Player (free) or a copy of Mathematica (discounted price for students, click here).