MIT: non-contact heart rate measurement

Ming-Zher Poh, Phd candidate of MIT Media Lab, develops a non-invasive technique  to measure the heart rate by using camera, accurate to within 3 beats/minute.

• Low resolution camera
• Non-contact
• face motion tolerant
• measurement on more than 1 person at a time

Amazing!!! But how to achieve?????

Ans: By Software, Post image processing, Algorithm


Principle of measurement
"Different skin tones might yield a different baseline level of brightness, bu the pulse information is contained within the phasic information," he told.

"Volumetric changes in the facial blood vessel during the cardiac cycle modify the path length of the incident ambient light such that the subsequent changes in amount of reflected light indicate the timing of cardiovascular event." ~from Poh's paper.

What a small signal!!!
They use powerful image processing algorithm to extract signal from accumulated image.

Operation
Face Tracking (open library) ->RGB Frame (30s accumulated) -->FFT -->Signal Processing

FFT transform RGB frames to frequency domain data.
Compare these spectrum of RGB, correlation analysis the frequency component in spectrum to identify the frequency of blood volume pulse(BVP).

But
Avoid large, rapid motion