What if checking your pulse, respiration, and blood pressure were as easy as glancing at the mirror while brushing your teeth? A system being developed by a graduate student in the Harvard-MIT Health Sciences and Technology program could make that possible.
Ming-Zher Poh, SM ‘07, has demonstrated that he can get accurate pulse measurements from ordinary low-resolution webcam imagery—and he’s created a prototype that’s built into a mirror and displays results at the bottom. (Another researcher’s photographic pulse-detection system, described in a 2005 paper, required expensive camera equipment.) Now he’s working on extending the technology to measure respiration and blood-oxygen levels, and eventually blood pressure as well. A paper describing initial results of his work, carried out with Media Lab grad student Daniel McDuff and professor of media arts and sciences Rosalind Picard, SM ‘86, ScD ‘91, was published recently in the journal Optics Express.
Such a monitoring system could be especially helpful when attaching sensors to the body would be difficult or uncomfortable, Poh says, such as with burn victims or newborns. It could also be used for routine monitoring or for initial screening tests conducted over the Internet.
The system tracks pulse by measuring slight variations in brightness produced by the flow of blood through blood vessels in the face; with further processing, the researchers hope, the same data could be used to measure respiration and blood pressure. Software locates the face in the video image, and then the digital information from this area is broken down into separate data from the red, green, and blue sensors that produce the image. The brightness variations show up mostly in the green-light data, so separating the color channels helps make the signal stand out.
In tests, the team compared the readings they got from this setup with those from a commercially available, FDA-approved sensor, which measures pulse by bouncing infrared light off the skin. The results agreed to within about three beats per minute—even when the subject was moving a bit in front of the camera. The system can deal with the movement issue and variations in ambient lighting thanks to Poh’s adaptation of signal-processing techniques originally developed to extract a single voice from a roomful of conversations.
Keep Reading
Most Popular
Large language models can do jaw-dropping things. But nobody knows exactly why.
And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.
The problem with plug-in hybrids? Their drivers.
Plug-in hybrids are often sold as a transition to EVs, but new data from Europe shows we’re still underestimating the emissions they produce.
Google DeepMind’s new generative model makes Super Mario–like games from scratch
Genie learns how to control games by watching hours and hours of video. It could help train next-gen robots too.
How scientists traced a mysterious covid case back to six toilets
When wastewater surveillance turns into a hunt for a single infected individual, the ethics get tricky.
Stay connected
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.