What can 3D imaging tell us about the health of the human body? Lots. The motion capture of one’s gait has the potential to serve as a digital biomarker that identifies disease progression and rehabilitation outcomes, while 3D tracking can identify problematic postures — such as spending a day hunched over a computer — that aggravate things including chronic neck pain. The ability to monitor and measure human movements can aid with detecting and managing a range of musculoskeletal disorders and much more. Great tech if you can get it, but 3D motion capture systems tend to be complicated (with multiple cameras and cables) and expensive. Until now. Researchers at Cornell University say they’ve come up with a way to capture body posture in 3D using a simple wrist band.
They’ve aptly named it BodyTrak, as the dime-sized wearable can track complete 3D body positions using only one miniature RGB camera worn on the wrist. To capture body silhouettes, the camera merely has to point to the wearer’s body to infer the 3D positions of 14 joints on legs, arms, the torso, and the head. In a recently published paper, Cornell researchers detail a study in which 9 participants wore BodyTrak wristbands while performing daily activities, including exercising, walking, and sitting. The results show that the cameras recorded accurate body positions with an average error of just 6.9 centimeters: less than 3 inches.
Much of BodyTrak’s success is thanks to its deep neural network that uses AI to train computers to learn from mistakes. The network can read rudimentary images and complete them to recreate 3D body positions virtually in real time. Researcher Hyunchul Lim says, “Our research shows that we don’t need our body frames to be fully within camera view for body sensing. If we are able to capture just a part of our bodies, that is a lot of information to infer to reconstruct the full body.”
And the research team thinks this technology is almost ready for integration with today’s commercially available wearables, which currently lack the battery life and camera power to handle full body sensing. But Cornell’s Cheng Zhang, senior author of the paper, says, “Since smartwatches already have a camera, technology like BodyTrak could understand the user’s pose and give real-time feedback. That’s handy, affordable and does not limit the user’s moving area.” While we’re not there yet, one day your smartwatch could almost literally track every move you make and offer a wealth of health information through what it captures.
