Using synchronized video captured by a pair of smartphones, scientists have created an open-source motion capture application that collects human movement data and quickly analyzes it through an artificial intelligence system, which is then used for clinical rehabilitation, pre-surgery planning and disease diagnosis at a cost of only 1% of traditional technology.
OpenCap, created by Stanford researchers with funding from the National Institutes of Health, uses two calibrated iPhones to work together to measure human movement and the underlying musculoskeletal mechanisms that power the movement. What's more, it's faster than traditional techniques used to gather the same information and costs a fraction of the $150,000 equipment that uses about eight high-tech cameras in specialty clinics.
"OpenCap democratizes human movement analysis," said senior author Scott Delp, a professor of bioengineering and mechanical engineering at Stanford University. "We hope it will put this once out-of-reach tool in the hands of more people."
This analysis can inform the treatment of patients with movement problems, help clinicians develop surgical plans, and review the effects of various treatments. It also has the potential to be used to screen for disease, since changes in gait or balance may not be easily observed during a routine physical exam.
They conducted OpenCap tests on 100 participants, recording videos that were then analyzed via web-based artificial intelligence to assess muscle activation, joint loading and joint motion. Data collection for all 100 participants took less than 10 hours, and analysis results were returned within 31 hours. Data collection takes about 10 minutes per person, and processing starts automatically on a cloud platform that is free for researchers to use.
"It would have taken expert engineers days to collect and process the biomechanical data that OpenCap provides in minutes," said Scott Uhlrich, research director of the Stanford Human Performance Laboratory. "We collected data from 100 people in less than 10 hours, which previously would have taken a year."
The data examines the body's "landmarks" -- knees, hips, shoulders and other joints -- and how they move in three-dimensional space. It then uses complex physics and biological models of the musculoskeletal system to assess how the body moves and what forces are at play. This provides important information about joint angles and loads, and it can even tell you which specific muscles are being activated.
The researchers believe this data collection and deep learning analysis will usher in a new era of biomechanical research.
"The human genome is now available, but this will really be a movement genome that quantitatively captures the full sequence of human movement," Delp said. "We hope that, in the process of democratizing human movement analysis, OpenCap will accelerate the incorporation of key biomechanical metrics into a growing number of studies, trials, and clinical practices to improve outcomes for patients around the world."
The research was published in PLOS Computational Biology. For more information, watch this video of Stan Dodd's team demonstrating OpenCap.