We recently partnered up with the computer vision platform Physmodo. Here's everything you need to know about how computer vision is making an impact in movement health, and where we see it going.
What is computer vision, anyway?
If you follow the trends in digital health and fitness, the term "computer vision" is likely not a new one to you. Many of the hottest startups and fastest growing companies in the space are powered by computer vision, which enables individualized training, assessments and fitness at scale.
Peloton, Mirror, Tonal and all of the major players are either acquiring or building computer vision components of their system to increase personalization for (and engagement from) their users.
In a nutshell, computer vision allows cameras to use images and videos to collect meaningful information about objects moving in their environment. In relation to human movement, it's the first step in the role of digitizing the collection of data and information that our bodies provide when we demonstrate basic tasks (like a toe touch, for example).
We are still in the very early innings of bringing computer vision fully into consumer fitness and health experiences. Currently, movement assessments captured via computer vision are common in physical therapy products such as Kaia, Hinge and Sword Health to assist physical therapists in program creation for people who have experienced injury.
As we look forward, the emerging technology will put a robust, personalized fitness experience into the hands of anyone with a smartphone.
Just by setting up a phone camera, people will be able to complete a movement assessment or workout and receive real-time feedback and recommendations. How it is all implemented and used at scale remains to be seen.
Computer vision and movement health
Computer vision has the potential to be a core part of effective exercise recommendations in the near future, with its inherent ability to capture objective and unique information about human movement, leading to personalized exercise prescription.
The interesting piece of current computer vision technologies is the subtle differences in the software and capture technique. 3d modelling, room view, the number and type of cameras and the software powering them all play a role in distinguishing how different companies are innovating.
There has been a race to determine the best method for capturing and reporting movement data.
To date, the approach has largely been to find out how someone is moving, create metrics around that, and report them. While capturing human movement, correcting form and counting reps is very impressive technology, these features alone are not unique – nor are they going to innovate the exercise technology experience for the end consumer.
Manual assessments, especially hands-on assessments, will always have a place in relation to treating injuries or acute pain. Where computer vision fills a gap is in providing a high-throughput and objective measure of what is or isn't moving properly. The objective and consistent recommendations enabled by this technology go far beyond the everyday capabilities of the human eye, allowing a truly scalable solution to emerge.
The future of movement health
Current limitations of computer vision include the set up, ensuring appropriate distance from device (computer, tablet or phone), type of hardware, and ultimately delivering accurate, valuable and personalized insights beyond the superficial.
As a rep counter, scoring system or form feedback provider, the long-term stickiness of the technology remains to be seen.
As an assessment tool (whether the consumer is moving through an intentional assessment or is being passively captured during a workout), understanding how and where the body is moving in its environment provides an unprecedented opportunity for personalization.
Massive valuations for companies disrupting and scaling the physical therapy space (Hinge, Kaia, Sword) spiked in 2021 because of the work-from-home shift. This prompted fitness technologies to catch up by exploring (and in some cases, implementing) computer vision as a new feature product, like Peleton's introduction of the Guide, or Nautilus' acquiring Vay.
The next step with CV capture is figuring out what to do with that information. Understanding where something is in space doesn't actually tell us a lot about why the body is or isn't moving the way we'd like it to. The real value lies in synthesizing those data points.
To fully meet movement health demands, it will be essential for computer vision technologies to effectively translate their movement capture into meaningful recommendations for every individual.
movr is an assessment-based movement health technology company. Our goal is to help measurably improve musculoskeletal health with innovative technology solutions within an ever-changing health space.
Interested in learning more? Reach out here.