Even the strongest of people can tweak their backs mid-lift or leave the gym with a newfound limp if they aren’t correctly doing their workout routine. Enter Perch, an MIT startup that’s developing a system of 3D cameras that tracks workouts and gives athletes critical data on their performance and health.

Perch is led by Bowen Baker, Jacob Rothman, Nate Rodman, Jordan Lucier and Zach Churukian. The co-founders initially came up with the idea for Perch because of their personal experiences as varsity athletes at the Institute and they recently completed the MIT Delta V accelerator program, an educational accelerator for MIT student entrepreneurs.

In a world where wearables and fitness trackers are now the norm, the Perch team wanted to cater to athletes who need more specific metrics to guide their training. They have a small amount of funding and a prototype device, which is now in testing with gyms and coaches.

“I have a Fitbit. I wore it for a while, but then I put it in a drawer and haven’t touched it since,” Rothman, who played baseball at MIT, told me. “Fitness wearables track your steps, heart rate and activity. But those metrics aren’t super valuable for people who go to the gym to do workouts involving strength training and conditioning, plyometrics or yoga.”

Perch aims to go beyond accelerometers that strength coaches and weightlifters use to monitor workout routines, tracking how fast a person is lifting, as well as form and power output.

The goal is to push athletes to their max performance while working out, while helping them avoid injury – a cause near and dear to Rothman. Three years ago, he had a herniated disk. He told me, “The doctors weren’t sure what caused it, but I was clearly not training in the right way. It could have been my form, that my body was overworked or that I wasn’t getting enough sleep… I want our technology to look at people, how they’re lifting and keep them safe.”

I want our technology to look at people, how they’re lifting and keep them safe.

Perch will be using machine learning when analyzing users’ workout data, so it will can predict injuries and predict which workout would render the best results for them. The venture wants its technology to give users’ the most accurate information as possible without impeding their workouts. Perch uses 3D cameras attached to the weight rack and a small computer that processes the images to analyze a person’s movements. It will also feature a touchscreen display to give users’ instant feedback on what they’re doing.

That feedback will include whether any detection of deviations from what’s normal for a lifter. For example, Perch will give an alert when an individual is lifting at a more strained pace than they typically do, recommending that they adjust the weight so they don’t hurt themselves. Users will also be able to set goals and Perch will track their progress. And there’ll be a web application where they can access all of their workout data.

To start, Perch will target the strength and conditioning market. They’ve tested their prototypes in college and commercial gyms and they’re looking to have more commercial facilities and weightlifters try them, to validate the commercial market and see how less seasoned athletes use the product.

For the most part, the startup has been bootstrapping. It has received small amounts in funding from MIT resources, like the Delta V Accelerator and MIT Sandbox.

Technical development continues, aiming toward a full release of the Perch product early next year. Once Perch has a good grasp on its joint-tracking technology, it will look to expand its product into other markets, like yoga and physical therapy.

“There’s a lot of avenues we can take,” Rothman said.

Update: This article initially listed two former team members instead of Jordan Lucier and Zach Churukian.