Scientists create an AI fitness trainer that might actually work

Soon after you start researching exercise routines, ads start popping up. These days, it’s mostly videos of people claiming they got insanely fit in no time with the help of an AI trainer.

But I’m sure you haven’t heard about this AI-based training technique just yet.

Researchers from the Universidad de Concepción in Chile are proposing a technology-based method for optimal training and injury prevention.

They’ve tested their method with weightlifting but will start to experiment with more sports soon.

They say a method combining thermal imaging and AI processing is supposed to enable targeted training and recovery.

Thermal images are used to track muscles, detect their activation, and highlight areas of strain or fatigue.

Researchers say that by continuously tracking even the smallest changes in body temperature, athletes and their coaches can better understand what happens to the body throughout the entire exercise. This should allow optimized training and prevent injuries.

“We also aim to refine this technology for applications in the healthcare field by providing specialists with thermal data and information about body position, which could contribute to more effective rehabilitation after an illness or injury,”

Laura Viafora, a research team lead, said.

Muscle exertion during weightlifting can be tracked by simply attaching a thermal camera to a smartphone. It doesn’t even have to be an expensive thermal camera, they say.

To identify individuals, their body parts, activated and strained areas, researchers developed data processing algorithms and also used Google MediaPipe artificial intelligence software.

MediaPipe is an open-source project for developers that provides libraries and tools for applying AI and machine learning techniques to applications.

“We can see whether one side of the body is being exerted more than the other, and thermal images make it easier to identify individuals compared to conventional images, which can be influenced by background colors, clothing, and skin tone,” Viafora said.

You won’t see any ads for this AI trainer just yet.

After tests with weightlifting showed promising results, scientists want to move to other disciplines in both conventional and paralympic sports. They will also work on refining the algorithm so it could be more useful, for example, telling an athlete whether the exercise was performed correctly.

Their research was published in a peer-reviewed scientific journal Applied Optics.