Machine learning meets ultrasound. How osteopath Jana Maes aims to reduce the risk of injury and improve rehabilitation for top athletes.
What exactly happened on 24 July 2018 at the Hotel Beau Rivage Palace in Lausanne?
I attended a bar talk by the networking organization Swiss Ladies Drive. That’s when I met physicist Sophia Borowka. We got into conversation and talked about her plan to develop a massage robot, among other things. One thought led to another, and we finally came up with the idea that we are still pursuing today – a combination of ultrasound diagnostics with self-learning algorithms.
What medical problem does this address?
Today, for example, top athletes with soft-tissue injuries such as a suspected torn muscle fiber are given a magnetic resonance tomogram. Not only is it expensive, but it only provides snapshots. We rely on ultrasound and have developed a device that makes it possible to periodically take ultrasound images of a body part in the same way every time. Our algorithms compare these images and document the healing process.
You are working on the prototype right now. Are you in contact with any potential end-users?
We’re talking to two professional football clubs – Grasshopper Club Zurich and the Belgian club Bruges. Both are keen to test our system.
In the long term, your system should help not only rehabilitation but also prevention. How will this work?
In time, we will have a vast database of tissue images. Our AI platform will recognize patterns that indicate susceptibility to specific injuries. Athletes can then modify their training programs accordingly.