Your arm lifts a cup from the shelf — a familiar motion you don't notice until muscles tire. Then the trajectory changes: the elbow shifts slightly, the hand trembles. A doctor usually catches these changes by hand, feeling the patient's muscles. But what if a diagnosis could be made from video?
A team of researchers created an algorithm that watches arm movement from the outside and compares it to a digital twin of the person. This twin isn't just a cartoon skeleton but a detailed model where each muscle can contract and fatigue according to the laws of physics. The algorithm 'fatigues' different muscle groups in the simulation one by one and sees which version of fatigue most closely matches the patient's real movement. This way, it finds the true culprit — for example, fatigue in the elbow flexors, not the shoulder.
Tests on real people showed the algorithm reliably distinguishes fatigue in different muscle groups. This matters not only for hospitals. In the future, such an approach could help assistive robots notice human fatigue in time and adapt to prevent injuries. Contactless diagnosis removes barriers — no need to travel to a doctor, just a camera.
The cup still rises to the lips, but now an impassive observer watches the movement, seeing fatigue before the arm itself feels it.
A tired muscle not only weakens but also relaxes more slowly after contraction — one of the key signs the algorithm catches.
In the film 'Elysium', a medical pod instantly scans the body and cures any disease — our algorithm takes a tiny step toward diagnosis without touch.