In Russia, AI Has Learned to Diagnose Diseases by Gait
Researchers have developed an AI‑powered gait‑analysis system that identifies musculoskeletal disorders using computer vision and geometric modeling
A Breakthrough in Contactless Diagnostics
Scientists at the Novosibirsk Research Institute of Hygiene, part of Rospotrebnadzor, have developed and patented a new method for diagnosing musculoskeletal disorders based on a person’s gait using artificial intelligence. According to the authors, any changes in muscle condition—whether caused by impaired innervation or blood supply—manifest themselves in subtle changes in how a person walks. Over time, particularly with chronic conditions, these altered movement patterns become fixed.
How the System Works
The diagnostic process is non‑contact and requires only a camera and a computer, making it both accessible and scalable. The method captures a side‑view video of a person walking. A neural network identifies 25 key body points, calculates biomechanical parameters, and compares them to established norms. The system then determines the precise coordinates of body segments involved in locomotion.
Affordable Technology With Real Clinical Value
One of the system’s key advantages is its affordability: the hardware requirements are minimal, and the diagnostics do not require specialized medical devices. This significantly increases the accessibility of early screening for muscle‑related disorders. For patients, this means earlier detection, more accurate monitoring of disease progression, and the ability to evaluate treatment outcomes without costly procedures. For clinicians, it offers an objective, data‑driven tool that reduces subjectivity and improves diagnostic accuracy.
