Novosibirsk State Technical University (NSTU) has proposed a digital solution for the primary rehabilitation of patients with musculoskeletal disorders. The developers are creating a digital platform with adaptive algorithms based on computer vision and VR, which, according to their estimates, will be several times cheaper than analogues. Work on the application is being carried out on the basis of the youth laboratory of neuropsychological developments, created within the framework of the Priority 2030 program.
The solution proposed by the developers does not require specialized equipment, assumes a flexible approach system for patients with varying degrees of impairment and the possibility of remote monitoring.
According to the project leader, Svetlana Nedovyesova, Associate Professor of the Department of Psychology and Pedagogy at NSTU-NETI, Candidate of Biological Sciences, traditional rehabilitation methods require expensive equipment and inpatient monitoring. With the digital platform, nothing is required except a personal computer with a camera, and a person can do exercises at home. Progress is recorded in the form of the number of exercises that the patient must perform. The doctor can monitor the rehabilitation process directly or indirectly based on reports that are generated based on the completed session. Both the correctness of the exercises and errors are read. The exercise will not be counted until it is completed correctly.
"The developed solution is a software package for upper limb rehabilitation, which includes three key exercises: hand compression, arm lifting and towel folding. These exercises are selected based on recommendations for restoring fine motor skills and coordination in patients after a stroke or injury. The system uses a webcam and computer vision algorithms (MediaPipe) to track movements in real time, providing automatic repetition counting and adaptive feedback. Our solution does not require specialized equipment, assumes a flexible approach system for patients with varying degrees of impairment and provides the possibility of remote monitoring. The algorithms automatically adapt to each user," Svetlana Nedovyesova said.
The development uses standard hardware (webcam, PC) and open libraries (OpenCV, MediaPipe), which eliminates the cost of specialized hardware or licenses. This reduces the cost of the platform to the level of regular software, while analogues require investments from hundreds of thousands to millions of rubles. Thus, the solution becomes available for use in small clinics and at home.
The practical value of the development is the minimization of barriers to access to rehabilitation and the possibility of scaling standardized protocols to remote areas. The target group is patients with post—stroke conditions (the coverage potential in the Novosibirsk region alone exceeds 120,000 cases per year).
Currently, a demonstration prototype has been created: Python-based software shows exercises and captures the movements of a person's hand using a camera in order to improve the quality of exercises in the absence of correction by a specialist. The plans include testing the digital platform on patients after stroke and expert evaluation of the application by specialists in the field of medicine.