Design of a Real-Time and Dependable Hybrid Computer System for the Recovery From the Medical Cubital Tunnel Syndrome

Resumen

Computer systems are playing an increasingly vital role as core components in the design of dependable and medical cyber-physical systems (CPS) applications, which demand high levels of safety, reliability and performance. Cubital tunnel syndrome is a widespread upper limb peripheral nerve entrapment neuropathy that can cause significant discomfort and disability for the affected person. This critical medical condition frequently involves physical therapy activities aimed at reducing nerve compression. This paper presents a real-time and reliable hybrid computer system to optimize the efficacy of exercises for Cubital tunnel syndrome (CuTS) through automated video analysis, principles of computer vision, computer architecture and real-time dependable systems. The system compares exercises completed by the user with a reference image to verify proper joint angle alignment and improve therapeutic results. Specific joint angles are extracted and analyzed in real time by processing user videos and providing continuous feedback to users. The proposed computer system dynamically tracks angles and records them to facilitate consistent guidance and tracking of user progress through speech synthesis. OpenCV, MediaPipe, and Google Text-to-Speech (gTTS) were combined together for video processing, pose estimation, and text-to-speech conversion, respectively. Landmarks for specific joints are reliably detected and extracted using MediaPipe combined with precise custom angle calculation functions are consistently applied. The system compares calculated angles with predefined reference ranges and provides auditory feedback to users about the correctness of their joint angles, as well as guidance for adjustments. The proposed system is tested and validated to show promising result for this integration of hardware computer processing and speech synthesis for enhancing patient monitoring and rehabilitation.