Background: Hip osteoarthritis (OA) affects a significant percentage of the world population, impairing functionality, quality of life and increasing years of life lost due to disability, causing high costs and overwhelming healthcare systems. The integration of Artificial Intelligence (AI) and Computer Vision offers promise in the early detection of this condition. Objective: The study aimed to assess two key aspects: first, to evaluate the accuracy of Aictive's pose estimation computer vision system and its agreement with the VICON® system, and second, to appraise the diagnostic capacity using data obtained from the pose detector and diagnosis algorithm. Methods: 17 participants, comprising 10 individuals with diagnosed hip OA and 7 healthy subjects, completed a battery of 5 functional tests each. Data from both systems were concurrently collected to enable measurement comparisons. Results: The evaluation revealed that the pose detection system demonstrated an accuracy of 6.81° root mean squared error compared to the gold standard. Additionally, the pose detector's data yielded a remarkable 93% accuracy in classifying the pathology. Conclusions: This study underscores that Aictive's technology, rooted in computer vision and motion analysis, can serve as a valuable tool for early detection and precise diagnostic assistance in musculoskeletal pathologies, including hip OA.
|