APPLICATION OF ARTIFICIAL INTELLIGENCE IN THE EVALUATION AND MONITORING OF CORRECTIVE EXERCISES IN PATIENTS WITH FLATFOOT (VIDEO-BASED FOOT MOVEMENT ANALYSIS USING COMPUTER VISION

Paper ID : 1508-SPORTCONGRESS

Authors

Seyede rezvane Sajjadi *1, Mohamad Klantarian2 1اموزش پرورش 2وزارت علوم

Abstract

Introduction: Flatfoot (Pes Planus) is a common musculoskeletal deformity characterized by a collapsed medial longitudinal arch, which can alter lower-limb biomechanics and impair functional movement. Continuous and objective monitoring of rehabilitation progress is essential for optimizing intervention strategies and preventing secondary complications. Methods: A marker-less 2D pose estimation approach was employed to extract skeletal keypoints from smartphone-recorded videos of subjects performing standardized foot movements. From these keypoints, biomechanical indices such as the arch index, foot angles, and plantar pressure–related features were computed. A hierarchical deep learning architecture was then used to classify foot posture types and track temporal changes across rehabilitation sessions. Validation data included synchronized plantar pressure measurements and radiographic assessments as reference standards. Results: The proposed system demonstrated high accuracy in identifying flatfoot deformities and detecting biomechanical changes corresponding to therapeutic improvement. The video-based framework achieved comparable diagnostic performance to laboratory-grade 3D motion capture and pressure mapping systems, while significantly reducing cost and setup complexity. Conclusion: This study demonstrates the feasibility of using AI-driven video analysis for objective and continuous monitoring of flatfoot rehabilitation. The approach provides a low-cost, accessible, and scalable solution suitable for clinical and tele-rehabilitation applications, supporting data-driven and personalized treatment design.

Keywords

Flatfoot, Computer Vision, Pose Estimation, Deep Learning, Rehabilitation Monitoring, Biomechanics

Status: Abstract Accepted (Poster Presentation)