Gait Analysis—A Tool for Medical Inferences

Abstract

Gait analysis is a valuable tool for making medical inferences and improving the diagnosis and treatment of mobility issues. This project aims to leverage gait analysis in addressing two important challenges: detecting knock knees and monitoring patients with Parkinson’s disease for falls. The project proposes the integration of gait analysis with yoga therapy to provide a unique and effective approach for correcting knock knees. A web user interface is developed to enable individuals to access the system, receive accurate feedback on their gait, and access yoga postures tailored to target knock knees. Additionally, a fall detection system is designed to monitor patients with Parkinson’s disease and notify caregivers or guardians in case of a fall. The implementation involves utilizing deep learning models, such as OpenPose model, a widely adopted deep learning framework for pose estimation and MediaPipe, another recognized framework used for building multimodal applied machine learning pipelines, to analyze gait patterns and detect knock knees and falls. The project aims to empower individuals in improving their gait, correcting knock knees, and enhancing their physical health, ultimately improving their quality of life and well-being.