Mobile Gait Analysis

Abstract

This project aims to develop an extension to automated gait analysis that makes gait analysis available on smart devices. The alternative may serve as a baseline for future implementations that are cheaper, user-friendly and accessible to an ordinary smartphone or web browser. Accessibility of gait analysis on an application encourages people to check their walking patterns more regularly, and if the issue is very severe, they can take the next step of contacting a specialist. By collaborating with the Podiatry Department of the University of Malta and the Chinese Academy of Sciences Institute of Automation (CASIA), a considerable amount of gait data was acquired. The data consists of videos of people walking regularly or irregularly. But videos are not enough for the development of our system. The videos were inputted into a pose estimator whose goal was to outline the skeleton of the person throughout the video. Additionally, the pose estimator was modified to record the coordinates of the main joints concerning a gait cycle (hip, knee and ankle). These coordinates were then plotted as a scatter plot where the gait cycle is generated. With the coordinates extracted, kinematics were also extracted to create another model which detects different features for gait analysis. After the gait cycle of each video was extracted, the next step was to classify whether that gait cycle was either regular or irregular. This goal is achieved by passing the extracted data through the VGG16 architecture. The application was tested out on people which have either bad, good or slightly bad gaits to investigate the rigidity of the system. After a series of experiments, it can be concluded that the system performs with 94% accuracy just by using a mobile phone.