Abstract

AbstractHuman gait is essential for long-term health monitoring as it reflects physical and neurological aspects of a person’s health status. In this paper, we propose a non-invasive video-based gait analysis system to detect abnormal gait, and record gait and postural parameters framework on a day-to-day basis. It takes videos captured from a single camera mounted on a robot as input. Open Pose, a deep learning-based 2D pose estimator is used to localize skeleton and joints in each frame. Angles of body parts form multivariate time series. Then, we employ time series analysis for normal and abnormal gait classification. Dynamic time warping (DTW)-based support vector machine (SVM)-based classification module is proposed and developed. We classify normal and abnormal gait by characterizing subjects’ gait pattern and measuring deviation from their normal gait. In the experiment, we capture videos of our volunteers showing normal gait as well as simulated abnormal gait to validate the proposed methods. From the gait and postural parameters, we observe a distinction between normal and abnormal gait groups. It shows that by recording and tracking these parameters, we can quantitatively analyze body posture. People can see on the display results of the evaluation after walking through a camera mounted on a companion robot.KeywordsMultidimensional dataDeep learningHuman gait analysis

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call