Date Thesis Awarded

5-2020

Document Type

Honors Thesis

Degree Name

Bachelors of Science (BS)

Department

Interdisciplinary Studies

Advisor

William Cooke

Committee Members

Keith Griffioen

Evie Burnet

Dennis Manos

Abstract

The World Health Organization reports that falls are the second-leading cause of accidental death among senior adults around the world. Currently, a research team at William & Mary’s Department of Kinesiology & Health Sciences attempts to recognize and correct aging-related factors that can result in falling. To meet this goal, the members of that team videotape walking tests to examine individual gait parameters of older subjects. However, they undergo a slow, laborious process of analyzing video frame by video frame to obtain such parameters. This project uses computer vision software to reconstruct walking models from residents of an independent living retirement community. Those subjects have agreed to be tested bi-annually and to report their fall history. Videos previously recorded demonstrate a variety of walks. Our procedures use several OpenCV-Python functions to detect, label, and follow markers that have been placed on the subjects’ shoes and knees. The trajectories followed by these markers allow us to generate walking models with gait parameters, such as the step height and the ankle dorsiflexion angle. This computer vision video analysis runs unsupervised to reduce processing time dramatically while enhancing the accuracy of a variety of measurements. Therefore, our data processing techniques will enable our kinesiology investigators to quickly generate a more extensive data set to learn how falling problems develop. This outcome will allow them to develop and to test exercises that can reduce those problems and prevent future falls for older subjects.

Keywords: computer vision, falls, gait, seniors

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Share

COinS