Privacy-Preserved Video Monitoring Method with 3D Human Pose Estimation

With the fast growth of aging population and the spread of various chronic diseases such as heart disease and arthritis among older adults, elderly care has become an urgent topic facing today’s society. Consequently, technologies mediated remote care has become a widely-used method, with the significant promise of reducing cost and improving the efficiency and quality of healthcare. Yet, most remote-caring technologies, especially surveillance video based remote care, face the challenge of privacy issues. For addressing this issue, this paper proposes a privacy- preserved remote care method. Specially, we use ROMP to extract the 3D human model of the elderly in the surveillance video, and use KNN pose estimation algorithm to detect the potential abnormal behaviors. Compared to existing methods, which mainly replace the privacy information with totally different contents, our method not only protects the personal privacy information of the elderly, but also provides clear and identifiable posture information which could better support remote care.