Advanced algorithms of motion estimation for an elder at living room using Principal Component Analysis

Abstract

Objective: Development of telemedicine for an elder has been an important research area in an aging society, and effective Personal Emergency Response System(PERS) can provide exact medical decision and prompt treatment under emergency conditions. Previous studies have been focused on adapting troublesome sensors or passive calling system to monitor the old in their house. However, these previous systems might have limited applications due to its difficulties in usage and restraints in their daily activities, especially in the emergency. Methods: In this study, the real time algorithms using surveillance camera was developed to monitor their pose change, such as emergency and falling motion. To estimate the motion of elder people, this research use a ratio of eigenvectors of the Principal Component Analysis (PCA) technique. Results: In this system, no additional motion sensors or devices were applied to the object and it can be automatically controlled and monitor the old from a distance. It was found that this system can successfully monitor the old in living room regardless of surveillance camera angles and a silhouette size depending camera distance as using image processing and PCA. Conclusion: This algorithm was validated by experiments in a living room and this technique can be applicable to home monitoring and further applications. (Journal of Korean Society of Medical Informatics 13-3, 299-308, 2007)