Design, Implementation, and Performance Analysis of In-Home Video based Monitoring System for Patients with Dementia

Abstract

Dementia is a major public health problem affecting 35 million people in USA. The caregivers of dementia patients experience many types of physical and psychological stress while dealing with disruptive behaviors of dementia patients. This will also result in frequent hospitalizations and re-admissions. In this project we design, implement, and measure the performance of an advanced video based monitoring system to aide the caregivers in managing the behavioral symptoms of dementia patients. The caregivers will be able to easily capture and share the antecedents, consequences, and the function of behavior, through a video clip, and get the real-time feedback from clinical experts. Overall the system will help in reducing the hospital admission/readmission, improve the quality of life for caregivers, and in general result in reduced cost of health care systems. System is developed using python scripts, open source web frameworks, FFmpeg tool chain, and commercial off-the-shelf IP camera and mini-PC. WebRTC is used for video based coaching of caregivers. A framework has been developed to evaluate the storage and retrieval latency of video clips to public and On-premise clouds, video streaming performance in LAN and WLAN environments, and WebRTC performance in different types of access networks. InstaGENIrack, a GENI rack in KU is used as on-premise cloud infrastructure for the evaluation. OpenSSL utilities are employed for secured transport and storage of captured video clips. We conducted the trials in Google fiber ISP in Kansas city, and compared the performance with other traditional ISPs.