CamCare
Cameras for Smart Healthcare
Goal: Currently the AlarmNet system does not integrate cameras due to their power and privacy challenges. Research in this project focuses on using cameras efficiently and securely to fulfill the following objectives:
Tracking: The tracking GUI of AlarmNet can show only one person. But if there are multiple persons in a room or different persons in different rooms, then the GUI does not behave correctly. The CamCare subsystem incorporates this feature into the current system by using cameras in addition to the existing motion sensors.
Data Association: If multiple persons live in a home, then it becomes difficult to recognize activities of each individual. Because the sensors can only detect activities, but cannot tell us which individual triggered the detection. With the help of image analysis, CamCare can associate each activity to the original individual.
Verification: Cameras can be used to verify certain events recognized by the AlarmNet system (like fall detection). Without cameras there are always possibilities of errors in fall detection. But with images captured by the camera, it is possible to detect such events more accurately.
Challenges: The main research challenges are:
- To perform complex image analysis algorithms. We expect cameras to help us in precise indoor localization.
- To ensure acceptable response time.
- To ensure security and privacy.
Lead Researchers
Enamul Hoque, Anthony Wood