Patient Associated Motion Detection with Optical Flow Using Microsoft Kinect V2

Liang Liu, Sanjay Mehrotra

Abstract: This work describes our recent work of detecting the patient associated motions in a hospital room. After we had installed the designed Kinect V2 sensor-based health system in the hospital, we began to face big data challenges. The acquired data is big in both size and content. In this paper, we will propose a method to filter the big data using optical flow methods. As a result, we can discard the unnecessary data and quickly target on the data including valuable motion information about the patient. The proposed methodology facilitates the follow-up activity detection and serves for evaluating the amount of the movement the patient generates to allow the caregiver to improve the treatment plan.