Development of an intelligent system for the prediction of dangerous patient situations: A diploma – A startup

Abstract

In this Master’s thesis, we presented a project within the framework of monitoring patients staying in a hospital. Each room can contain several patients where a patient is associated with a bed with all the medical facilities. The rooms are occupied by surveillance cameras and the video sequences will be transmitted to our intelligent system called Aidy. The Aidy system first tries to detect object pairs of patients and beds in order to take an image that relates patient to bed using a classification model for object detection. These cropped images from the original image will be sent to a second classification model to predict the situation of patients. If the prediction indicates a dangerous case of a patient, a notification will be transmitted to a mobile application installed on the mobile device of a doctor or a nurse, which is connected to our intelligent system Aidy.