Understanding of Patient's Dangerous Behavior by Image Recognition

Abstract

In medical facilities, a surveillance system that monitors around a bed and detects some accidental movements of a patient is strongly in demand. The purpose of our research is to detect the patient″s dangerous behaviors such that he falls down from own bed in hospital room. In this paper, we propose a system for understanding patient″s dangerous behaviors by image recognition. This leads to reassurance of both the patient″s safety and reduce of health professions responsibility. We evaluate the effectiveness of the proposed system by the simulation with experimental data.