Research On Monitoring System For Elderly People Living Alone Based On Facial Expression Recognition

As the domestic population ages,the growing demand for elderly care and home-based services is inconsistent with the shortage of nursing assistant.For the elderly people living alone,it is easy to cause fall accidents due to arthritis attacks and other painful causes,and fall accidents are prone to cause poor prognosis such as hip fractures.A monitoring system is provided to automatically send the lives of many elderly people living alone to the nursing assistant to identify the mental state of whether it is painful,and alert the nursing assistant before the fall accident may occur.The nursing assistant can specifically understand and intervene in the situation of the elderly who are being supervised by the guardian to avoid the abnormal mental state.Therefore,the development of an automated integrated monitoring system for such elderly people living alone is of great significance.This paper mainly studies an integrated monitoring system for elderly people living alone based on expression recognition.In order to prevent the occurrence of injury accidents,the system uses computer vision algorithms to express the facial expressions of the daily mental state of the elderly living alone,and classifies the results of the expression recognition by machine learning.When abnormal state of expression is detected(such as from time to time there is a painful expression)to promptly issue an early warning to the nursing assistant,and at the same time to increase the monitoring of harmful environmental factors in the living room to achieve remote comprehensive monitoring of the elderly living alone,to avoid accidents such as accidental falls.The monitored terminal module placed in the elderly living room alone is composed of a plurality of sensors,and monitors environmental data such as temperature and humidity,illuminance,and abnormal gas concentration in the room;In the mental state recognition of the elderly,based on the machine vision algorithm,the face of the elderly The expression features are used for facial expression recognition.If the painful expression is recognized,and the system promptly warns the nursing assistant to avoid serious accidents such as falls caused by abnormal mental state.For the recognition of painful expressions,this paper studies the face detection algorithm based on color features to quickly judge that the images captured by the camera contain appropriate face features and are suitable for expression recognition,so as to re-sampling and increase the monitoring efficiency.The facial key-points are extracted from the images containing facial features,and the alignment algorithms of these key-points are studied.The machine learning based on support vector machine is performed by using the key-points after alignment to achieve the purpose of mental state expression recognition.Finally,this paper uses Arduino and Raspberry Pi platform to integrate and test the above modules,verifying the practicability of the recognition algorithm,and using the WeChat public platform to establish the connection between the monitoring terminal and the monitored terminal,and realize the comprehensive monitoring of the elderly living alone.