Recognition and Visualization of Facial Expression and Emotion in Healthcare

Abstract

To make the SenseCare KM-EP system more useful and smart, we integrated emotion recognition from facial expression. People with dementia have capricious feelings; the target of this paper is measuring and predicting these facial expressions. Analysis of data from emotional monitoring of dementia patients at home or during medical treatment will help healthcare professionals to judge the behavior of people with dementia in an improved and more informed way. In relation to the research project, SenseCare, this paper describes methods of video analysis focusing on facial expression and visualization of emotions, in order to implement an “Emotional Monitoring” web tool, which facilitates recognition and visualization of facial expression, in order to raise the quality of therapy. In this study, we detail the conceptual design of each process of the proposed system, and we describe our methods chosen for the implementation of the prototype using face-api.js and tensorflow.js for detection and recognition of facial expression and the PAD space model for 3D visualization of emotions.