Real-Time Emotion Detection and Mental Health Monitoring for Remote Employees

Abstract

Remote work provides flexibility, but it also magnifies concerns about employee well-being in terms of emotional instability, stress, and mental health. The traditional way of monitoring mental health and performance evaluation is flawed for a remote working environment since it lacks both real-time ability and integration of emotional and physiological data. This acknowledges the need for an Intelligent System that will help in emotion detection, mental health monitoring, and personalized intervention to enhance the productivity and well-being of remote employees.

In this proposed solution, state-of-the-art machine learning algorithms, including Convolutional Neural Networks (CNN) for emotion classification in real time, and Random Forest Classifiers for predicting stress levels through HRV data, were employed. Natural Language Processing (NLP) is used to detect psychological patterns in terms of sentiment. It uses a combination of heart rate data, webcam-based facial recognition for mental health monitoring. Developing a web-based application serves as the primary interface for the monitoring and emotion detection of a user's mental state for employees working remotely in real-time, and includes the visualization of emotion history, stress trends analysis, and automated personalized music recommendations & therapy suggestions. A lightweight mobile companion app interfaces with to provide real-time heart rate data.

The prototype achieved an overall accuracy of 92%. It was evaluated using a dataset of over 20,000 facial images, heart rate data, and stress prediction based on these inputs. The final output included music recommendations and therapy suggestions. Real-time evaluations demonstrated the system’s responsiveness and usability across various platforms, highlighting its potential to significantly improve remote employee well-being and enhance organizational performance.