EmotionSense: A Real-Time Emotion Monitoring and Analysis System Based on Physiological Signals

Real-time emotional monitoring and visualization is an increasingly important topic in the field of human-machine interaction. Most current emotion monitoring studies focus on visual- or speech-based methods. However, due to multiple reasons, these methods can be unstable and unreliable in realworld applications. Compared to visual and speech data, physiological signals offer a more direct and involuntary reflection of emotional states, as they are less influenced by conscious control. However, capturing and interpreting physiological data for emotion analysis in real-time is challenging. To address this issue, this paper presents EmotionSense, a real-time emotion analysis system based on physiological data, which includes realtime signal processing for emotion recognition and a user-friendly emotion visualization interface. For physiological data, the developed system utilizes a wearable sensor to capture key physiological signals including heart rate, electrodermal activity, temperature, motion, and photoplethysmography in real-time. Specifically, this paper introduces (1) the development of a realtime emotion analysis algorithm that maps physiological signals to dimensional emotional arousal-valence-intensity values, discrete emotional states, and human stress levels; (2) the creation of a realtime emotion visualization interface displaying emotion analyzing results for enhancing user engagement and understanding; and (3) evaluation experiments to demonstrate the effectiveness of the developed systems. The results highlight the system's ability in analyzing human emotions in real-time, and guide future improvements for biofeedback systems focused on enhancing human-machine interaction.