Personalized Stress Relief System Using Emotion Recognition and Machine Learning

In today’s fast-paced society, stress has become a pervasive issue, driven by demanding environments and societal pressures. Music, with its unique ability to connect with human emotions, has emerged as a powerful tool for stress relief. However, its effectiveness is limited when the selected tracks do not align with the listener’s emotional state. Existing music players lack the capability to dynamically select content based on the user’s emotional needs. This paper proposes an adaptive emotion-based stress management system that leverages facial expression analysis and real-time content recommendations. The system utilizes the Google Vision API to detect the user’s emotional state (e.g., sad, happy, angry, or neutral) and recommends personalized content, such as songs, movies, or books, tailored to their current mood. By simplifying emotion recognition and providing real-time, personalized recommendations, the system offers a practical and user-friendly solution for stress relief. The system achieves an accuracy of 92.19% in emotion detection, outperforming traditional methods like EEG-based systems, which are complex and resource-intensive. Key findings demonstrate the system’s ability to reduce stress levels across various emotional states, such as anger, fear, and sadness, while reinforcing positive emotions like happiness and neutrality. The system’s high accuracy, maintainability, and intuitive design enhance user satisfaction and accessibility.