A Deep Learning Approach for YouTube Music Recommendation Based on Facial Emotion

In the contemporary digital era, the intersection of emotion recognition technology and music recommendation systems has paved the way for a transformative user experience. This research introduces an innovative “Emotion-Based Music Recommendation System,” designed to enhance the personalization of music consumption. Leveraging real-time facial expression analysis and a curated music dataset, the system dynamically adapts to users' emotional states, providing a sealess and engaging music discovery process. The integration of Facial Recognition (FR) technology enables the system to capture and interpret users' moods instantaneously, eliminating the need for manual classification of songs. This novel approach not only streaDlines the user experience but also holds significant potential for mitigating revenue loss associated with prolonged music search times. The research utilizes Python and PyCharm, employing Haar cascades for face detection and collaborative filtering and content-based filtering methods for music recommendation. The datasets utilized include FER, containing grayscale images of faces classified into seven emotions, and YouTube Music data, incorporating diverse attributes for mood-based classification. Through this synthesis of facial recognition and music datasets, the proposed system stands poised to redefine the landscape of music recommendation, emphasizing real-time adaptability and user-centricity.