Published Paper Details:

Music plays a central role in everything we do. It is a way to express our feelings; it can bring us comfort and relieve stress. Traditional recommendation systems for music typically consider a user's prior activity (listening history, searching patterns) and use predefined playlists. As a result, these systems do not necessarily reflect the current emotional state of the user, creating a mismatch between the recommendations provided and what users want to listen to at that moment. To resolve this problem, we are proposing a project called The Face-Based Song Recommendation System. Through the use of computer vision and artificial intelligence, our system will utilize facial expression recognition to deliver recommendations in real time. Users can capture their facial expressions with webcams or smartphones, and the system will utilize machine learning algorithms to classify their facial expressions into several emotions (happy, sad, angry, surprised, neutral, and afraid). After identifying the user's emotion, an engine will map the emotion to a genre or playlist that corresponds to that emotional state (for example, an uplifting playlist would be suggested for someone who is feeling sad, while a calm and peaceful playlist would be recommended for someone experiencing stress or anger). This innovative, emotion-based recommendation method provides a much more personalized experience than typical recommendation systems, which only provide a set of playlists with no personalization. This project will demonstrate the importance of using emotion-aware computing in song recommendations.