Syncing Music With Emotions Using Computer Vision

With more consumers turning to digital platforms to find new music, music recommendation engines have become popular in recent years. Nevertheless, most recommendation engines rely on collaborative filtering methods or a user's historical listening habits, which might not fully represent the user's emotional state or mood. This study suggests a music recommendation engine integrating emotion recognition to deliver more individualized recommendations. The proposed system analyzes the user's tone of voice and facial expressions using machine-learning algorithms to identify enthusiasm, rage, sadness, and happiness. A playlist is then generated by mapping the user's emotional state to a series of musical parameters, such as key, pace, and rhythm. When the system was tested with a dataset of user-generated music playlists, it could predict the user's emotional state with more than 60% accuracy. Additionally, consumers expressed a generally favorable opinion of the suggested playlists, saying they had been introduced to new music they enjoyed and that the playlists had appropriately mirrored their current emotional state. The future scope of this study is to cater to diverse audiences with varying tastes in music and language.