Music Recommendation System Based on Facial Mood Detection

A person’s mood is mainly affected by the surrounding atmosphere and the type of company they keep around them. Its repercussions are seen in the person’s day-to-day life, socially and career-wise, leading to a positive or negative mindset. Our proposed system recommends appropriate songs to the customers of their liking based on their current dominant mood to emphasize this idea. Here, the customer’s face is detected using the OpenCV library, and different expressions are analyzed using CNN (Convolutional Neural Network) classifier and FER (Facial Expression Recognizer). In the end, appropriate songs get recommended to the customer based on their dominant mood detected. To make sure the songs recommended are of the customer’s liking, the customer selects their playlists at the starting. The system’s overall idea is to ensure that the customer gets to listen to their preferred songs to help them relate to their current feelings and help them reduce them to some extent by relaxing and stabilizing their emotions.