Published Paper Details:

The internet and mobile technology have developed quickly and made it possible for us to freely access various music resources. While the music industry might lean more toward certain genres of music. Our current playlists in the music listening apps are static and user explicitly have to change it based on their likings. Music recommendation systems have become a crucial part of the music listening experience. However, most traditional recommendation systems rely on user behavior data or metadata, which fail to capture the emotional content of music. As music has a strong emotional impact on listeners, personalized music recommendation systems that take into account the emotional state of users are highly desired. For this purpose there were two independent models and existing systems, one for detecting the mood from the facial expression that is FER(Facial Emotion Recognition) and another was Music Classification Models which were used to recommend a song. We are attempting to combine these two systems to make efficient and accurate recommendation to the user. So in this project, we are going to develop a system which will capture the real time emotion of user by conversating with user or by other means and based on that emotion related songs will be recommended. We are going to categorize songs into the groups based on the categories like Happy, Sad, Energetic and calm. Then according to the captured emotion from the user, the songs related to that emotion will be recommended. Key factor of the song recommendation is classification of songs based on their acoustic features. In this way, user can listen the songs according to the mood.