Face Mood Juke Box

Music has a huge impact on our lives, it makes us feel incredible and can make every situation that little bit better. Lively musicgets us going during a workout, and chill songs help us relax and unwind. Recognising the relation between music and mood, music recommendation systems have been proposed. But these systems tend to take into account elements such as listening history or genre not always indicative of a user’s here-and-now emotional state.Here, we present such a novel jukebox recommendation system that will make use of sophisticated facial emotion detection. The system provides personalised song recommendations according to a user’s real-time emotions through the use of the AI-based facial emotion recognition tool, Gemini-provision. For instance, if someone is feeling happy, the system might interpret their facial expressions and propose an upbeat playlist to elevate their mood. Instead, if a user sounded depressed, the system would play soothing music. The system is also capable of incorporating the user's language preferences, thus ensuring that recommendations are not only consistent with their emotion, rather that it is also consistent with their geographical origin and their musical preferences. This paper describes the system, explains the functionality of Gemini-provision, and discusses its architecture. We also discuss the advantages of this approach, in terms of the possibility of allowing the generation of a stronger relationship between music and an emotionalstatus.