The facial expressions of a user can be used to infer the user's state of mind or emotion. These expressions can be generated using the live feed provided by the system's camera. There is a significant amount of research being carried out in the field of computer vision and machine learning (ML), which is the process of teaching computers to recognise different human emotions and states of mind. The ability to read people's feelings is made possible by a number of different approaches provided by machine learning. The use of the mobilenet model with Keras is one method of this kind. This method generates a trained model with a small size and simplifies the process of Android-ML integration. People frequently turn to music as a technique for regulating their moods, particularly when they want to shift a poor mood, improve their level of energy, or alleviate tension. Additionally, if you listen to the appropriate genre of music at the appropriate moment, this may be beneficial to your mental health. Therefore, there is a close connection between the emotions of humans and music. In the system that we have proposed, a mood-based music player will be constructed. This player will do real-time mood detection and will make song recommendations depending on the detected mood. This adds a new function to the standard music player applications that are already pre-installed on our mobile devices. Customer satisfaction is an essential advantage that may be gained through mood detection. The purpose of this system is to analyse the user's image, make an educated guess as to the user's expression, and then make song recommendations that are relevant to the user's current state of mind.

Machine Learning, Deep Learning, mobilenet, keras, Haar face detector, spotify.