Mood Based Music Generation

Abstract

The field of audio processing and generation has seen quite an interest in the rise of deep learning domain. This Project aims to generate music using Deep learning based on mood. It implies the nature of output to be generated. It is possible to generate music without the need for working with instruments artists. The process followed is to predict the expression of the human face as fast and as accurate as possible and generate the music based on their mood. Face recognition is done with CNN Music is composed initially using RNN but suffers from lack of global structure as it cannot keep track of events so approached discussed here is using RNN and LSTM which is a good mechanism which takes input from emotion generated to compose music with accuracy and timing Constraints.