Human emotion detection is an immediate need so that modern AI systems can simulate and measure facial responses. It also has advantages in the identification of intent, promotion of products, and security verification. Real-time emotion recognition from images and video is a very simple task for the human eyes and brain, but it proves to be very difficult for machines and similar machine learning tools. Basically, Image Processing techniques are needed for feature extraction supported by a reliable database trained in a Machine Learning model for the system. Several machine learning algorithms and tools such as Convolutional Neural Networks, OpenCV, Deep Learning, Eigen values, and Eigen vectors are suitable for this job and I intend to use these methods in this project. For machines development of different modules and then training them using various images and real-time feed is essential. Various leading institutions and researchers have trained their own models for an accuracy of approximately 50% and above. This project explores the ML algorithms as well as emotion detection techniques which would help us in the correct identification of human emotion and furthermore, implementation of the system into a useful application of a music recommendation system.