Design and Implementation of AI Based Efficient Emotion Detection and Music Recommendation System

In recent times, with the development and operation of big data, deep learning has taken more and more attention. As deep learning neural network, a convolutional neural network plays an extremely important part in face image recognition. In this paper, a combination of expression recognition technology of an convolutional neural network and automatic music recommendation algorithm is developed to identify a model that recognizes facial expressions and recommends music according to corresponding mood. The facial expression recognition model established in this paper uses FER2013 with a recognition rate of 62.1. After relating the matching expression, a content- grounded music recommendation algorithm is used to prize the point vector of the song and a cosine similarity algorithm is used to make the music recommendation. This exploration helps to ameliorate the practicality of the music recommendation system, and the affiliated results will also serve as a reference for the operation of the music recommendation system in areas similar as emotion parameter.