Music Emotion Recognition Based on Deep Learning: A Review

In recent years, with the development of the digital era, music emotion recognition technology has been widely used in the fields of music recommendation system, music classification, psychotherapy, music visualization, background music generation, smart home, and other applications of music emotion recognition, and has received attention from all walks of life. Especially the rapid development of artificial intelligence and deep learning, the music emotion recognition model using efficient deep neural network composition has become the mainstream model. This paper provides a more detailed overview of music emotion recognition, first introducing the background of music and emotion, and briefly summarizing the content of related works as well as the content framework. In the process, we also compare the similarities and differences in the content of other researchers’ reviews of related research areas. And in the middle section, we provide a detailed account of datasets, emotion models, feature extraction, and emotion recognition algorithms. Finally, we discuss the current challenges in music emotion recognition and explore future research priorities.