Improvisational Dance Piano Accompaniment System Based on BP Neural Network

In recent years, the rapid development of artificial intelligence and its applications in various fields have led to the rapid development of computer music. Among them, algorithmic composition is an important research field of computer music, which aims to use computer technology to achieve the purpose of automatic composition in the process of composition. Impromptu accompaniment of piano has a very important position in dance art performance. Therefore, this article aims to study the improvisational dance piano accompaniment system based on BP neural network, and analyze the basic principles of BP neural network, the characteristics of BP neural network and dance performance. Based on the role of impromptu accompaniment, an impromptu piano accompaniment system is designed combined with the BP neural network algorithm. Finally, three professional composers were invited to conduct a systematic evaluation of the results of the accompaniment system. The evaluation concluded that the 10 types of results generated by the piano accompaniment system performed well in the accuracy of harmony, coherence, and technical complexity of the accompaniment, and the overall effect also meet the needs of this article.