Feature Extraction Method of Piano Performance Technique Based on Recurrent Neural Network

Abstract

In order to solve the problem of low efficiency in traditional feature extraction methods of piano performance techniques, a feature extraction method of piano performance techniques based on recurrent neural network is proposed. Analyze the types of piano playing techniques, and establish the hand model. On this basis, the hand action signals of piano performance are collected from the two aspects of finger key strength and hand action video image. Finally, the feature extraction of piano performance techniques is realized from the time domain and frequency domain. Through the comparison with the traditional extraction method, it is concluded that the extraction efficiency of the optimized design of piano performance technique feature extraction method has been significantly improved, and it has obvious application advantages in the identification of piano performance techniques.