Human-Computer Interaction Facial Recognition in Music Teaching System Based on Parameter extraction algorithm

Abstract

Speech recognition technology becomes the focus of attention of major research institutions in the world. People hope that machines can understand human voice commands and that they can control machines with voice. Therefore, the research and development of voice recognition will make people's lives easier in the near future. The feature parameter selected in this paper is the Mel-Zepstr coefficient. It analyzes the MFCC parameter extraction process in detail and proposes an improved MFCC extraction algorithm. Compared with the traditional MFCC extraction algorithm, the calculation amount is reduced by nearly 50% and the feature is improved. With the increasing use of mobile robots in the service industry, technologies for human-computer interaction (such as face recognition) have attracted more and more attention. In order to meet the real-time and efficiency requirements of real robot platforms, and to improve the accuracy of feature extraction under various lighting conditions and facial recognition technology positions, this paper, facial recognition and feature tracking are analyzed. In setting the voice recognition and face recognition module function of human-computer interaction, this article will combination of traditional music teaching and related new technologies, the facial recognition and voice recognition, strengthen the function of music teaching system, in the traditional classroom teaching, homework management, music practice, on the basis of the information to inform, From two aspects of voice and face, to meet the music classroom online when the need for information function. It can not only enable music teachers to observe each student's expression and action through facial recognition, strengthen expression management, but also enable music teachers to analyze each student's musical expression, so as to provide more comprehensive guidance for students.