Application of GMM Model Adaptation Algorithm Based on Embedded System Architecture in Vocal Music Segmentation

With the development of embedded system, people have more and more in-depth research on its application in the field of vocal music. This paper mainly introduces a variety of optimization methods based on fuzzy neural network theory, improved algorithm and support vector machine technology to deal with multi-dimensional nonlinear planning problems in the process of vocal music product processing, and combines optimization solutions for linear weighting factors under different parameters to achieve the optimal segmentation model. Finally, numerical simulation experiments show that the scheme can better solve the complex feature points in the processing system, The embedded vocal music segmentation system based on GMM model adaptation algorithm has fast segmentation time, fast adaptation time and high matching degree. The system architecture has good stability and controllability.