Detection and classification of electronic music signals using multiple models

Abstract

Abstract This paper briefly introduces the process of detecting and classifying electronic music signals, the support vector machine (SVM) classifier, and the convolutional neural network (CNN) classifier. Additionally, the CNN classifier was optimized by incorporating particle swarm optimization (PSO). The study then conducted simulation experiments to compare the performance of SVM, back-propagation neural network (BPNN), and the improved CNN. The noise immunity of the three algorithms was also tested. The results of the experiments demonstrated that the improved CNN algorithm outperformed the SVM and BPNN algorithms in recognizing music signals, regardless of the presence or absence of noise interference. Furthermore, the improved CNN algorithm exhibited the best noise immunity, followed by the BPNN and SVM algorithms. The interference of noise increased the time consumption of the detection and classification algorithm, and the detection of the improved CNN algorithm took the least time among these classifiers in the face of the same music signal.