Implementation of Voiceprint Recognition Using SVM and SimpleVGGNet Method based on MFCC
Chengwei Lu, Lin Fu, Changxiao Liu, Yazhuo Zhao · 2022 IEEE International Conference on Advances in Electrical Engineering and Computer Applications (AEECA) · 2022
Voiceprint Recognition is a practical Recognition Technique that is a part of Biometric Technology. In recent years, both Machine Learning and Deep Learning have produced remarkable results in Pattern Classification, giving Voiceprint Recognition Technology a fresh strategy. As a result, utilizing the Support Vector Machine (SVM) in Machine Learning and Deep Learning framework of SimpleVGGNet, this study retrieved the MFCC characteristics of voiceprint and performed Voiceprint Recognition. We conducted experiments on two different Datasets respectively. These trials show that the Accuracy obtained by SimpleVGGNet is generally better than that of SVM, and SimpleVGGNet performed better in a large dataset, which can reach the Accuracy of 0.88. In the process of using different combinations of data as training data, we found that the closeness and gender of individuals within the data can also impact the quality of the model, with high intimacy between people reducing the accuracy of recognition, and the difference in vocal patterns between girls appearing to be smaller than the distinction between men.