期刊名称：International Journal of Computer Science & Technology
印刷版ISSN：2229-4333
电子版ISSN：0976-8491
出版年度：2012
卷号：3
期号：4
页码：47-50
语种：English
出版社：Ayushmaan Technologies
摘要：In this paper, Zak transform is used for feature extraction of speaker identification system. Earlier Fourier transform is widely used to extract speaker specific voice characteristics, but Fourier transform cannot be used for frequency analysis that is local in time. This limitation of Fourier transform can be overcome by Zak Transform (ZT) which gives more detailed information about speaker’s voice by means of time frequency analysis of the speech signal. The probabilistic model for this feature set is created by Generalized Gaussian Mixer Model (GGMM) Model and model parameters are estimated using Expectation Maximization (EM) algorithm. The performance of the model is evaluated by drawing Detection Error Tradeoff (DET) curves and finding the minimum detection cost function. This speaker model performs better than the other existing speaker models.
关键词：Zak transform;EM algorithm;GGM;DET curves;Detection cost Function (DCF)
