Title :
Voiceprint Verification Based on Two-Level Decision HMM-UBM
Author :
Chen, Yan ; Hong, Qingyang
Author_Institution :
Inst. of Artificial Intell., Xiamen Univ., Xiamen, China
Abstract :
Voiceprint recognition is a new technology of biology authentication, and the most important task of it is to reduce both the false rejection rate (FRR) and false acceptation rate (FAR). To build a highly reliable voiceprint recognition system, this paper proposes three novel methods: two-level decision based on HMM-UBM, Model Distance and Dynamic Threshold. The evaluation results demonstrated that our methods had good performance.
Keywords :
hidden Markov models; speaker recognition; biology authentication; false acceptation rate; false rejection rate; model distance-dynamic threshold; two-level decision HMM-UBM; voiceprint verification; Artificial intelligence; Authentication; Automatic speech recognition; Biological system modeling; Hidden Markov models; Loudspeakers; Sequences; Spectrogram; Speech processing; Speech recognition;
Conference_Titel :
Information Science and Engineering (ICISE), 2009 1st International Conference on
Conference_Location :
Nanjing
Print_ISBN :
978-1-4244-4909-5
DOI :
10.1109/ICISE.2009.1342