Title :
Speaker verification based on prosodic features
Author :
Szymon Drgas;Damian Cetnarowicz;Adam Dąbrowski
Author_Institution :
Poznan University of Technology, Poland
Abstract :
In this paper accuracy of the speaker verification system based on prosodic features was evaluated. The efficiency of bigram models of the speaker rhythm was assessed. The results showed that in speech the rhythm carries speaker specific information. The optimal number quantization levels for of segments´ duration was determined.
Keywords :
"Random access memory","Context","Mel frequency cepstral coefficient","Context modeling","Histograms"
Conference_Titel :
Signal Processing Algorithms, Architectures, Arrangements, and Applications (SPA), 2008
Print_ISBN :
978-1-4577-1660-7
