| Paper: | SLP-P14.5 | ||
| Session: | Speaker Recognition: Models and Methods | ||
| Time: | Thursday, May 18, 14:00 - 16:00 | ||
| Presentation: | Poster | ||
| Topic: | Speech and Spoken Language Processing: Speaker Verification | ||
| Title: | An Efficient GMM Classification Post-Processing Method for Structural Gaussian Mixture Model Based Speaker Verification | ||
| Authors: | R. Saeidi, Iran University of Science and Technology, Iran; H. R. Sadegh Mohammadi, Iranian Research Institute for Electrical Engineering, Iran; M. Khalaj Amirhosseini, Iran University of Science and Technology, Iran | ||
| Abstract: | In this paper a Gaussian mixture model (GMM) classifier, called GMM identifier, is proposed as an efficient post-processing method to enhance the performance of a GMM-based speaker verification system; such as Gaussian mixture model universal background model (GMM-UBM) and structural Gaussian mixture models with structural background model (SGMM-SBM) speaker verification schemes. The proposed classifier shows good performance while its computational load is almost negligible compared to the main GMM system. Experimental results show the superior performance of this post-processing method in comparison with a neural-network post-processor for such applications. | ||
©2018 Conference Management Services, Inc. -||- email: webmaster@icassp2006.org -||- Last updated Friday, August 17, 2012