Files in this item

FilesDescriptionFormat

application/pdf

application/pdfECE499-Sp2014-qianKaizhi.pdf (342kB)
(no description provided)PDF

Description

Title:Regularized Estimation of Gaussian Mixture Models for SVM Based Speaker Recognition
Author(s):Qian, Kaizhi
Contributor(s):Hasegawa-Johnson, Mark
Subject(s):speaker recognition
elastic net
sparsity
Gaussian mixture model
supervector
Abstract:Speaker adaptation based on the Universal Background Model (UBM) has become a standard approach for speaker recognition. A GMM supervector is constructed by normalizing and stacking the means of the adapted mixture components, which provides a compact representation of the speaker-dependent model in speaker recognition tasks. The estimation of the unknown GMM parameters is usually obtained by the method of maximum a posteriori estimation (MAP), which can be regularized to increase the model interpretability with insufficient training data. In this thesis, the speaker-adapted models are estimated using the MAP with L1-regularization, referred to as the elastic net, based on the assumption that the distinctions between any two speakers are sparse. Experiments on the NIST2008 speaker recognition evaluation task show error rate reduction with the elastic net.
Issue Date:2014-05
Genre:Other
Type:Text
Language:English
URI:http://hdl.handle.net/2142/55486
Date Available in IDEALS:2014-10-24


This item appears in the following Collection(s)

Item Statistics