Speaker verification using text-constrained Gaussian Mixture Models

D.E. Sturim; D.A. Reynolds; R.B. Dunn; T.F. Quatieri

doi:10.1109/ICASSP.2002.5743808

Proceedings of International Conference on Acoustics, Speech and Signal Processing (CASSP'02)

Speaker verification using text-constrained Gaussian Mixture Models

Year: 2002, Volume: 1, Pages: I-677-I-680

DOI Bookmark: 10.1109/ICASSP.2002.5743808

Authors

D.E. Sturim, MIT Lincoln Laboratory, Lexington, MA USA
D.A. Reynolds, MIT Lincoln Laboratory, Lexington, MA USA
R.B. Dunn, MIT Lincoln Laboratory, Lexington, MA USA
T.F. Quatieri, MIT Lincoln Laboratory, Lexington, MA USA

Abstract

In this paper we present an approach to close the gap between text-dependent and text-independent speaker verification performance. Text-constrained GMM-UBM systems are created using word segmentations produced by a LVCSR system on conversational speech allowing the system to focus on speaker differences over a constrained set of acoustic units. Results on the 2001 NIST extended data task show this approach can be used to produce an equal error rate of < 1 %.

Like what you’re reading?

Already a member?

Get this article FREE with a new membership!

Type-2 Fuzzy GMMs for Robust Text-Independent Speaker Verification in Noisy Environments
2014 22nd International Conference on Pattern Recognition (ICPR)
Speaker adaptation of tied-mixture-based phoneme models for text-prompted speaker recognition
Acoustics, Speech, and Signal Processing, IEEE International Conference on
Speaker identification using Gaussian mixture models based on multi-space probability distribution
Acoustics, Speech, and Signal Processing, IEEE International Conference on
Real-Time Speaker Verification Based on GMM-UBM for PDA
Embedded Computing, IEEE International Symposium on
Speaker Verification Using Adapted Bounded Gaussian Mixture Model
2018 IEEE International Conference on Information Reuse and Integration (IRI)
Methods to Improve Gaussian Mixture Model for Language Identification
2010 International Conference on Measuring Technology and Mechatronics Automation
Comparing maximum a posteriori vector quantization and Gaussian mixture models in speaker verification
Acoustics, Speech, and Signal Processing, IEEE International Conference on
Using prosodic and lexical information for speaker identification
Proceedings of International Conference on Acoustics, Speech and Signal Processing (CASSP'02)
A comparison between sequence kernels for SVM speaker verification
Acoustics, Speech, and Signal Processing, IEEE International Conference on
A Novel Embedded Speaker Verification on System on Chip
Fuzzy Systems and Knowledge Discovery, Fourth International Conference on

Speaker verification using text-constrained Gaussian Mixture Models

Authors

Abstract

Related Articles