Faculty of Information Technology, BUT

Publication Details

Speaker Recognition With Random Digit Strings Using Uncertainty Normalized HMM-Based i-Vectors

MAGHSOODI Nooshin, SAMETI Hossein, ZEINALI Hossein and STAFYLAKIS Themos. Speaker Recognition With Random Digit Strings Using Uncertainty Normalized HMM-Based i-Vectors. IEEE/ACM TRANSACTIONS ON AUDIO, SPEECH AND LANGUAGE PROCESSING, vol. 2019, no. 11, pp. 1815-1825. ISSN 2329-9290. Available from: https://ieeexplore.ieee.org/document/8759963?source=authoralert
Czech title
Rozpoznávání mluvčího s náhodnými sekvencemi číslovek pomocí i-vektorů založených na HMM s normalizací neurčitosti
Type
journal article
Language
english
Authors
Maghsoodi Nooshin (SHARIF)
Sameti Hossein (SHARIF)
Zeinali Hossein, Ph.D. (DCGM FIT BUT)
Stafylakis Themos (OMILIA)
URL
Keywords
text-dependent speaker verification, uncertainty compensation, HMM, i-vector
Abstract
In this paper, we combine Hidden Markov Models (HMMs) with i-vector extractors to address the problem of text-dependent speaker recognition with random digit strings. We employ digit-specific HMMs to segment the utterances into digits, to perform frame alignment to HMM states and to extract Baum-Welch statistics. By making use of the natural partition of input features into digits, we train digit-specific i-vector extractors on top of each HMM and we extract well-localized i-vectors, each modelling merely the phonetic content corresponding to a single digit. We then examine ways to perform channel and uncertainty compensation, and we propose a novel method for using the uncertainty in the i-vector estimates. The experiments on RSR2015 part III show that the proposed method attains 1.52% and 1.77% Equal Error Rate (EER) for male and female respectively, outperforming state-of-the-art methods such as x-vectors, trained on vast amounts of data. Furthermore, these results are attained by a single system trained entirely on RSR2015, and by a simple score-normalized cosine distance. Moreover, we show that the omission of channel compensation yields only a minor degradation in performance, meaning that the system attains state-of-the-art results even without recordings from multiple handsets per speaker for training or enrolment. Similar conclusions are drawn from our experiments on the RedDots corpus, where the same method is evaluated on phrases. Finally, we report results with bottleneck features and show that further improvement is attained when fusing them with spectral features.
Published
2019
Pages
1815-1825
Journal
IEEE/ACM TRANSACTIONS ON AUDIO, SPEECH AND LANGUAGE PROCESSING, vol. 2019, no. 11, ISSN 2329-9290
Publisher
IEEE Signal Processing Society
DOI
BibTeX
@ARTICLE{FITPUB12059,
   author = "Nooshin Maghsoodi and Hossein Sameti and Hossein Zeinali and Themos Stafylakis",
   title = "Speaker Recognition With Random Digit Strings Using Uncertainty Normalized HMM-Based i-Vectors",
   pages = "1815--1825",
   journal = "IEEE/ACM TRANSACTIONS ON AUDIO, SPEECH AND LANGUAGE PROCESSING",
   volume = 2019,
   number = 11,
   year = 2019,
   ISSN = "2329-9290",
   doi = "10.1109/TASLP.2019.2928143",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/12059"
}
Back to top