Neural Acoustic-Phonetic Approach for Speaker Verification with Phonetic Attention Mask

Page view(s)
82
Checked on Aug 10, 2025
Neural Acoustic-Phonetic Approach for Speaker Verification with Phonetic Attention Mask
Title:
Neural Acoustic-Phonetic Approach for Speaker Verification with Phonetic Attention Mask
Journal Title:
IEEE Signal Processing Letters
Publication Date:
13 January 2022
Citation:
Liu, Das, R. K., Lee, K. A., & Li, H. (2022). Neural Acoustic-Phonetic Approach for Speaker Verification with Phonetic Attention Mask. IEEE Signal Processing Letters, 1–1. https://doi.org/10.1109/lsp.2022.3143036
Abstract:
Traditional acoustic-phonetic approach makes use of both spectral and phonetic information when comparing the voice of speakers. While phonetic units are not equally informative, the phonetic context of speech plays an important role in speaker verification (SV). In this paper, we propose a neural acoustic phonetic approach that learns to dynamically assign differentiated weights to spectral features for SV. Such differentiated weights form a phonetic attention mask (PAM). The neural acoustic-phonetic framework consists of two training pipelines, one for SV and another for speech recognition. Through the PAM, we leverage the phonetic information for SV. We evaluate the proposed neural acoustic-phonetic framework on the RSR2015 database Part III corpus, that consists of random digit strings. We show that the proposed framework with PAM consistently outperforms baseline with an equal error rate reduction of 13.45% and 10.20% for female and male data, respectively.
License type:
Attribution 4.0 International (CC BY 4.0)
Funding Info:
This research / project is supported by the National Research Foundation - RIE 2020 -Advanced Manufacturing and Engineering
Grant Reference no. : A1687b0033

This research / project is supported by the Agency for Science, Technology and Research - Feasibility Study Scheme
Grant Reference no. : FS-2021-001

This research / project is supported by the Agency for Science, Technology and Research - National Robotics Programme
Grant Reference no. : 192 25 00054
Description:
ISSN:
1558-2361
1070-9908