ICSpk: Interpretable Complex Speaker Embedding Extractor from Raw Waveform

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26230%2F21%3APU142958" target="_blank" >RIV/00216305:26230/21:PU142958 - isvavai.cz</a>

Výsledek na webu

<a href="https://www.isca-speech.org/archive/interspeech_2021/peng21_interspeech.html" target="_blank" >https://www.isca-speech.org/archive/interspeech_2021/peng21_interspeech.html</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.21437/Interspeech.2021-2016" target="_blank" >10.21437/Interspeech.2021-2016</a>

Jazyk výsledku

angličtina

Název v původním jazyce

ICSpk: Interpretable Complex Speaker Embedding Extractor from Raw Waveform

Popis výsledku v původním jazyce

Recently, extracting speaker embedding directly from raw waveform has drawn increasing attention in the field of speaker verification. Parametric real-valued filters in the first convolutional layer are learned to transform the waveform into time-frequency representations. However, these methods only focus on the magnitude spectrum and the poor interpretability of the learned filters limits the performance. In this paper, we propose a complex speaker embedding extractor, named ICSpk, with higher interpretability and fewer parameters. Specifically, at first, to quantify the speaker-related frequency response of waveform, we modify the original short-term Fourier transform filters into a family of complex exponential filters, named interpretable complex (IC) filters. Each IC filter is confined by a complex exponential filter parameterized by frequency. Then, a deep complex-valued speaker embedding extractor is designed to operate on the complex-valued output of IC filters. The proposed ICSpk is evaluated onVoxCeleb andCNCeleb databases. Experimental results demonstrate the IC filters-based system exhibits a significant improvement over the complex spectrogram based systems. Furthermore, the proposed ICSpk outperforms existing raw waveform based systems by a large margin.

Název v anglickém jazyce

ICSpk: Interpretable Complex Speaker Embedding Extractor from Raw Waveform

Popis výsledku anglicky

Recently, extracting speaker embedding directly from raw waveform has drawn increasing attention in the field of speaker verification. Parametric real-valued filters in the first convolutional layer are learned to transform the waveform into time-frequency representations. However, these methods only focus on the magnitude spectrum and the poor interpretability of the learned filters limits the performance. In this paper, we propose a complex speaker embedding extractor, named ICSpk, with higher interpretability and fewer parameters. Specifically, at first, to quantify the speaker-related frequency response of waveform, we modify the original short-term Fourier transform filters into a family of complex exponential filters, named interpretable complex (IC) filters. Each IC filter is confined by a complex exponential filter parameterized by frequency. Then, a deep complex-valued speaker embedding extractor is designed to operate on the complex-valued output of IC filters. The proposed ICSpk is evaluated onVoxCeleb andCNCeleb databases. Experimental results demonstrate the IC filters-based system exhibits a significant improvement over the complex spectrogram based systems. Furthermore, the proposed ICSpk outperforms existing raw waveform based systems by a large margin.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

Výsledek vznikl pri realizaci vícero projektů. Více informací v záložce Projekty.

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2021

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

ISBN

—

ISSN

1990-9772

e-ISSN

—

Počet stran výsledku

5

Strana od-do

511-515

Název nakladatele

International Speech Communication Association

Místo vydání

Brno

Místo konání akce

Brno

Datum konání akce

30. 8. 2021

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

000841879500103