Bayesian HMM clustering of x-vector sequences (VBx) in speaker diarization: Theory, implementation and analysis on standard tasks

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26230%2F22%3APU142975" target="_blank" >RIV/00216305:26230/22:PU142975 - isvavai.cz</a>

Výsledek na webu

<a href="https://www.sciencedirect.com/science/article/pii/S0885230821000619" target="_blank" >https://www.sciencedirect.com/science/article/pii/S0885230821000619</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1016/j.csl.2021.101254" target="_blank" >10.1016/j.csl.2021.101254</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Bayesian HMM clustering of x-vector sequences (VBx) in speaker diarization: Theory, implementation and analysis on standard tasks

Popis výsledku v původním jazyce

The recently proposed VBx diarization method uses a Bayesian hidden Markov model to find speaker clusters in a sequence of x-vectors. In this work we perform an extensive comparison of performance of the VBx diarization with other approaches in the literature and we show that VBx achieves superior performance on three of the most popular datasets for evaluating diarization: CALLHOME, AMI and DIHARD II datasets. Further, we present for the first time the derivation and update formulae for the VBx model, focusing on the efficiency and simplicity of this model as compared to the previous and more complex BHMM model working on frame-by-frame standard Cepstral features. Together with this publication, we release the recipe for training the x-vector extractors used in our experiments on both wide and narrowband data, and the VBx recipes that attain state-of-the-art performance on all three datasets. Besides, we point out the lack of a standardized evaluation protocol for AMI dataset and we propose a new protocol for both Beamformed and Mix-Headset audios based on the official AMI partitions and transcriptions.

Název v anglickém jazyce

Bayesian HMM clustering of x-vector sequences (VBx) in speaker diarization: Theory, implementation and analysis on standard tasks

Popis výsledku anglicky

The recently proposed VBx diarization method uses a Bayesian hidden Markov model to find speaker clusters in a sequence of x-vectors. In this work we perform an extensive comparison of performance of the VBx diarization with other approaches in the literature and we show that VBx achieves superior performance on three of the most popular datasets for evaluating diarization: CALLHOME, AMI and DIHARD II datasets. Further, we present for the first time the derivation and update formulae for the VBx model, focusing on the efficiency and simplicity of this model as compared to the previous and more complex BHMM model working on frame-by-frame standard Cepstral features. Together with this publication, we release the recipe for training the x-vector extractors used in our experiments on both wide and narrowband data, and the VBx recipes that attain state-of-the-art performance on all three datasets. Besides, we point out the lack of a standardized evaluation protocol for AMI dataset and we propose a new protocol for both Beamformed and Mix-Headset audios based on the official AMI partitions and transcriptions.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

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

Projekt

<a href="/cs/project/GX19-26934X" target="_blank" >GX19-26934X: Neuronové reprezentace v multimodálním a mnohojazyčném modelování</a><br>

Návaznosti

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

Rok uplatnění

2022

Kód důvěrnosti údajů

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

Název periodika

COMPUTER SPEECH AND LANGUAGE

ISSN

0885-2308

e-ISSN

1095-8363

Svazek periodika

71

Číslo periodika v rámci svazku

101254

Stát vydavatele periodika

GB - Spojené království Velké Británie a Severního Irska

Počet stran výsledku

16

Strana od-do

1-16

Kód UT WoS článku

000761599000019

EID výsledku v databázi Scopus

2-s2.0-85109214006