Deep Speaker Embeddings Based Online Diarization

This paper describes our experiments with the Unbounded Interleaved-State Recurrent Neural Network (UIS-RNN) model for development of an online diarization system. For this task several UIS-RNN models based on different speaker embeddings extractors were trained. These systems were evaluated in term...

Full description

Saved in:
Bibliographic Details
Published inSpeech and Computer Vol. 13721; pp. 24 - 32
Main Authors Avdeeva, Anastasia, Novoselov, Sergey
Format Book Chapter
LanguageEnglish
Published Switzerland Springer International Publishing AG 2022
Springer International Publishing
SeriesLecture Notes in Computer Science
Subjects
Deep learning
Diarization
Online diarization
Online AccessGet full text
ISBN3031209796
9783031209796
ISSN0302-9743
1611-3349
DOI10.1007/978-3-031-20980-2_3

Cover

More Information
Summary:This paper describes our experiments with the Unbounded Interleaved-State Recurrent Neural Network (UIS-RNN) model for development of an online diarization system. For this task several UIS-RNN models based on different speaker embeddings extractors were trained. These systems were evaluated in terms of Diarization Error Rate (DER) metric on public and private test datasets. Also systems were tested on real dialogue data recorded in a bank office. Proposed online models outperform standard offline Agglomerative Hierarchical Clustering (AHC) approach and are compatible with the state-of-the-art Bayesian HMM (VBx) offline method.
ISBN:3031209796
9783031209796
ISSN:0302-9743
1611-3349
DOI:10.1007/978-3-031-20980-2_3