A study on end-to-end speaker diarization system using single-label classification

Jaehee Jung; Wooil Kim

doi:10.7776/ASK.2023.42.6.536

All Issue

2023 Vol.42, Issue 6 Preview Page Next Page

Research Article

A study on end-to-end speaker diarization system using single-label classification 단일 레이블 분류를 이용한 종단 간 화자 분할 시스템 성능 향상에 관한 연구

30 November 2023. pp. 536-543

PDF XML

Abstract

Speaker diarization, which labels for “who spoken when?” in speech with multiple speakers, has been studied on a deep neural network-based end-to-end method for labeling on speech overlap and optimization of speaker diarization models. Most deep neural network-based end-to-end speaker diarization systems perform multi-label classification problem that predicts the labels of all speakers spoken in each frame of speech. However, the performance of the multi-label-based model varies greatly depending on what the threshold is set to. In this paper, it is studied a speaker diarization system using single-label classification so that speaker diarization can be performed without thresholds. The proposed model estimate labels from the output of the model by converting speaker labels into a single label. To consider speaker label permutations in the training, the proposed model is used a combination of Permutation Invariant Training (PIT) loss and cross-entropy loss. In addition, how to add the residual connection structures to model is studied for effective learning of speaker diarization models with deep structures. The experiment used the Librispech database to generate and use simulated noise data for two speakers. When compared with the proposed method and baseline model using the Diarization Error Rate (DER) performance the proposed method can be labeling without threshold, and it has improved performance by about 20.7 %.

Keywords

End-to-End speaker diarization

Single label

Residual connection

Loss combination

다수의 화자가 존재하는 음성에서 “누가 언제 발화했는가?”에 대해 레이블링하는 화자 분할은 발화 중첩 구간에 대한 레이블링과 화자 분할 모델의 최적화를 위해 심층 신경망 기반의 종단 간 방법에 대해 연구되었다. 대부분 심층 신경망 기반의 종단 간 화자 분할 시스템은 음성의 각 프레임에서 발화한 모든 화자의 레이블들을 추정하는 다중 레이블 분류 문제로 분할을 수행한다. 다중 레이블 기반의 화자 분할 시스템은 임계값을 어떤 값으로 설정하는지에 따라 모델의 성능이 많이 달라진다. 본 논문에서는 임계값 없이 화자 분할을 수행할 수 있도록 단일 레이블 분류를 이용한 화자 분할 시스템에 대해 연구하였다. 제안하는 화자 분할 시스템은 기존의 화자 레이블을 단일 레이블 형태로 변환하여 모델의 출력으로부터 레이블을 바로 추정한다. 훈련에서는 화자 레이블 순열을 고려하기 위해 Permutation Invariant Training(PIT) 손실함수와 교차 엔트로피 손실함수를 조합하여 사용하였다. 또한 심층 구조를 갖는 모델의 효과적인 학습을 위해 화자 분할 모델에 잔차 연결 구조를 추가하였다. 실험은 Librispeech 데이터베이스를 이용해 화자 2명에 대한 시뮬레이션 잡음 데이터를 생성하여 사용하였다. Diarization Error Rate(DER) 성능 평가 지수를 이용해 제안한 방법과 베이스라인 모델을 비교 평가했을 때, 제안한 방법이 임계값 없이 분할이 가능하며, 약 20.7 %만큼 향상된 성능을 보였다.

키워드

종단 간 화자 분할

단일 레이블

잔차 연결 구조

손실함수 조합

References

D. Garcia-Romero, D. Snyder, G. Sell, D. Povey, and A. McCree, "Speaker diarization using deep neural network embeddings," Proc. ICASSP, 4930-4934 (2017). 10.1109/ICASSP.2017.7953094

Q. Wang, C. Downey, L. Wan, P. A. Mansfield, and I. L. Moreno, "Speaker diarization with LSTM," Proc. ICASSP, 5239-5243 (2018). 10.1109/ICASSP.2018.8462628

M. Diez, L. Burget, S. Wang, J. Rohdin, and H. Cernocký, "Bayesian HMM based x-Vector clustering for speaker diarization," Proc. Interspeech, 346-350 (2019). 10.21437/Interspeech.2019-2813

I. Medennikov, M. Korenevsky, T. Prisyach, Y. Khokhlov, M. Korenevskaya, I. Sorokin, T. Timofeeva, A. Mitrofanov, A. Andrusenko, I. Podluzhny, A. Laptev, and A. Romanenko, "Target-speaker voice activity detection: a novel approach for multispeaker diarization in a dinner party scenario," Proc. Interspeech, 274-278 (2020). 10.21437/Interspeech.2020-1602

Y. C. Liu, E. Han, C. Lee, and A. Stolcke, "End-to-end neural diarization: From transformer to conformer," Proc. Interspeech, 3081-3085 (2021). 10.21437/Interspeech.2021-1909

Z. Du, S. Zhang, S. Zheng, and Z. Yan, "Speaker embedding-aware neural diarization: A novel framework for overlapping speech diarization in the meeting scenario," arXiv preprint arXiv:2203.09767 (2022).

Y. Fujita, N. Kanda, S. Horiguchi, K. Nagamatsu, and S. Watanabe, "End-to-end neural speaker diarization with permutation-free objectives," Proc. Interspeech, 4300-4304 (2019). 10.21437/Interspeech.2019-2899

Y. Fujita, N. Kanda, S. Horiguchi, Y. Xue, K. Nagamatsu, and S. Watanabe, "End-to-end neural speaker diarization with self-attention," Proc. ASRU, 296-303 (2019). 10.1109/ASRU46091.2019.9003959

Y. Yu, D. Park, and H. K. Kim, "Auxiliary loss of transformer with residual connection for end-to-end speaker diarization," Proc. ICASSP, 8377-8381 (2022). 10.1109/ICASSP43922.2022.9746602

V. Panayotov, G. Chen, D. Povey, and S. Khudanpur, "Librispeech: An asr corpus based on public domain audio books," Proc. ICASSP, 5206-5210 (2015). 10.1109/ICASSP.2015.7178964

D. Snyder, G. Chen, and D. Povey, "Musan: A music, speech, and noise corpus," arXiv preprint arXiv:1510. 08484 (2015).

T. Ko, V. Peddinti, D. Povey, M. L. Seltzer, and S. Khudanpur, "A study on data augmentation of reverberant speech for robust speech recognition," Proc. ICASSP, 5220-5224 (2017). 10.1109/ICASSP.2017.7953152

J. Carletta, "Unleashing the killer corpus: experiences in creating the multi-everything AMI Meeting Corpus," Lang. Resour. Eval. 41, 181-190 (2007). 10.1007/s10579-007-9040-x

A. Janin, D. Baron, J. Edwards, D. Ellis, D. Gelbart, N. Morgan, B. Peskin, T. Pfau, E. Shriberg, A. Stolcke, and C. Wooters, "The ICSI meeting corpus," Proc. ICASSP, 364-367 (2003).

A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. Kaiser, and I. Polosukhin, "Attention is all you need," Proc. NIPS, 5998-6008 (2017).

J. G. Fiscus, J. Ajot, and J. S. Garofolo, The Rich Transcription 2007 Meeting Recognition Evaluation (Springer, Maryland, 2007), pp. 373-389. 10.1007/978-3-540-68585-2_36

H. Bredin, R. Yin, J. M. Coria, G. Gelly, P. Korshunov, M. Lavechin, D. Fustes, H. Titeux, W, Bouaziz, and M. P. Gill, "Pyannote. audio: Neural building blocks for speaker diarization," Proc. ICASSP, 7124-7128 (2020). 10.1109/ICASSP40776.2020.9052974

H. Bredin and A. Laurent, "End-to-end speaker segmentation for overlap-aware resegmentation," Proc. Interspeech, 3111-3115 (2021). 10.21437/Interspeech.2021-560

Information

Publisher :The Acoustical Society of Korea
Publisher(Ko) :한국음향학회
Journal Title :The Journal of the Acoustical Society of Korea
Journal Title(Ko) :한국음향학회지
Volume : 42
No :6
Pages :536-543
Received Date : 2023-08-08
Accepted Date : 2023-09-06
DOI :https://doi.org/10.7776/ASK.2023.42.6.536

The Journal of the Acoustical Society of KoreaISSN:1225-4428(Print) 2287-3775(Online)한국음향학회

All Issue