---
_id: '57085'
abstract:
- lang: eng
text: We propose an approach for simultaneous diarization and separation of meeting
data. It consists of a complex Angular Central Gaussian Mixture Model (cACGMM)
for speech source separation, and a von-Mises-Fisher Mixture Model (VMFMM) for
diarization in a joint statistical framework. Through the integration, both spatial
and spectral information are exploited for diarization and separation. We also
develop a method for counting the number of active speakers in a segment of a
meeting to support block-wise processing. While the total number of speakers in
a meeting may be known, it is usually not known on a per-segment level. With the
proposed speaker counting, joint diarization and source separation can be done
segment-by-segment, and the permutation problem across segments is solved, thus
allowing for block-online processing in the future. Experimental results on the
LibriCSS meeting corpus show that the integrated approach outperforms a cascaded
approach of diarization and speech enhancement in terms of WER, both on a per-segment
and on a per-meeting level.
author:
- first_name: Tobias
full_name: Cord-Landwehr, Tobias
id: '44393'
last_name: Cord-Landwehr
- first_name: Christoph
full_name: Boeddeker, Christoph
id: '40767'
last_name: Boeddeker
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Cord-Landwehr T, Boeddeker C, Haeb-Umbach R. Simultaneous Diarization and
Separation of Meetings through the Integration of Statistical Mixture Models.
In: ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and
Signal Processing (ICASSP). ; 2024. doi:10.1109/ICASSP49660.2025.10888445'
apa: Cord-Landwehr, T., Boeddeker, C., & Haeb-Umbach, R. (2024). Simultaneous
Diarization and Separation of Meetings through the Integration of Statistical
Mixture Models. ICASSP 2025 - 2025 IEEE International Conference on Acoustics,
Speech and Signal Processing (ICASSP). 2025 IEEE International Conference
on Acoustics, Speech and Signal Processing (ICASSP), Hyderabad, India. https://doi.org/10.1109/ICASSP49660.2025.10888445
bibtex: '@inproceedings{Cord-Landwehr_Boeddeker_Haeb-Umbach_2024, title={Simultaneous
Diarization and Separation of Meetings through the Integration of Statistical
Mixture Models}, DOI={10.1109/ICASSP49660.2025.10888445},
booktitle={ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech
and Signal Processing (ICASSP)}, author={Cord-Landwehr, Tobias and Boeddeker,
Christoph and Haeb-Umbach, Reinhold}, year={2024} }'
chicago: Cord-Landwehr, Tobias, Christoph Boeddeker, and Reinhold Haeb-Umbach. “Simultaneous
Diarization and Separation of Meetings through the Integration of Statistical
Mixture Models.” In ICASSP 2025 - 2025 IEEE International Conference on Acoustics,
Speech and Signal Processing (ICASSP), 2024. https://doi.org/10.1109/ICASSP49660.2025.10888445.
ieee: 'T. Cord-Landwehr, C. Boeddeker, and R. Haeb-Umbach, “Simultaneous Diarization
and Separation of Meetings through the Integration of Statistical Mixture Models,”
presented at the 2025 IEEE International Conference on Acoustics, Speech and Signal
Processing (ICASSP), Hyderabad, India, 2024, doi: 10.1109/ICASSP49660.2025.10888445.'
mla: Cord-Landwehr, Tobias, et al. “Simultaneous Diarization and Separation of Meetings
through the Integration of Statistical Mixture Models.” ICASSP 2025 - 2025
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),
2024, doi:10.1109/ICASSP49660.2025.10888445.
short: 'T. Cord-Landwehr, C. Boeddeker, R. Haeb-Umbach, in: ICASSP 2025 - 2025 IEEE
International Conference on Acoustics, Speech and Signal Processing (ICASSP),
2024.'
conference:
location: Hyderabad, India
name: 2025 IEEE International Conference on Acoustics, Speech and Signal Processing
(ICASSP)
date_created: 2024-11-14T09:32:38Z
date_updated: 2025-08-14T08:12:22Z
ddc:
- '000'
department:
- _id: '54'
doi: 10.1109/ICASSP49660.2025.10888445
file:
- access_level: closed
content_type: application/pdf
creator: cord
date_created: 2025-08-14T08:11:57Z
date_updated: 2025-08-14T08:11:57Z
file_id: '60930'
file_name: main.pdf
file_size: 259907
relation: main_file
success: 1
file_date_updated: 2025-08-14T08:11:57Z
has_accepted_license: '1'
keyword:
- diarization
- source separation
- mixture model
- meeting
language:
- iso: eng
main_file_link:
- open_access: '1'
url: https://arxiv.org/pdf/2410.21455
oa: '1'
project:
- _id: '52'
name: 'PC2: Computing Resources Provided by the Paderborn Center for Parallel Computing'
- _id: '508'
name: Automatische Transkription von Gesprächssituationen
publication: ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech
and Signal Processing (ICASSP)
status: public
title: Simultaneous Diarization and Separation of Meetings through the Integration
of Statistical Mixture Models
type: conference
user_id: '44393'
year: '2024'
...
---
_id: '49109'
abstract:
- lang: eng
text: "We propose a diarization system, that estimates “who spoke when” based on
spatial information, to be used as a front-end of a meeting transcription system
running on the signals gathered from an acoustic sensor network (ASN). Although
the\r\nspatial distribution of the microphones is advantageous, exploiting the
spatial diversity for diarization and signal enhancement is challenging, because
the microphones’ positions are typically unknown, and the recorded signals are
initially unsynchronized in general. Here, we approach these issues by first blindly
synchronizing the signals and then estimating time differences of arrival (TDOAs).
The TDOA information is exploited to estimate the speakers’ activity, even in
the presence of multiple speakers being simultaneously active. This speaker activity
information serves as a guide for a spatial mixture model, on which basis the
individual speaker’s signals are extracted via beamforming. Finally, the extracted
signals are forwarded to a speech recognizer. Additionally, a novel initialization
scheme for spatial mixture models based on the TDOA estimates is proposed. Experiments
conducted on real recordings from the LibriWASN data set have shown that our proposed
system is advantageous compared to a system using a spatial mixture model, which
does not make use\r\nof external diarization information."
author:
- first_name: Tobias
full_name: Gburrek, Tobias
id: '44006'
last_name: Gburrek
- first_name: Joerg
full_name: Schmalenstroeer, Joerg
id: '460'
last_name: Schmalenstroeer
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Gburrek T, Schmalenstroeer J, Haeb-Umbach R. Spatial Diarization for Meeting
Transcription with Ad-Hoc Acoustic Sensor Networks. In: Proc. Asilomar Conference
on Signals, Systems, and Computers. ; 2023.'
apa: Gburrek, T., Schmalenstroeer, J., & Haeb-Umbach, R. (2023). Spatial Diarization
for Meeting Transcription with Ad-Hoc Acoustic Sensor Networks. Proc. Asilomar
Conference on Signals, Systems, and Computers. 57th Asilomar Conference on
Signals, Systems, and Computers.
bibtex: '@inproceedings{Gburrek_Schmalenstroeer_Haeb-Umbach_2023, title={Spatial
Diarization for Meeting Transcription with Ad-Hoc Acoustic Sensor Networks}, booktitle={Proc.
Asilomar Conference on Signals, Systems, and Computers}, author={Gburrek, Tobias
and Schmalenstroeer, Joerg and Haeb-Umbach, Reinhold}, year={2023} }'
chicago: Gburrek, Tobias, Joerg Schmalenstroeer, and Reinhold Haeb-Umbach. “Spatial
Diarization for Meeting Transcription with Ad-Hoc Acoustic Sensor Networks.” In
Proc. Asilomar Conference on Signals, Systems, and Computers, 2023.
ieee: T. Gburrek, J. Schmalenstroeer, and R. Haeb-Umbach, “Spatial Diarization for
Meeting Transcription with Ad-Hoc Acoustic Sensor Networks,” presented at the
57th Asilomar Conference on Signals, Systems, and Computers, 2023.
mla: Gburrek, Tobias, et al. “Spatial Diarization for Meeting Transcription with
Ad-Hoc Acoustic Sensor Networks.” Proc. Asilomar Conference on Signals, Systems,
and Computers, 2023.
short: 'T. Gburrek, J. Schmalenstroeer, R. Haeb-Umbach, in: Proc. Asilomar Conference
on Signals, Systems, and Computers, 2023.'
conference:
end_date: 2023-11-01
name: 57th Asilomar Conference on Signals, Systems, and Computers
start_date: 2023-10-31
date_created: 2023-11-22T07:52:29Z
date_updated: 2023-11-22T07:58:49Z
ddc:
- '004'
department:
- _id: '54'
file:
- access_level: open_access
content_type: application/pdf
creator: schmalen
date_created: 2023-11-22T07:51:18Z
date_updated: 2023-11-22T07:58:49Z
file_id: '49110'
file_name: asilomar.pdf
file_size: 212317
relation: main_file
file_date_updated: 2023-11-22T07:58:49Z
has_accepted_license: '1'
keyword:
- Diarization
- time difference of arrival
- ad-hoc acoustic sensor network
- meeting transcription
language:
- iso: eng
oa: '1'
publication: Proc. Asilomar Conference on Signals, Systems, and Computers
quality_controlled: '1'
status: public
title: Spatial Diarization for Meeting Transcription with Ad-Hoc Acoustic Sensor Networks
type: conference
user_id: '460'
year: '2023'
...
---
_id: '48281'
abstract:
- lang: eng
text: "\tWe propose a general framework to compute the word error rate (WER) of
ASR systems that process recordings containing multiple speakers at their input
and that produce multiple output word sequences (MIMO).\r\n\tSuch ASR systems
are typically required, e.g., for meeting transcription.\r\n\tWe provide an efficient
implementation based on a dynamic programming search in a multi-dimensional Levenshtein
distance tensor under the constraint that a reference utterance must be matched
consistently with one hypothesis output. \r\n\tThis also results in an efficient
implementation of the ORC WER which previously suffered from exponential complexity.\r\n\tWe
give an overview of commonly used WER definitions for multi-speaker scenarios
and show that they are specializations of the above MIMO WER tuned to particular
application scenarios. \r\n\tWe conclude with a discussion of the pros and cons
of the various WER definitions and a recommendation when to use which."
author:
- first_name: Thilo
full_name: von Neumann, Thilo
id: '49870'
last_name: von Neumann
orcid: https://orcid.org/0000-0002-7717-8670
- first_name: Christoph
full_name: Boeddeker, Christoph
id: '40767'
last_name: Boeddeker
- first_name: Keisuke
full_name: Kinoshita, Keisuke
last_name: Kinoshita
- first_name: Marc
full_name: Delcroix, Marc
last_name: Delcroix
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'von Neumann T, Boeddeker C, Kinoshita K, Delcroix M, Haeb-Umbach R. On Word
Error Rate Definitions and Their Efficient Computation for Multi-Speaker Speech
Recognition Systems. In: ICASSP 2023 - 2023 IEEE International Conference on
Acoustics, Speech and Signal Processing (ICASSP). IEEE; 2023. doi:10.1109/icassp49357.2023.10094784'
apa: von Neumann, T., Boeddeker, C., Kinoshita, K., Delcroix, M., & Haeb-Umbach,
R. (2023). On Word Error Rate Definitions and Their Efficient Computation for
Multi-Speaker Speech Recognition Systems. ICASSP 2023 - 2023 IEEE International
Conference on Acoustics, Speech and Signal Processing (ICASSP). https://doi.org/10.1109/icassp49357.2023.10094784
bibtex: '@inproceedings{von Neumann_Boeddeker_Kinoshita_Delcroix_Haeb-Umbach_2023,
title={On Word Error Rate Definitions and Their Efficient Computation for Multi-Speaker
Speech Recognition Systems}, DOI={10.1109/icassp49357.2023.10094784},
booktitle={ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech
and Signal Processing (ICASSP)}, publisher={IEEE}, author={von Neumann, Thilo
and Boeddeker, Christoph and Kinoshita, Keisuke and Delcroix, Marc and Haeb-Umbach,
Reinhold}, year={2023} }'
chicago: Neumann, Thilo von, Christoph Boeddeker, Keisuke Kinoshita, Marc Delcroix,
and Reinhold Haeb-Umbach. “On Word Error Rate Definitions and Their Efficient
Computation for Multi-Speaker Speech Recognition Systems.” In ICASSP 2023 -
2023 IEEE International Conference on Acoustics, Speech and Signal Processing
(ICASSP). IEEE, 2023. https://doi.org/10.1109/icassp49357.2023.10094784.
ieee: 'T. von Neumann, C. Boeddeker, K. Kinoshita, M. Delcroix, and R. Haeb-Umbach,
“On Word Error Rate Definitions and Their Efficient Computation for Multi-Speaker
Speech Recognition Systems,” 2023, doi: 10.1109/icassp49357.2023.10094784.'
mla: von Neumann, Thilo, et al. “On Word Error Rate Definitions and Their Efficient
Computation for Multi-Speaker Speech Recognition Systems.” ICASSP 2023 - 2023
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),
IEEE, 2023, doi:10.1109/icassp49357.2023.10094784.
short: 'T. von Neumann, C. Boeddeker, K. Kinoshita, M. Delcroix, R. Haeb-Umbach,
in: ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and
Signal Processing (ICASSP), IEEE, 2023.'
date_created: 2023-10-19T07:38:31Z
date_updated: 2025-02-12T09:16:34Z
ddc:
- '000'
department:
- _id: '54'
doi: 10.1109/icassp49357.2023.10094784
file:
- access_level: open_access
content_type: application/pdf
creator: tvn
date_created: 2023-10-19T07:39:57Z
date_updated: 2023-10-19T07:41:56Z
file_id: '48282'
file_name: ICASSP_2023_Meeting_Evaluation.pdf
file_size: 204994
relation: main_file
file_date_updated: 2023-10-19T07:41:56Z
has_accepted_license: '1'
keyword:
- Word Error Rate
- Meeting Recognition
- Levenshtein Distance
language:
- iso: eng
main_file_link:
- url: https://ieeexplore.ieee.org/document/10094784
oa: '1'
project:
- _id: '52'
name: 'PC2: Computing Resources Provided by the Paderborn Center for Parallel Computing'
- _id: '508'
grant_number: '448568305'
name: Automatische Transkription von Gesprächssituationen
publication: ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech
and Signal Processing (ICASSP)
publication_status: published
publisher: IEEE
quality_controlled: '1'
related_material:
link:
- relation: software
url: https://github.com/fgnt/meeteval
status: public
title: On Word Error Rate Definitions and Their Efficient Computation for Multi-Speaker
Speech Recognition Systems
type: conference
user_id: '40767'
year: '2023'
...
---
_id: '48275'
abstract:
- lang: eng
text: "MeetEval is an open-source toolkit to evaluate all kinds of meeting transcription
systems.\r\nIt provides a unified interface for the computation of commonly used
Word Error Rates (WERs), specifically cpWER, ORC WER and MIMO WER along other
WER definitions.\r\nWe extend the cpWER computation by a temporal constraint to
ensure that only words are identified as correct when the temporal alignment is
plausible.\r\nThis leads to a better quality of the matching of the hypothesis
string to the reference string that more closely resembles the actual transcription
quality, and a system is penalized if it provides poor time annotations.\r\nSince
word-level timing information is often not available, we present a way to approximate
exact word-level timings from segment-level timings (e.g., a sentence) and show
that the approximation leads to a similar WER as a matching with exact word-level
annotations.\r\nAt the same time, the time constraint leads to a speedup of the
matching algorithm, which outweighs the additional overhead caused by processing
the time stamps."
author:
- first_name: Thilo
full_name: von Neumann, Thilo
id: '49870'
last_name: von Neumann
orcid: https://orcid.org/0000-0002-7717-8670
- first_name: Christoph
full_name: Boeddeker, Christoph
id: '40767'
last_name: Boeddeker
- first_name: Marc
full_name: Delcroix, Marc
last_name: Delcroix
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'von Neumann T, Boeddeker C, Delcroix M, Haeb-Umbach R. MeetEval: A Toolkit
for Computation of Word Error Rates for Meeting Transcription Systems. In: Proc.
CHiME 2023 Workshop on Speech Processing in Everyday Environments. ; 2023.'
apa: 'von Neumann, T., Boeddeker, C., Delcroix, M., & Haeb-Umbach, R. (2023).
MeetEval: A Toolkit for Computation of Word Error Rates for Meeting Transcription
Systems. Proc. CHiME 2023 Workshop on Speech Processing in Everyday Environments.
CHiME 2023 Workshop on Speech Processing in Everyday Environments, Dublin.'
bibtex: '@inproceedings{von Neumann_Boeddeker_Delcroix_Haeb-Umbach_2023, title={MeetEval:
A Toolkit for Computation of Word Error Rates for Meeting Transcription Systems},
booktitle={Proc. CHiME 2023 Workshop on Speech Processing in Everyday Environments},
author={von Neumann, Thilo and Boeddeker, Christoph and Delcroix, Marc and Haeb-Umbach,
Reinhold}, year={2023} }'
chicago: 'Neumann, Thilo von, Christoph Boeddeker, Marc Delcroix, and Reinhold Haeb-Umbach.
“MeetEval: A Toolkit for Computation of Word Error Rates for Meeting Transcription
Systems.” In Proc. CHiME 2023 Workshop on Speech Processing in Everyday Environments,
2023.'
ieee: 'T. von Neumann, C. Boeddeker, M. Delcroix, and R. Haeb-Umbach, “MeetEval:
A Toolkit for Computation of Word Error Rates for Meeting Transcription Systems,”
presented at the CHiME 2023 Workshop on Speech Processing in Everyday Environments,
Dublin, 2023.'
mla: 'von Neumann, Thilo, et al. “MeetEval: A Toolkit for Computation of Word Error
Rates for Meeting Transcription Systems.” Proc. CHiME 2023 Workshop on Speech
Processing in Everyday Environments, 2023.'
short: 'T. von Neumann, C. Boeddeker, M. Delcroix, R. Haeb-Umbach, in: Proc. CHiME
2023 Workshop on Speech Processing in Everyday Environments, 2023.'
conference:
location: Dublin
name: CHiME 2023 Workshop on Speech Processing in Everyday Environments
date_created: 2023-10-19T07:24:51Z
date_updated: 2025-02-12T09:12:05Z
ddc:
- '000'
department:
- _id: '54'
file:
- access_level: open_access
content_type: application/pdf
creator: tvn
date_created: 2023-10-19T07:19:59Z
date_updated: 2023-10-19T07:19:59Z
file_id: '48276'
file_name: Chime_7__MeetEval.pdf
file_size: 263744
relation: main_file
file_date_updated: 2023-10-19T07:19:59Z
has_accepted_license: '1'
keyword:
- Speech Recognition
- Word Error Rate
- Meeting Transcription
language:
- iso: eng
main_file_link:
- open_access: '1'
url: https://arxiv.org/abs/2307.11394
oa: '1'
project:
- _id: '52'
name: 'PC2: Computing Resources Provided by the Paderborn Center for Parallel Computing'
- _id: '508'
grant_number: '448568305'
name: Automatische Transkription von Gesprächssituationen
publication: Proc. CHiME 2023 Workshop on Speech Processing in Everyday Environments
quality_controlled: '1'
related_material:
link:
- relation: software
url: https://github.com/fgnt/meeteval
status: public
title: 'MeetEval: A Toolkit for Computation of Word Error Rates for Meeting Transcription
Systems'
type: conference
user_id: '40767'
year: '2023'
...