Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Grünert, David | - |
dc.contributor.author | de Spindler, Alexandre | - |
dc.contributor.author | Dellwo, Volker | - |
dc.date.accessioned | 2023-10-21T13:01:43Z | - |
dc.date.available | 2023-10-21T13:01:43Z | - |
dc.date.issued | 2023 | - |
dc.identifier.uri | https://digitalcollection.zhaw.ch/handle/11475/28923 | - |
dc.description.abstract | Forensic audio does not seldom consist of long recordings of multiple speakers engaged in a dialogue. An important task for forensic phoneticians is then to say (a) how many speakers are present in the recording and (b) who speaks when. This can also lead to the identification of audio segments that are relevant for detailed inspection. Speaker diarization provides the fundamental ability to automatically split audio streams into segments assigned to speakers (Tranter, 2003). Since current diarization systems do not require audio profiles of speakers and do not assume any given number of speakers, they support a wide range of applications. In this paper, we present our approach to leverages speaker diarization to support forensic audio analysis. This also implies the identification of challenges for diarization systems in this context and the necessity of a novel evaluation metric. | de_CH |
dc.language.iso | en | de_CH |
dc.publisher | Centre for Forensic Phonetics and Acoustics (CFPA) | de_CH |
dc.rights | Licence according to publishing contract | de_CH |
dc.subject.ddc | 410.285: Computerlinguistik | de_CH |
dc.title | Speaker diarization systems in the context of forensic audio analysis | de_CH |
dc.type | Konferenz: Sonstiges | de_CH |
dcterms.type | Text | de_CH |
zhaw.departement | School of Management and Law | de_CH |
zhaw.organisationalunit | Institut für Wirtschaftsinformatik (IWI) | de_CH |
zhaw.publisher.place | Zurich | de_CH |
zhaw.conference.details | 31st Annual Conference of the International Association for Forensic Phonetics and Acoustics (IAFPA), Zurich, Switzerland, 9-12 July 2023 | de_CH |
zhaw.funding.eu | No | de_CH |
zhaw.originated.zhaw | Yes | de_CH |
zhaw.pages.end | 75 | de_CH |
zhaw.pages.start | 74 | de_CH |
zhaw.publication.status | publishedVersion | de_CH |
zhaw.publication.review | Peer review (Abstract) | de_CH |
zhaw.title.proceedings | 31st IAFPA Conference : Book of Abstracts | de_CH |
zhaw.author.additional | No | de_CH |
zhaw.display.portrait | Yes | de_CH |
Appears in collections: | Publikationen School of Management and Law |
Files in This Item:
There are no files associated with this item.
Show simple item record
Grünert, D., de Spindler, A., & Dellwo, V. (2023). Speaker diarization systems in the context of forensic audio analysis [Conference presentation]. 31st IAFPA Conference : Book of Abstracts, 74–75.
Grünert, D., de Spindler, A. and Dellwo, V. (2023) ‘Speaker diarization systems in the context of forensic audio analysis’, in 31st IAFPA Conference : Book of Abstracts. Zurich: Centre for Forensic Phonetics and Acoustics (CFPA), pp. 74–75.
D. Grünert, A. de Spindler, and V. Dellwo, “Speaker diarization systems in the context of forensic audio analysis,” in 31st IAFPA Conference : Book of Abstracts, 2023, pp. 74–75.
GRÜNERT, David, Alexandre DE SPINDLER und Volker DELLWO, 2023. Speaker diarization systems in the context of forensic audio analysis. In: 31st IAFPA Conference : Book of Abstracts. Conference presentation. Zurich: Centre for Forensic Phonetics and Acoustics (CFPA). 2023. S. 74–75
Grünert, David, Alexandre de Spindler, and Volker Dellwo. 2023. “Speaker Diarization Systems in the Context of Forensic Audio Analysis.” Conference presentation. In 31st IAFPA Conference : Book of Abstracts, 74–75. Zurich: Centre for Forensic Phonetics and Acoustics (CFPA).
Grünert, David, et al. “Speaker Diarization Systems in the Context of Forensic Audio Analysis.” 31st IAFPA Conference : Book of Abstracts, Centre for Forensic Phonetics and Acoustics (CFPA), 2023, pp. 74–75.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.