Detail výsledku
Speaker recognition on mono-channel telephony recordings
Cohen Noa
Rohdin Johan Andréas, M.Sc., Ph.D., FIT (FIT), UPGM (FIT)
Madikeri Srikanth, FIT (FIT)
Černocký Jan, prof. Dr. Ing., UPGM (FIT)
Conversations stored as mono data is a common problem inmany real world speaker recognition applications. In this paper,we focus on investigative scenarios, where a number ofmono telephone conversations are available for a speaker of interest.For example, a human operator may have verified thatthe speaker is present in these conversations. We propose severalapproaches for automatically creating enrollment modelsfor the speaker of interest from such data. We then use the enrollmentmodels to search for appearances of the speaker of interestin other calls. We analyze the performance of the differentmethod on two dataset that matches our scenario, one is from asimulated case and one is from a real case.
speaker recognition, telephony recordings
@inproceedings{BUT179690,
author="Yosef {Solewicz} and Noa {Cohen} and Johan Andréas {Rohdin} and Srikanth {Madikeri} and Jan {Černocký}",
title="Speaker recognition on mono-channel telephony recordings",
booktitle="Proceedings of Odyssey 2022",
year="2022",
pages="193--199",
publisher="International Speech Communication Association",
address="Beijing",
doi="10.21437/Odyssey.2022-27",
url="https://www.isca-speech.org/archive/pdfs/odyssey_2022/solewicz22_odyssey.pdf"
}