Skip to content

Publications

Detail of publication

Citation

Zbynek Zajic and Jan Zelinka and Ludek Muller : Neural Network Speaker Descriptor in Speaker Diarization of Telephone Speech . Speech and Computer 19th International Conference (SPECOM 2017), p. 555-563, Springer, 2017.

Download PDF

PDF 2

Abstract

In this paper, we have been investigating an approach to a speaker representation for a diarization system that clusters short telephone conversation segments (produced by the same speaker). The proposed approach applies a neural-network-based descriptor that replaces a usual i-vector descriptor in the state-of-the-art diarization systems. The comparison of these two techniques was done on the English part of the CallHome corpus. The final results indicate the superiority of the i-vector’s approach although our proposed descriptor brings an additive information. Thus, the combined descriptor represents a speaker in a segment for diarization purpose with lower diarization error (almost 20% relative improvement compared with only i-vector application).

Detail of publication

Title: Neural Network Speaker Descriptor in Speaker Diarization of Telephone Speech
Author: Zbynek Zajic ; Jan Zelinka ; Ludek Muller
Language: English
Year: 2017
Type of publication: Conferences presentations outside the Czech Republic
Title of journal or book: Speech and Computer 19th International Conference (SPECOM 2017)
Page: 555 - 563
DOI: 10.1007/978-3-319-66429-3_55
Publisher: Springer
/ 2019-11-15 14:44:26 /

Keywords

Neural network, Speaker diarization, i-Vector

BibTeX

@INPROCEEDINGS{ZbynekZajic_2017_NeuralNetwork,
 author = {Zbynek Zajic and Jan Zelinka and Ludek Muller},
 title = {Neural Network Speaker Descriptor in Speaker Diarization of Telephone Speech},
 year = {2017},
 publisher = {Springer},
 journal = {Speech and Computer 19th International Conference (SPECOM 2017)},
 pages = {555-563},
 doi = {10.1007/978-3-319-66429-3_55},
 url = {http://www.kky.zcu.cz/en/publications/ZbynekZajic_2017_NeuralNetwork},
}