A Text-Independent Speaker Verification for SdSV Challenge 2020

Tran Dang Khoa, Tsung Han Tsai

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In this paper, we present an intelligent system for text-independent speaker verification. The system has been constructed based on the X-vector framework by expanding some TDNN layers to the baseline system. The work has been open evaluated by Short-Duration Speaker Verification (SdSV) Challenge 2020 for the task of speaker recognition. From the evaluaton, a fixed training condition of the dataset is provided to build the speaker verification (SV) system. In the end, the result can achieve the Equal Error Rate (EER) and minimum Detection Cost Function (minDCF) of 16.26% and 0.4794, respectively.

Original languageEnglish
Title of host publication2020 IEEE 5th International Conference on Computing Communication and Automation, ICCCA 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages459-461
Number of pages3
ISBN (Electronic)9781728163246
DOIs
StatePublished - 30 Oct 2020
Event5th IEEE International Conference on Computing Communication and Automation, ICCCA 2020 - Greater Noida, India
Duration: 30 Oct 202031 Oct 2020

Publication series

Name2020 IEEE 5th International Conference on Computing Communication and Automation, ICCCA 2020

Conference

Conference5th IEEE International Conference on Computing Communication and Automation, ICCCA 2020
Country/TerritoryIndia
CityGreater Noida
Period30/10/2031/10/20

Keywords

  • SdSV
  • speaker recognition
  • x-vector

Fingerprint

Dive into the research topics of 'A Text-Independent Speaker Verification for SdSV Challenge 2020'. Together they form a unique fingerprint.

Cite this