Dimension reduction of the modulation spectrogram for speaker verification

Tomi Kinnunen, Kong Aik Lee, Haizhou Li

Research output: Unpublished conference presentation (presented paper, abstract, poster)Conference presentation (not published in journal/proceeding/book)Academic researchpeer-review

38 Citations (Scopus)

Abstract

A so-called modulation spectrogram is obtained from the conventional speech spectrogram by short-term spectral analysis along the temporal trajectories of the frequency bins. In its original definition, the modulation spectrogram is a high-dimensional representation and it is not clear how to extract features from it. In this paper, we define a low-dimensional feature which captures the shape of the modulation spectra. The recognition accuracy of the modulation spectrogram based classifier is improved from our previous result of EER=25.1% to EER=17.4% on the NIST 2001 speaker recognition task.

Original languageEnglish
Publication statusPublished - Jan 2008
Externally publishedYes
EventSpeaker and Language Recognition Workshop, Odyssey 2008 - Stellenbosch, South Africa
Duration: 21 Jan 200824 Jan 2008

Conference

ConferenceSpeaker and Language Recognition Workshop, Odyssey 2008
Country/TerritorySouth Africa
CityStellenbosch
Period21/01/0824/01/08

Keywords

  • Modulation spectrum
  • Speaker recognition
  • Spectro-temporal features

ASJC Scopus subject areas

  • Human-Computer Interaction
  • Software
  • Signal Processing

Fingerprint

Dive into the research topics of 'Dimension reduction of the modulation spectrogram for speaker verification'. Together they form a unique fingerprint.

Cite this