Extraction of Speech from Mixture Signals

Research output: Chapter in Book/Report/Conference proceedingChapter

Original languageEnglish (US)
Title of host publicationTechniques for Noise Robustness in Automatic Speech Recognition
PublisherJohn Wiley and Sons
Pages87-108
Number of pages22
ISBN (Print)9781119970880
DOIs
StatePublished - Oct 5 2012

Keywords

  • ICA and multichannel source separation, spectral factorization
  • ICA-based multichannel and NMF-based single-channel approaches
  • Interference removal, in clean speech input operation
  • Mixed signal separation and clean version of speech preprocessing
  • Multichannel sound mixtures, recordings using microphone arrays
  • NMF-based source separation performance on mixtures
  • Nonnegative spectral sound factorization, using nonnegative models
  • Single-channel mixtures, challenging problems of practical deployment
  • Speaker's spectral character, to mixture decompositions in part
  • Speech extraction from mixture signals

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Smaragdis, P. (2012). Extraction of Speech from Mixture Signals. In Techniques for Noise Robustness in Automatic Speech Recognition (pp. 87-108). John Wiley and Sons. https://doi.org/10.1002/9781118392683.ch5