Skip to content

Underdetermined blind separation using learned basis function sets

Research output: Contribution to journalArticle

Original languageEnglish
Pages158 - 160
JournalElectronics Letters
Journal publication dateJan 2003
Volume39
Journal issue1
DOIs
StatePublished

Abstract

Blind separation from an underdetermined system is approached using a learned basis function dictionary. An overcomplete basis function set is learned from individual speech samples and the combined set reduced in size by eliminating similar functions. This library is then used as an effective overcomplete basis set for separation

Additional information

Publisher: Institution of Electrical Engineers (IEE) Rose publication type: Journal article Sponsorship: The authors wish to thank the UK MOD TGOl CRP Fellowship Project for support

View research connections

Related faculties, schools or groups