Full metadata record
DC FieldValueLanguage
dc.contributor.authorChien, Jen-Tzungen_US
dc.contributor.authorChang, You-Chengen_US
dc.date.accessioned2017-04-21T06:49:22Z-
dc.date.available2017-04-21T06:49:22Z-
dc.date.issued2016en_US
dc.identifier.isbn978-1-5090-0746-2en_US
dc.identifier.issn2161-0363en_US
dc.identifier.urihttp://hdl.handle.net/11536/134553-
dc.description.abstractThis study presents a Bayesian approach to enhance the magnitude spectra of single-channel reverberant speech signals. Speech dereverberation model is constructed by using a non-negative convolutive transfer function (NCTF) and a nonnegative matrix factorization (NMF). NCTF is used to characterize the magnitude spectra of speech signal and room impulse response while NMF is applied to represent the fine structure of speech spectra. Importantly, we deal with the variations of dereverberation model by introducing the exponential priors for reverberation kernel and noise signal. A full Bayesian solution to speech dereverberation is obtained according to the variational Bayesian inference algorithm. Using this algorithm, the room configuration and the speaker characteristics are automatically learned from data. Such a general model can be reduced to the previous methods. Experimental results on both simulated data and real recordings from 2014 RE-VERB Challenge show the merit of the proposed method for single-channel speech dereverberation.en_US
dc.language.isoen_USen_US
dc.subjectBayesian learningen_US
dc.subjectnonnegative matrix factorizationen_US
dc.subjectspeech dereverberationen_US
dc.titleBAYESIAN LEARNING FOR SPEECH DEREVERBERATIONen_US
dc.typeProceedings Paperen_US
dc.identifier.journal2016 IEEE 26TH INTERNATIONAL WORKSHOP ON MACHINE LEARNING FOR SIGNAL PROCESSING (MLSP)en_US
dc.contributor.department電機工程學系zh_TW
dc.contributor.departmentDepartment of Electrical and Computer Engineeringen_US
dc.identifier.wosnumberWOS:000392177200057en_US
dc.citation.woscount0en_US
Appears in Collections:Conferences Paper