MULTI-CHANNEL SPEECH ENHANCEMENT BASED ON INDEPENDENT VECTOR EXTRACTION

Simple item page
dc.contributor.authorČmejla, Jaroslav
dc.contributor.authorKoldovský, Zbyněk
dc.date.accessioned2019-10-07T06:30:28Z
dc.date.available2019-10-07T06:30:28Z
dc.date.issued2018
dc.description.abstractIn this paper, a gradient-based algorithm for Independent Vector Extraction is introduced and is first used for blind audio source extraction. We also propose novel modifications of the gradient learning rule in the algorithm based on preconditioning of input data and by using the AdaGrad update. Experiments with six-channel CHiME-4 recordings are conducted where the modified algorithms show significant improvement in terms of convergence speed.cs
dc.format.extent6 strancs
dc.identifier.urihttps://dspace.tul.cz/handle/15240/153844
dc.identifier.urihttps://asap.ite.tul.cz/wp-content/uploads/sites/3/2018/09/iwaenc2018c.pdf
dc.language.isocscs
dc.publisherIEEE, 345 E 47TH ST, NEW YORK, NY 10017 USA
dc.relation.ispartof2018 16TH INTERNATIONAL WORKSHOP ON ACOUSTIC SIGNAL ENHANCEMENT (IWAENC)
dc.relation.ispartofseriesInternational Workshop on Acoustic Signal Enhancement
dc.subjectBlind Audio Source Separationcs
dc.subjectSpeech Enhancementcs
dc.subjectIndependent Component Analysiscs
dc.subjectIndependent Vector Analysiscs
dc.subjectIndependent Vector Extractioncs
dc.titleMULTI-CHANNEL SPEECH ENHANCEMENT BASED ON INDEPENDENT VECTOR EXTRACTIONcs
dc.typeProceedings Paper
local.citation.epage529
local.citation.spage525
local.event.edate2018-09-20
local.event.locationTokyo, JAPAN
local.event.sdate2018-09-17
local.event.title16th International Workshop on Acoustic Signal Enhancement (IWAENC)
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
MULTI-CHANNEL SPEECH ENHANCEMENT BASED ON INDEPENDENT VECTOR EXTRACTION.pdf
Size:
827.03 KB
Format:
Adobe Portable Document Format
Description:
článek
Download
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
1.71 KB
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections