Robust Speech Enhancement Based on NPHMM Under Unknown Noise

In this paper, a new speech enhancement based on the nonlinear H ∞ filtering and neural predictive HMM (NPHMM) is presented. In H ∞ filtering, no a priorknowledge of the noise source statistics is required. Speech is modeled as the output of a neural predictive HMM combining MLP neural network and...

Ausführliche Beschreibung

Gespeichert in:

Bibliographische Detailangaben
Hauptverfasser:	Lee, Ki Yong, Rheem, Jae Yeol
Format:	Tagungsbericht
Sprache:	eng
Schlagworte:	Applied sciences Artificial intelligence Computer science control theory systems Exact sciences and technology Noisy Speech Noisy Speech Signal Speech and sound recognition and synthesis. Linguistics Speech Enhancement Unknown Noise Vocal Tract
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

container_end_page	431
container_issue
container_start_page	427
container_title
container_volume
creator	Lee, Ki Yong Rheem, Jae Yeol
description	In this paper, a new speech enhancement based on the nonlinear H ∞ filtering and neural predictive HMM (NPHMM) is presented. In H ∞ filtering, no a priorknowledge of the noise source statistics is required. Speech is modeled as the output of a neural predictive HMM combining MLP neural network and HMM. The proposed enhancement method consists of multiple nonlinear H ∞ filters with parameter of the NPHMM. The switching between the nonlinear H ∞ filters is governed by a finite state Markov chain according to the transition probabilities. An approximate improvement of 0.4-1.8dB in output SNR is achieved at various input SNRs compared with conventional Kalman method.
doi_str_mv	10.1007/11520153_29
format	Conference Proceeding
fullrecord	<record><control><sourceid>pascalfrancis_sprin</sourceid><recordid>TN_cdi_pascalfrancis_primary_17027335</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>17027335</sourcerecordid><originalsourceid>FETCH-LOGICAL-p219t-58e94756a17533d25012b3a2232617c675cbd197217bd4c09b8d327d84479ca13</originalsourceid><addsrcrecordid>eNpNUE1LAzEUjF9gqT35B_biwcNqXl6y2YAXLdUKrYrac8gmqV3bZpfNivjvjVTEd3gP3gzDzBByCvQCKJWXAIJREKiZ2iMjJUsUnCKUZVHskwEUADkiVwd_GJOcAz8kA4qU5UpyPCajGN9pGgTFAAfk6rmpPmKfvbTe21U2CSsTrN_60Gc3JnqXNSF7eJrO59kiON-lvQ7NZ_o1dfQn5GhpNtGPfu-QLG4nr-NpPnu8ux9fz_KWgepzUXrFpSgMSIHomKDAKjSMIStA2kIKWzlQkoGsHLdUVaVDJl3JuVTWAA7J2U63NdGazbJLHuuo267emu5Lg0xREUXine94MUHhzXe6app11ED1T4P6X4P4DfimWcE</addsrcrecordid><sourcetype>Index Database</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype></control><display><type>conference_proceeding</type><title>Robust Speech Enhancement Based on NPHMM Under Unknown Noise</title><source>Springer Books</source><creator>Lee, Ki Yong ; Rheem, Jae Yeol</creator><contributor>Chollet, Gérard ; Esposito, Anna ; Marinaro, Maria ; Faundez-Zanuy, Marcos</contributor><creatorcontrib>Lee, Ki Yong ; Rheem, Jae Yeol ; Chollet, Gérard ; Esposito, Anna ; Marinaro, Maria ; Faundez-Zanuy, Marcos</creatorcontrib><description>In this paper, a new speech enhancement based on the nonlinear H ∞ filtering and neural predictive HMM (NPHMM) is presented. In H ∞ filtering, no a priorknowledge of the noise source statistics is required. Speech is modeled as the output of a neural predictive HMM combining MLP neural network and HMM. The proposed enhancement method consists of multiple nonlinear H ∞ filters with parameter of the NPHMM. The switching between the nonlinear H ∞ filters is governed by a finite state Markov chain according to the transition probabilities. An approximate improvement of 0.4-1.8dB in output SNR is achieved at various input SNRs compared with conventional Kalman method.</description><identifier>ISSN: 0302-9743</identifier><identifier>ISBN: 9783540274414</identifier><identifier>ISBN: 3540274413</identifier><identifier>EISSN: 1611-3349</identifier><identifier>EISBN: 9783540318866</identifier><identifier>EISBN: 3540318860</identifier><identifier>DOI: 10.1007/11520153_29</identifier><language>eng</language><publisher>Berlin, Heidelberg: Springer Berlin Heidelberg</publisher><subject>Applied sciences ; Artificial intelligence ; Computer science; control theory; systems ; Exact sciences and technology ; Noisy Speech ; Noisy Speech Signal ; Speech and sound recognition and synthesis. Linguistics ; Speech Enhancement ; Unknown Noise ; Vocal Tract</subject><ispartof>Lecture notes in computer science, 2005, p.427-431</ispartof><rights>Springer-Verlag Berlin Heidelberg 2005</rights><rights>2005 INIST-CNRS</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://link.springer.com/content/pdf/10.1007/11520153_29$$EPDF$$P50$$Gspringer$$H</linktopdf><linktohtml>$$Uhttps://link.springer.com/10.1007/11520153_29$$EHTML$$P50$$Gspringer$$H</linktohtml><link.rule.ids>309,310,779,780,784,789,790,793,4050,4051,27925,38255,41442,42511</link.rule.ids><backlink>$$Uhttp://pascal-francis.inist.fr/vibad/index.php?action=getRecordDetail&idt=17027335$$DView record in Pascal Francis$$Hfree_for_read</backlink></links><search><contributor>Chollet, Gérard</contributor><contributor>Esposito, Anna</contributor><contributor>Marinaro, Maria</contributor><contributor>Faundez-Zanuy, Marcos</contributor><creatorcontrib>Lee, Ki Yong</creatorcontrib><creatorcontrib>Rheem, Jae Yeol</creatorcontrib><title>Robust Speech Enhancement Based on NPHMM Under Unknown Noise</title><title>Lecture notes in computer science</title><description>In this paper, a new speech enhancement based on the nonlinear H ∞ filtering and neural predictive HMM (NPHMM) is presented. In H ∞ filtering, no a priorknowledge of the noise source statistics is required. Speech is modeled as the output of a neural predictive HMM combining MLP neural network and HMM. The proposed enhancement method consists of multiple nonlinear H ∞ filters with parameter of the NPHMM. The switching between the nonlinear H ∞ filters is governed by a finite state Markov chain according to the transition probabilities. An approximate improvement of 0.4-1.8dB in output SNR is achieved at various input SNRs compared with conventional Kalman method.</description><subject>Applied sciences</subject><subject>Artificial intelligence</subject><subject>Computer science; control theory; systems</subject><subject>Exact sciences and technology</subject><subject>Noisy Speech</subject><subject>Noisy Speech Signal</subject><subject>Speech and sound recognition and synthesis. Linguistics</subject><subject>Speech Enhancement</subject><subject>Unknown Noise</subject><subject>Vocal Tract</subject><issn>0302-9743</issn><issn>1611-3349</issn><isbn>9783540274414</isbn><isbn>3540274413</isbn><isbn>9783540318866</isbn><isbn>3540318860</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>2005</creationdate><recordtype>conference_proceeding</recordtype><recordid>eNpNUE1LAzEUjF9gqT35B_biwcNqXl6y2YAXLdUKrYrac8gmqV3bZpfNivjvjVTEd3gP3gzDzBByCvQCKJWXAIJREKiZ2iMjJUsUnCKUZVHskwEUADkiVwd_GJOcAz8kA4qU5UpyPCajGN9pGgTFAAfk6rmpPmKfvbTe21U2CSsTrN_60Gc3JnqXNSF7eJrO59kiON-lvQ7NZ_o1dfQn5GhpNtGPfu-QLG4nr-NpPnu8ux9fz_KWgepzUXrFpSgMSIHomKDAKjSMIStA2kIKWzlQkoGsHLdUVaVDJl3JuVTWAA7J2U63NdGazbJLHuuo267emu5Lg0xREUXine94MUHhzXe6app11ED1T4P6X4P4DfimWcE</recordid><startdate>2005</startdate><enddate>2005</enddate><creator>Lee, Ki Yong</creator><creator>Rheem, Jae Yeol</creator><general>Springer Berlin Heidelberg</general><general>Springer</general><scope>IQODW</scope></search><sort><creationdate>2005</creationdate><title>Robust Speech Enhancement Based on NPHMM Under Unknown Noise</title><author>Lee, Ki Yong ; Rheem, Jae Yeol</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-p219t-58e94756a17533d25012b3a2232617c675cbd197217bd4c09b8d327d84479ca13</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>2005</creationdate><topic>Applied sciences</topic><topic>Artificial intelligence</topic><topic>Computer science; control theory; systems</topic><topic>Exact sciences and technology</topic><topic>Noisy Speech</topic><topic>Noisy Speech Signal</topic><topic>Speech and sound recognition and synthesis. Linguistics</topic><topic>Speech Enhancement</topic><topic>Unknown Noise</topic><topic>Vocal Tract</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Lee, Ki Yong</creatorcontrib><creatorcontrib>Rheem, Jae Yeol</creatorcontrib><collection>Pascal-Francis</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Lee, Ki Yong</au><au>Rheem, Jae Yeol</au><au>Chollet, Gérard</au><au>Esposito, Anna</au><au>Marinaro, Maria</au><au>Faundez-Zanuy, Marcos</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>Robust Speech Enhancement Based on NPHMM Under Unknown Noise</atitle><btitle>Lecture notes in computer science</btitle><date>2005</date><risdate>2005</risdate><spage>427</spage><epage>431</epage><pages>427-431</pages><issn>0302-9743</issn><eissn>1611-3349</eissn><isbn>9783540274414</isbn><isbn>3540274413</isbn><eisbn>9783540318866</eisbn><eisbn>3540318860</eisbn><abstract>In this paper, a new speech enhancement based on the nonlinear H ∞ filtering and neural predictive HMM (NPHMM) is presented. In H ∞ filtering, no a priorknowledge of the noise source statistics is required. Speech is modeled as the output of a neural predictive HMM combining MLP neural network and HMM. The proposed enhancement method consists of multiple nonlinear H ∞ filters with parameter of the NPHMM. The switching between the nonlinear H ∞ filters is governed by a finite state Markov chain according to the transition probabilities. An approximate improvement of 0.4-1.8dB in output SNR is achieved at various input SNRs compared with conventional Kalman method.</abstract><cop>Berlin, Heidelberg</cop><pub>Springer Berlin Heidelberg</pub><doi>10.1007/11520153_29</doi><tpages>5</tpages></addata></record>
fulltext	fulltext
identifier	ISSN: 0302-9743
ispartof	Lecture notes in computer science, 2005, p.427-431
issn	0302-9743 1611-3349
language	eng
recordid	cdi_pascalfrancis_primary_17027335
source	Springer Books
subjects	Applied sciences Artificial intelligence Computer science control theory systems Exact sciences and technology Noisy Speech Noisy Speech Signal Speech and sound recognition and synthesis. Linguistics Speech Enhancement Unknown Noise Vocal Tract
title	Robust Speech Enhancement Based on NPHMM Under Unknown Noise
url	https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-03T23%3A10%3A08IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-pascalfrancis_sprin&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=Robust%20Speech%20Enhancement%20Based%20on%20NPHMM%20Under%20Unknown%20Noise&rft.btitle=Lecture%20notes%20in%20computer%20science&rft.au=Lee,%20Ki%20Yong&rft.date=2005&rft.spage=427&rft.epage=431&rft.pages=427-431&rft.issn=0302-9743&rft.eissn=1611-3349&rft.isbn=9783540274414&rft.isbn_list=3540274413&rft_id=info:doi/10.1007/11520153_29&rft_dat=%3Cpascalfrancis_sprin%3E17027335%3C/pascalfrancis_sprin%3E%3Curl%3E%3C/url%3E&rft.eisbn=9783540318866&rft.eisbn_list=3540318860&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true