High-performance low-complexity wordspotting using neural networks

A high-performance low-complexity neural network wordspotter was developed using radial basis function (RBF) neutral networks in a hidden Markov model (HMM) framework. Two new complementary approaches substantially improve performance on the talker-independent Switchboard corpus. Figure of merit (FO...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on signal processing 1997-11, Vol.45 (11), p.2864-2870
Hauptverfasser: Chang, E.I., Lippmann, R.P.
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page 2870
container_issue 11
container_start_page 2864
container_title IEEE transactions on signal processing
container_volume 45
creator Chang, E.I.
Lippmann, R.P.
description A high-performance low-complexity neural network wordspotter was developed using radial basis function (RBF) neutral networks in a hidden Markov model (HMM) framework. Two new complementary approaches substantially improve performance on the talker-independent Switchboard corpus. Figure of merit (FOM) training adapts wordspotter parameters to directly improve the FOM performance metric, and voice transformations generate additional training examples by warping the spectra of training data to mimic across-talker vocal tract length variability.
doi_str_mv 10.1109/78.650114
format Article
fullrecord <record><control><sourceid>proquest_RIE</sourceid><recordid>TN_cdi_proquest_miscellaneous_28332069</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>650114</ieee_id><sourcerecordid>28332069</sourcerecordid><originalsourceid>FETCH-LOGICAL-c266t-ed096e2479630da036ddb057100f6994333f255a156293a961f56f912d6c67953</originalsourceid><addsrcrecordid>eNo9kL1PwzAUxC0EEqUwsDJ1QEgMKc9fz_EIFVCkSiwgsUUmsUsgiYOdqPS_J1WqLu-edL-74Qi5pDCnFPSdSucogVJxRCZUC5qAUHg8_CB5IlP1cUrOYvwGoEJonJCHZbn-SlobnA-1aXI7q_wmyX3dVvav7LazjQ9FbH3Xlc161sfdbWwfTDVIN5g_8ZycOFNFe7HXKXl_enxbLJPV6_PL4n6V5AyxS2wBGi0TSiOHwgDHovgEqSiAQ60F59wxKQ2VyDQ3GqmT6DRlBeaotORTcjP2tsH_9jZ2WV3G3FaVaazvY8ZSzhmgHsDbEcyDjzFYl7WhrE3YZhSy3UqZSrNxpYG93peamJvKhWGDMh4CDJRAmg7Y1YiV1tqDu-_4B0Arbf0</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>28332069</pqid></control><display><type>article</type><title>High-performance low-complexity wordspotting using neural networks</title><source>IEEE Electronic Library (IEL)</source><creator>Chang, E.I. ; Lippmann, R.P.</creator><creatorcontrib>Chang, E.I. ; Lippmann, R.P.</creatorcontrib><description>A high-performance low-complexity neural network wordspotter was developed using radial basis function (RBF) neutral networks in a hidden Markov model (HMM) framework. Two new complementary approaches substantially improve performance on the talker-independent Switchboard corpus. Figure of merit (FOM) training adapts wordspotter parameters to directly improve the FOM performance metric, and voice transformations generate additional training examples by warping the spectra of training data to mimic across-talker vocal tract length variability.</description><identifier>ISSN: 1053-587X</identifier><identifier>EISSN: 1941-0476</identifier><identifier>DOI: 10.1109/78.650114</identifier><identifier>CODEN: ITPRED</identifier><language>eng</language><publisher>New York, NY: IEEE</publisher><subject>Acoustic signal detection ; Applied sciences ; Artificial intelligence ; Computer science; control theory; systems ; Connectionism. Neural networks ; Control systems ; Exact sciences and technology ; Hidden Markov models ; Maximum likelihood detection ; Measurement ; Neural networks ; NIST ; Speech ; Speech and sound recognition and synthesis. Linguistics ; Testing ; Training data</subject><ispartof>IEEE transactions on signal processing, 1997-11, Vol.45 (11), p.2864-2870</ispartof><rights>1998 INIST-CNRS</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><cites>FETCH-LOGICAL-c266t-ed096e2479630da036ddb057100f6994333f255a156293a961f56f912d6c67953</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/650114$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,780,784,796,27924,27925,54758</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/650114$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc><backlink>$$Uhttp://pascal-francis.inist.fr/vibad/index.php?action=getRecordDetail&amp;idt=2074618$$DView record in Pascal Francis$$Hfree_for_read</backlink></links><search><creatorcontrib>Chang, E.I.</creatorcontrib><creatorcontrib>Lippmann, R.P.</creatorcontrib><title>High-performance low-complexity wordspotting using neural networks</title><title>IEEE transactions on signal processing</title><addtitle>TSP</addtitle><description>A high-performance low-complexity neural network wordspotter was developed using radial basis function (RBF) neutral networks in a hidden Markov model (HMM) framework. Two new complementary approaches substantially improve performance on the talker-independent Switchboard corpus. Figure of merit (FOM) training adapts wordspotter parameters to directly improve the FOM performance metric, and voice transformations generate additional training examples by warping the spectra of training data to mimic across-talker vocal tract length variability.</description><subject>Acoustic signal detection</subject><subject>Applied sciences</subject><subject>Artificial intelligence</subject><subject>Computer science; control theory; systems</subject><subject>Connectionism. Neural networks</subject><subject>Control systems</subject><subject>Exact sciences and technology</subject><subject>Hidden Markov models</subject><subject>Maximum likelihood detection</subject><subject>Measurement</subject><subject>Neural networks</subject><subject>NIST</subject><subject>Speech</subject><subject>Speech and sound recognition and synthesis. Linguistics</subject><subject>Testing</subject><subject>Training data</subject><issn>1053-587X</issn><issn>1941-0476</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>1997</creationdate><recordtype>article</recordtype><recordid>eNo9kL1PwzAUxC0EEqUwsDJ1QEgMKc9fz_EIFVCkSiwgsUUmsUsgiYOdqPS_J1WqLu-edL-74Qi5pDCnFPSdSucogVJxRCZUC5qAUHg8_CB5IlP1cUrOYvwGoEJonJCHZbn-SlobnA-1aXI7q_wmyX3dVvav7LazjQ9FbH3Xlc161sfdbWwfTDVIN5g_8ZycOFNFe7HXKXl_enxbLJPV6_PL4n6V5AyxS2wBGi0TSiOHwgDHovgEqSiAQ60F59wxKQ2VyDQ3GqmT6DRlBeaotORTcjP2tsH_9jZ2WV3G3FaVaazvY8ZSzhmgHsDbEcyDjzFYl7WhrE3YZhSy3UqZSrNxpYG93peamJvKhWGDMh4CDJRAmg7Y1YiV1tqDu-_4B0Arbf0</recordid><startdate>19971101</startdate><enddate>19971101</enddate><creator>Chang, E.I.</creator><creator>Lippmann, R.P.</creator><general>IEEE</general><general>Institute of Electrical and Electronics Engineers</general><scope>IQODW</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7SC</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope></search><sort><creationdate>19971101</creationdate><title>High-performance low-complexity wordspotting using neural networks</title><author>Chang, E.I. ; Lippmann, R.P.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c266t-ed096e2479630da036ddb057100f6994333f255a156293a961f56f912d6c67953</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>1997</creationdate><topic>Acoustic signal detection</topic><topic>Applied sciences</topic><topic>Artificial intelligence</topic><topic>Computer science; control theory; systems</topic><topic>Connectionism. Neural networks</topic><topic>Control systems</topic><topic>Exact sciences and technology</topic><topic>Hidden Markov models</topic><topic>Maximum likelihood detection</topic><topic>Measurement</topic><topic>Neural networks</topic><topic>NIST</topic><topic>Speech</topic><topic>Speech and sound recognition and synthesis. Linguistics</topic><topic>Testing</topic><topic>Training data</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Chang, E.I.</creatorcontrib><creatorcontrib>Lippmann, R.P.</creatorcontrib><collection>Pascal-Francis</collection><collection>CrossRef</collection><collection>Computer and Information Systems Abstracts</collection><collection>Technology Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts – Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><jtitle>IEEE transactions on signal processing</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Chang, E.I.</au><au>Lippmann, R.P.</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>High-performance low-complexity wordspotting using neural networks</atitle><jtitle>IEEE transactions on signal processing</jtitle><stitle>TSP</stitle><date>1997-11-01</date><risdate>1997</risdate><volume>45</volume><issue>11</issue><spage>2864</spage><epage>2870</epage><pages>2864-2870</pages><issn>1053-587X</issn><eissn>1941-0476</eissn><coden>ITPRED</coden><abstract>A high-performance low-complexity neural network wordspotter was developed using radial basis function (RBF) neutral networks in a hidden Markov model (HMM) framework. Two new complementary approaches substantially improve performance on the talker-independent Switchboard corpus. Figure of merit (FOM) training adapts wordspotter parameters to directly improve the FOM performance metric, and voice transformations generate additional training examples by warping the spectra of training data to mimic across-talker vocal tract length variability.</abstract><cop>New York, NY</cop><pub>IEEE</pub><doi>10.1109/78.650114</doi><tpages>7</tpages></addata></record>
fulltext fulltext_linktorsrc
identifier ISSN: 1053-587X
ispartof IEEE transactions on signal processing, 1997-11, Vol.45 (11), p.2864-2870
issn 1053-587X
1941-0476
language eng
recordid cdi_proquest_miscellaneous_28332069
source IEEE Electronic Library (IEL)
subjects Acoustic signal detection
Applied sciences
Artificial intelligence
Computer science
control theory
systems
Connectionism. Neural networks
Control systems
Exact sciences and technology
Hidden Markov models
Maximum likelihood detection
Measurement
Neural networks
NIST
Speech
Speech and sound recognition and synthesis. Linguistics
Testing
Training data
title High-performance low-complexity wordspotting using neural networks
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-07T21%3A39%3A14IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_RIE&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=High-performance%20low-complexity%20wordspotting%20using%20neural%20networks&rft.jtitle=IEEE%20transactions%20on%20signal%20processing&rft.au=Chang,%20E.I.&rft.date=1997-11-01&rft.volume=45&rft.issue=11&rft.spage=2864&rft.epage=2870&rft.pages=2864-2870&rft.issn=1053-587X&rft.eissn=1941-0476&rft.coden=ITPRED&rft_id=info:doi/10.1109/78.650114&rft_dat=%3Cproquest_RIE%3E28332069%3C/proquest_RIE%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=28332069&rft_id=info:pmid/&rft_ieee_id=650114&rfr_iscdi=true