Robust speech recognition based on joint model and feature space optimization of hidden Markov models
The hidden Markov model (HMM) inversion algorithm, based on either the gradient search or the Baum-Welch reestimation of input speech features, is proposed and applied to the robust speech recognition tasks under general types of mismatch conditions. This algorithm stems from the gradient-based inve...
Saved in:
| Published in | IEEE transactions on neural networks Vol. 8; no. 2; pp. 194 - 204 |
|---|---|
| Main Authors | , |
| Format | Journal Article |
| Language | English |
| Published |
New York, NY
IEEE
01.03.1997
Institute of Electrical and Electronics Engineers |
| Subjects | |
| Online Access | Get full text |
| ISSN | 1045-9227 1941-0093 |
| DOI | 10.1109/72.557656 |
Cover
Loading…
| Abstract | The hidden Markov model (HMM) inversion algorithm, based on either the gradient search or the Baum-Welch reestimation of input speech features, is proposed and applied to the robust speech recognition tasks under general types of mismatch conditions. This algorithm stems from the gradient-based inversion algorithm of an artificial neural network (ANN) by viewing an HMM as a special type of ANN. Given input speech features s, the forward training of an HMM finds the model parameters /spl lambda/ subject to an optimization criterion. On the other hand, the inversion of an HMM finds speech features, s, subject to an optimization criterion with given model parameters /spl lambda/. The gradient-based HMM inversion and the Baum-Welch HMM inversion algorithms can be successfully integrated with the model space optimization techniques, such as the robust MINIMAX technique, to compensate the mismatch in the joint model and feature space. The joint space mismatch compensation technique achieves better performance than the single space, i.e. either the model space or the feature space alone, mismatch compensation techniques. It is also demonstrated that approximately 10-dB signal-to-noise ratio (SNR) gain is obtained in the low SNR environments when the joint model and feature space mismatch compensation technique is used. |
|---|---|
| AbstractList | The hidden Markov model (HMM) inversion algorithm, based on either the gradient search or the Baum-Welch reestimation of input speech features, is proposed and applied to the robust speech recognition tasks under general types of mismatch conditions. This algorithm stems from the gradient-based inversion algorithm of an artificial neural network (ANN) by viewing an HMM as a special type of ANN. Given input speech features s, the forward training of an HMM finds the model parameters lambda subject to an optimization criterion. On the other hand, the inversion of an HMM finds speech features, s, subject to an optimization criterion with given model parameters lambda. The gradient-based HMM inversion and the Baum-Welch HMM inversion algorithms can be successfully integrated with the model space optimization techniques, such as the robust MINIMAX technique, to compensate the mismatch in the joint model and feature space. The joint space mismatch compensation technique achieves better performance than the single space, i.e. either the model space or the feature space alone, mismatch compensation techniques. It is also demonstrated that approximately 10-dB signal-to-noise ratio (SNR) gain is obtained in the low SNR environments when the joint model and feature space mismatch compensation technique is used. The hidden Markov model (HMM) inversion algorithm, based on either the gradient search or the Baum-Welch reestimation of input speech features, is proposed and applied to the robust speech recognition tasks under general types of mismatch conditions. This algorithm stems from the gradient-based inversion algorithm of an artificial neural network (ANN) by viewing an HMM as a special type of ANN. Given input speech features s, the forward training of an HMM finds the model parameters /spl lambda/ subject to an optimization criterion. On the other hand, the inversion of an HMM finds speech features, s, subject to an optimization criterion with given model parameters /spl lambda/. The gradient-based HMM inversion and the Baum-Welch HMM inversion algorithms can be successfully integrated with the model space optimization techniques, such as the robust MINIMAX technique, to compensate the mismatch in the joint model and feature space. The joint space mismatch compensation technique achieves better performance than the single space, i.e. either the model space or the feature space alone, mismatch compensation techniques. It is also demonstrated that approximately 10-dB signal-to-noise ratio (SNR) gain is obtained in the low SNR environments when the joint model and feature space mismatch compensation technique is used. The hidden Markov model (HMM) inversion algorithm, based on either the gradient search or the Baum-Welch reestimation of input speech features, is proposed and applied to the robust speech recognition tasks under general types of mismatch conditions. This algorithm stems from the gradient-based inversion algorithm of an artificial neural network (ANN) by viewing an HMM as a special type of ANN. Given input speech features s, the forward training of an HMM finds the model parameters lambda subject to an optimization criterion. On the other hand, the inversion of an HMM finds speech features, s, subject to an optimization criterion with given model parameters lambda. The gradient-based HMM inversion and the Baum-Welch HMM inversion algorithms can be successfully integrated with the model space optimization techniques, such as the robust MINIMAX technique, to compensate the mismatch in the joint model and feature space. The joint space mismatch compensation technique achieves better performance than the single space, i.e. either the model space or the feature space alone, mismatch compensation techniques. It is also demonstrated that approximately 10-dB signal-to-noise ratio (SNR) gain is obtained in the low SNR environments when the joint model and feature space mismatch compensation technique is used.The hidden Markov model (HMM) inversion algorithm, based on either the gradient search or the Baum-Welch reestimation of input speech features, is proposed and applied to the robust speech recognition tasks under general types of mismatch conditions. This algorithm stems from the gradient-based inversion algorithm of an artificial neural network (ANN) by viewing an HMM as a special type of ANN. Given input speech features s, the forward training of an HMM finds the model parameters lambda subject to an optimization criterion. On the other hand, the inversion of an HMM finds speech features, s, subject to an optimization criterion with given model parameters lambda. The gradient-based HMM inversion and the Baum-Welch HMM inversion algorithms can be successfully integrated with the model space optimization techniques, such as the robust MINIMAX technique, to compensate the mismatch in the joint model and feature space. The joint space mismatch compensation technique achieves better performance than the single space, i.e. either the model space or the feature space alone, mismatch compensation techniques. It is also demonstrated that approximately 10-dB signal-to-noise ratio (SNR) gain is obtained in the low SNR environments when the joint model and feature space mismatch compensation technique is used. The hidden Markov model (HMM) inversion algorithm, based on either the gradient search or the Baum-Welch reestimation of input speech features, is proposed and applied to the robust speech recognition tasks under general types of mismatch conditions. This algorithm stems from the gradient-based inversion algorithm of an artificial neural network (ANN) by viewing an HMM as a special type of ANN. Given input speech features s, the forward training of an HMM finds the model parameters lambda subject to an optimization criterion. On the other hand, the inversion of an HMM finds speech features, s, subject to an optimization criterion with given model parameters lambda. The gradient-based HMM inversion and the Baum-Welch HMM inversion algorithms can be successfully integrated with the model space optimization techniques, such as the robust MINIMAX technique, to compensate the mismatch in the joint model and feature space. The joint space mismatch compensation technique achieves better performance than the single space, i.e. either the model space or the feature space alone, mismatch compensation techniques. It is also demonstrated that approximately 10-dB signal-to-noise ratio (SNR) gain is obtained in the low SNR environments when the joint model and feature space mismatch compensation technique is used The hidden Markov model (HMM) inversion algorithm, based on either the gradient search or the Baum-Welch reestimation of input speech features, is proposed and applied to the robust speech recognition tasks under general types of mismatch conditions. This algorith11m stems from the gradient-based inversion algorithm of an artificial neural network (ANN) by viewing an HMM as a special type of ANN. Given input speech features s, the forward training of an HMM finds the model parameters lambda subject to an optimization criterion. On the other hand, the inversion of an HMM finds speech features, s, subject to an optimization criterion with given model parameters lambda . The gradient-based HMM inversion and the Baum-Welch HMM inversion algorithms can be successfully integrated with the model space optimization techniques, such as the robust MINIMAX technique, to compensate the mismatch in the joint model and feature space. The joint space mismatch compensation technique achieves better performance than the single space, i.e., either the model space or the feature space alone, mismatch compensation techniques. It is also demonstrated that approximately 10-dB signal-to-noise ratio (SNR) gain is obtained in the low SNR environments when the joint model and feature space mismatch compensation technique is used. |
| Author | Seokyong Moon Jenq-Neng Hwang |
| Author_xml | – sequence: 1 givenname: S surname: MOON fullname: MOON, S organization: Communication Systems Research and Development, Samsung Electronics, Seoul, Korea, Republic of – sequence: 2 givenname: J.-N surname: HWANG fullname: HWANG, J.-N organization: Information Processing Laboratory, Department of Electrical Engineering, University of Washington, Seattle, WA 98195, United States |
| BackLink | http://pascal-francis.inist.fr/vibad/index.php?action=getRecordDetail&idt=2631536$$DView record in Pascal Francis https://www.ncbi.nlm.nih.gov/pubmed/18255624$$D View this record in MEDLINE/PubMed |
| BookMark | eNqF0UtrFjEUBuAgFXvRhVsXkoUoLqbNZXJbSvEGFUF0PWSSE5s6M_lMMoL-etPOZwURXeVAnveF5ByjgyUtgNBDSk4pJeZMsVMhlBTyDjqipqcdIYYftJn0ojOMqUN0XMoVIbQXRN5Dh1QzISTrjxB8SONaKi47AHeJM7j0eYk1pgWPtoDHbbhKcal4Th4mbBePA9i6ZmgZ6wCnXY1z_GFvMingy-g9LPidzV_Sty1V7qO7wU4FHuzPE_Tp1cuP52-6i_ev356_uOgcN6x2RjnbA4wk9FQS4NprHQRwIbW0VIGHoCkPRFvtjRGhvcFIoDBaUNT4wE_Qs613l9PXFUod5lgcTJNdIK1lULxnghrCm3z6T8m0YIor8n-oOKWGyQYf7-E6zuCHXY6zzd-HX3_dwJM9sMXZKWS7uFhuXauggl_3PN-Yy6mUDOF3Exmu1z0oNmzrbvbsD-tivdlEzTZOf0082hIRAG6b95c_AaGns50 |
| CODEN | ITNNEP |
| CitedBy_id | crossref_primary_10_1109_TMM_2005_843341 crossref_primary_10_1109_TSA_2002_805644 crossref_primary_10_1016_S0925_2312_00_00308_8 crossref_primary_10_1002__SICI_1098_111X_200005_15_5_365__AID_INT1_3_0_CO_2_P crossref_primary_10_1109_TCSVT_2004_826760 crossref_primary_10_1109_TNN_2002_1021891 crossref_primary_10_1109_TSP_2004_827153 crossref_primary_10_1109_5_880082 crossref_primary_10_1016_S1052_5149_25_00543_X crossref_primary_10_1109_TCBB_2016_2520919 crossref_primary_10_1016_j_entcom_2009_09_005 crossref_primary_10_1109_5_687838 crossref_primary_10_1109_TNN_2003_820838 crossref_primary_10_1109_LSP_2011_2179647 crossref_primary_10_1109_TSMCA_2010_2069094 crossref_primary_10_1186_1687_6180_2011_4 crossref_primary_10_1016_S0167_6393_00_00053_4 crossref_primary_10_1109_91_873579 crossref_primary_10_1016_j_neucom_2005_12_130 crossref_primary_10_1109_72_809074 crossref_primary_10_1109_LSP_2006_891326 crossref_primary_10_1109_TSA_2002_800558 |
| Cites_doi | 10.1109/89.221371 10.21236/ADA188529 10.1109/29.46548 10.1109/78.157214 10.1109/ICASSP.1993.319183 10.1109/IJCNN.1989.118277 10.1109/72.125870 10.1162/neco.1989.1.4.425 10.1109/ICASSP.1988.196643 10.1109/NNSP.1992.253704 10.1002/j.1538-7305.1985.tb00273.x 10.1016/0885-2308(91)90011-E 10.1109/TASSP.1979.1163209 10.1214/aoms/1177699803 10.1109/78.175747 10.1109/29.45543 10.1109/72.80299 10.1007/978-1-84628-615-5 10.1109/89.260341 10.1109/TASSP.1978.1163086 10.1109/36.239907 10.1109/89.260357 10.1109/NNSP.1991.239512 10.1109/PROC.1979.11540 10.1109/ICASSP.1990.115971 10.1109/72.125866 |
| ContentType | Journal Article |
| Copyright | 1997 INIST-CNRS |
| Copyright_xml | – notice: 1997 INIST-CNRS |
| DBID | AAYXX CITATION IQODW NPM 7SC 8FD JQ2 L7M L~C L~D 7X8 |
| DOI | 10.1109/72.557656 |
| DatabaseName | CrossRef Pascal-Francis PubMed Computer and Information Systems Abstracts Technology Research Database ProQuest Computer Science Collection Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Academic Computer and Information Systems Abstracts Professional MEDLINE - Academic |
| DatabaseTitle | CrossRef PubMed Computer and Information Systems Abstracts Technology Research Database Computer and Information Systems Abstracts – Academic Advanced Technologies Database with Aerospace ProQuest Computer Science Collection Computer and Information Systems Abstracts Professional MEDLINE - Academic |
| DatabaseTitleList | PubMed MEDLINE - Academic Computer and Information Systems Abstracts Computer and Information Systems Abstracts |
| Database_xml | – sequence: 1 dbid: NPM name: PubMed url: https://proxy.k.utb.cz/login?url=http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Engineering Anatomy & Physiology Computer Science Applied Sciences |
| EISSN | 1941-0093 |
| EndPage | 204 |
| ExternalDocumentID | 18255624 2631536 10_1109_72_557656 557656 |
| Genre | orig-research Journal Article |
| GroupedDBID | --- -~X .DC 0R~ 29I 4.4 53G 5GY 5VS 6IK 97E AAJGR AASAJ AAWTH ABAZT ABJNI ABQJQ ABVLG ACGFS AETIX AGQYO AGSQL AHBIQ AI. AIBXA ALLEH ALMA_UNASSIGNED_HOLDINGS ASUFR ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ CS3 DU5 EBS EJD F5P HZ~ H~9 ICLAB IFIPE IFJZH IPLJI JAVBF LAI M43 MS~ O9- OCL P2P RIA RIE RNS S10 TAE TN5 VH1 AAYOK AAYXX CITATION RIG IQODW NPM 7SC 8FD JQ2 L7M L~C L~D 7X8 |
| ID | FETCH-LOGICAL-c392t-97ca4eeb0f4160e38d88f5e35686a17edef813f08a8d995f55696e1ebae719df3 |
| IEDL.DBID | RIE |
| ISSN | 1045-9227 |
| IngestDate | Sun Aug 24 04:12:27 EDT 2025 Fri Jul 11 11:10:13 EDT 2025 Fri Jul 11 04:15:34 EDT 2025 Mon Jul 21 05:53:09 EDT 2025 Mon Jul 21 09:16:50 EDT 2025 Thu Apr 24 22:53:15 EDT 2025 Tue Jul 01 02:44:29 EDT 2025 Wed Aug 27 02:29:44 EDT 2025 |
| IsPeerReviewed | false |
| IsScholarly | false |
| Issue | 2 |
| Keywords | Statistical method Speech recognition Maximum likelihood Markov model Neural network Algorithm Gradient method Optimization Mean square error |
| Language | English |
| License | https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html CC BY 4.0 |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c392t-97ca4eeb0f4160e38d88f5e35686a17edef813f08a8d995f55696e1ebae719df3 |
| Notes | ObjectType-Article-2 SourceType-Scholarly Journals-1 ObjectType-Feature-1 content type line 23 ObjectType-Article-1 ObjectType-Feature-2 |
| PMID | 18255624 |
| PQID | 27311926 |
| PQPubID | 23500 |
| PageCount | 11 |
| ParticipantIDs | crossref_primary_10_1109_72_557656 proquest_miscellaneous_28527370 pascalfrancis_primary_2631536 pubmed_primary_18255624 crossref_citationtrail_10_1109_72_557656 proquest_miscellaneous_734251903 proquest_miscellaneous_27311926 ieee_primary_557656 |
| ProviderPackageCode | CITATION AAYXX |
| PublicationCentury | 1900 |
| PublicationDate | 1997-03-01 |
| PublicationDateYYYYMMDD | 1997-03-01 |
| PublicationDate_xml | – month: 03 year: 1997 text: 1997-03-01 day: 01 |
| PublicationDecade | 1990 |
| PublicationPlace | New York, NY |
| PublicationPlace_xml | – name: New York, NY – name: United States |
| PublicationTitle | IEEE transactions on neural networks |
| PublicationTitleAbbrev | TNN |
| PublicationTitleAlternate | IEEE Trans Neural Netw |
| PublicationYear | 1997 |
| Publisher | IEEE Institute of Electrical and Electronics Engineers |
| Publisher_xml | – name: IEEE – name: Institute of Electrical and Electronics Engineers |
| References | hwang (ref11) 1990 ref13 ref12 isidori (ref14) 1995 white (ref33) 1989; 1 ref32 ref10 brown (ref4) 1987 ref2 rabiner (ref30) 1993 ref1 ref17 ref16 ref19 rumelhart (ref31) 1986; 1 ref18 juang (ref15) 1985; 64 niles (ref29) 1991 lim (ref21) 1979; 67 ref24 ref23 ref20 ref22 moon (ref28) 1995 ref27 ref8 ref7 ref9 ref3 moon (ref25) 1992 ref6 ref5 moon (ref26) 1993; 2 |
| References_xml | – ident: ref24 doi: 10.1109/89.221371 – year: 1987 ident: ref4 publication-title: Acoustic-phonetic modeling problem in automatic speech recognition doi: 10.21236/ADA188529 – ident: ref23 doi: 10.1109/29.46548 – ident: ref17 doi: 10.1109/78.157214 – ident: ref27 doi: 10.1109/ICASSP.1993.319183 – year: 1993 ident: ref30 publication-title: Fundamentals of speech recognition – ident: ref22 doi: 10.1109/IJCNN.1989.118277 – ident: ref8 doi: 10.1109/72.125870 – year: 1991 ident: ref29 publication-title: Modeling and learning in speech recognition The relationship between stochastic pattern classifiers and neural networks – start-page: 145 year: 1995 ident: ref28 article-title: noisy speech recognition using robust inversion of hidden markov models publication-title: Proc EEE Int Conf Acoust Speech Signal Processing – volume: 1 start-page: 425 year: 1989 ident: ref33 article-title: learning in artificial neural networks: a statistical perspective publication-title: Neural Computa doi: 10.1162/neco.1989.1.4.425 – ident: ref32 doi: 10.1109/ICASSP.1988.196643 – ident: ref13 doi: 10.1109/NNSP.1992.253704 – start-page: 1086 year: 1992 ident: ref25 article-title: noisy speech recognition via wavelet coefficient enhancement publication-title: Proc 27th Asilomar Conf Signals Syst Comput – volume: 64 start-page: 1235 year: 1985 ident: ref15 article-title: maximum-likelihood estimation for mixture multivariate stochastic observations of markov chains publication-title: AT& T Tech J doi: 10.1002/j.1538-7305.1985.tb00273.x – ident: ref16 doi: 10.1016/0885-2308(91)90011-E – ident: ref3 doi: 10.1109/TASSP.1979.1163209 – ident: ref9 doi: 10.1214/aoms/1177699803 – ident: ref18 doi: 10.1109/78.175747 – ident: ref10 doi: 10.1109/29.45543 – ident: ref12 doi: 10.1109/72.80299 – year: 1995 ident: ref14 publication-title: Nonlinear Control Systems doi: 10.1007/978-1-84628-615-5 – ident: ref5 doi: 10.1109/89.260341 – volume: 2 year: 1993 ident: ref26 article-title: robust noisy speech enhancement using wavelets publication-title: Proc 1st Asia Pacific Conf Commun – volume: 1 start-page: 318 year: 1986 ident: ref31 article-title: learning internal representations by error propagation publication-title: Parallel Distributed Processing (PDP) Exploration in the Microstructure of Cognition – ident: ref20 doi: 10.1109/TASSP.1978.1163086 – ident: ref6 doi: 10.1109/36.239907 – ident: ref7 doi: 10.1109/89.260357 – start-page: 754 year: 1990 ident: ref11 article-title: iterative constrained inversion of neural networks and its applications publication-title: Proc 24th Conf Inform Syst Sci – ident: ref19 doi: 10.1109/NNSP.1991.239512 – volume: 67 start-page: 1586 year: 1979 ident: ref21 article-title: enhancement and bandwidth compression of noisy speech publication-title: Proceedings of the IEEE doi: 10.1109/PROC.1979.11540 – ident: ref1 doi: 10.1109/ICASSP.1990.115971 – ident: ref2 doi: 10.1109/72.125866 |
| SSID | ssj0014506 |
| Score | 1.3024223 |
| Snippet | The hidden Markov model (HMM) inversion algorithm, based on either the gradient search or the Baum-Welch reestimation of input speech features, is proposed and... |
| SourceID | proquest pubmed pascalfrancis crossref ieee |
| SourceType | Aggregation Database Index Database Enrichment Source Publisher |
| StartPage | 194 |
| SubjectTerms | Applied sciences Artificial neural networks Automatic speech recognition Exact sciences and technology Filtering Hidden Markov models Information, signal and communications theory Minimax techniques Moon Noise robustness Signal processing Speech enhancement Speech processing Speech recognition Telecommunications and information theory Wiener filter |
| Title | Robust speech recognition based on joint model and feature space optimization of hidden Markov models |
| URI | https://ieeexplore.ieee.org/document/557656 https://www.ncbi.nlm.nih.gov/pubmed/18255624 https://www.proquest.com/docview/27311926 https://www.proquest.com/docview/28527370 https://www.proquest.com/docview/734251903 |
| Volume | 8 |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwjV1Lb9QwELZoT3CgsOWxLQULIcQl28SOHzlWiKpCggOiUm-RY49VHk0qkiDRX1-PnU1b1CJuVjyWHHtsz9gz30fIG9sIdGUhEwopzKwus8aCzBTjkpsc4UMw3_nTZ3l0XH48EScTznbMhQGAGHwGKyzGt3zX2RGvyvZFMI6F3CAbwW9LqVrzg0EpIo1mcC5EVjGmJhChIq_2FVulhggQqhFyi5U3TqFIq4JBkaYP4-ITocXdFmc8eQ63Ukp3HwELMeDkx2ocmpW9-AvO8T9_6hF5OFmg9CCpzGNyD9oF2T5og_d99oe-pTEmNF62L8jWmvSBTnvAgjy4hmC4TeBL14z9QPtzAHtK53CkrqV4PjoaCt-7b-1AI-UONa2jHiKYaGhjLNAu7FlnUzIo7Tw9RUyTlmIOUfc7teqfkOPDD1_fH2UTc0Nmg701ZJWypgRoch_svRy4dlp7AVxILU2hwIHXBfe5NtpVlfBhTioJBTQGVFE5z5-SzbZr4TmhOTBQzHHLdVla5wyzSlilBEjudeOW5N16Jms7wZoju8bPOro3eVUrVqcxXpLXs-h5wvK4TWiBczQLrL_u3dCOuZpJHs6MUP9qrS11WJ_46GJa6Ma-DuZhEazof0loBMFT-ZLQOyQULzHBOOdL8ixp4lX_J4XeubXbu-R-gtvFmLkXZHP4NcJeMKKG5mVcPpe31RnL |
| linkProvider | IEEE |
| linkToHtml | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwjV1Nb9QwEB1BOUAPFLZQFii1EEJcsk3sOHaOFaJaoO0BtVJvkWOP1QJNKpIgwa_HH9lAUYu4WfFYcuyx59meeQPwStfcH2Ux4cKnMNMyT2qNRSIoK5hKPX2Ij3c-PCqWJ_mHU3468myHWBhEDM5nuPDF8JZvWj34q7Jd7sAxL27DHWf2eRaDtaYng5yHRJrueMGTklIx0ghlabkr6CI29RSh0pNu0fyKHQqJVbxbpOrcyNiY0uJmzBlsz_5GDOruAmWhdzn5shj6eqF__kXo-J-_9QDujxiU7EWleQi3sJnB5l7jzt8XP8hrErxCw3X7DDZWaR_IuAvMYP0PDsNNwE9tPXQ96S4R9RmZHJLahngLaYgrfG7Pm56EpDtENYZYDHSiro3SSFq3a12M4aCkteTMs5o0xEcRtd9jq-4RnOy_O367TMbcDYl2iKtPSqFVjlin1iG-FJk0UlqOjBeyUJlAg1ZmzKZSSVOW3Lo5KQvMsFYostJY9hjWmrbBJ0BSpCioYZrJPNfGKKoF10JwLJiVtZnDm9VMVnokNvf5Nb5W4YCTlpWgVRzjObycRC8jm8d1QjM_R5PA6uv2Fe2YqmnBnNVw9TsrbancCvXPLqrBdugqBxAzh6P_JSE9DZ5I50BukBAs9yHGKZvDVtTE3_0fFfrptd3egbvL48OD6uD90cdncC-S73oPuuew1n8bcNtBqr5-EZbSL4EnHRQ |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Robust+speech+recognition+based+on+joint+model+and+feature+space+optimization+of+hidden+Markov+models&rft.jtitle=IEEE+transactions+on+neural+networks&rft.au=Seokyong+Moon&rft.au=Jenq-Neng+Hwang&rft.date=1997-03-01&rft.issn=1045-9227&rft.eissn=1941-0093&rft.volume=8&rft.issue=2&rft.spage=194&rft.epage=204&rft_id=info:doi/10.1109%2F72.557656&rft.externalDBID=n%2Fa&rft.externalDocID=10_1109_72_557656 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1045-9227&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1045-9227&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1045-9227&client=summon |