000082955 001__ 82955
000082955 005__ 20181120161419.0
000082955 037__ $$aARTICLE
000082955 245__ $$aComparison and Combination of Features in a Hybrid HMM/MLP and a HMM/GMM Speech Recognition System
000082955 269__ $$a2003
000082955 260__ $$c2003
000082955 336__ $$aJournal Articles
000082955 500__ $$aIDIAP-RR 03-48
000082955 520__ $$aRecently, the advantages of the spectral parameters obtained by frequency filtering (FF) of the logarithmic filter-bank energies (logFBEs) have been reported. These parameters, which are frequency derivatives of the lofFBEs, lie in the frequency domain, and have shown good recognition performance with repect to the conventional MFCCs for HMM systems. In this paper, the FF features are first compared with the MFCCs and the Rasta-PLP features using both a hybrid HMM/MLP and a usual HMM/GMM recognition system, for both clean and noisy speech. Taking advantage of the ability of the hybrid system to deal with correlated features, the inclusion of both the frequency second-derivatives and the raw logFBes as additional features is proposed and tested. Moreover, the robustness of these features in noisy conditions is enhanced by combining the FF technique with the Rasta temporal filtering approach. Finally, a study of the FF features in the framework of multi-stram processing is presented. The best recognition results for both clean and noisy speech are obtained from the multi-stream combination of the J-Rasta-PLP features and the FF features.
000082955 6531_ $$aspeech
000082955 700__ $$aPujol, Pere
000082955 700__ $$aPol, Susagna
000082955 700__ $$aNadeu, Climent
000082955 700__ $$aHagen, Astrid
000082955 700__ $$0243348$$aBourlard, Hervé$$g117014
000082955 773__ $$k48$$tto be published in IEEE Transactions on Speech and Audio Processing
000082955 8564_ $$uhttp://publications.idiap.ch/downloads/reports/2003/rr03-48.pdf$$zURL
000082955 8564_ $$uhttp://publications.idiap.ch/index.php/publications/showcite/bourlard-03-48$$zRelated documents
000082955 8564_ $$s601652$$uhttps://infoscience.epfl.ch/record/82955/files/rr03-48.pdf$$zn/a
000082955 909C0 $$0252189$$pLIDIAP$$xU10381
000082955 909CO $$ooai:infoscience.tind.io:82955$$pSTI$$particle$$qGLOBAL_SET
000082955 937__ $$aEPFL-ARTICLE-82955
000082955 970__ $$abourlardRR48art/LIDIAP
000082955 973__ $$aEPFL$$sPUBLISHED
000082955 980__ $$aARTICLE