Hybrid combination of knowledge- and cepstral-based features for phoneme recognition

Simple item page
dc.contributor.authorv.d. Merwe Rudolph
dc.contributor.authordu Preez Johan A.
dc.date.accessioned2011-05-15T15:57:29Z
dc.date.available2011-05-15T15:57:29Z
dc.date.issued1998
dc.description.abstractIn this paper a new, general, mathematically sound technique is developed to integrate knowledge-based information with standard cepstral features into the formal HMM framework for phoneme recognition. By using these hybrid features, the maximum amount of information contained in the speech signal can be utilized. It is shown that a trivial extension of the statistical models used to model the cepstral features, cannot be used to model the hybrid feature vectors, as this results in a decrease in phoneme recognition accuracy. By using the proposed hybrid technique though, a statistically significant increase in phoneme recognition accuracy is achieved.
dc.description.versionArticle
dc.identifier.citationProceedings of the South African Symposium on Communications and Signal Processing, COMSIG
dc.identifier.urihttp://hdl.handle.net/10019.1/10432
dc.subjectFeature extraction
dc.subjectKnowledge based systems
dc.subjectMarkov processes
dc.subjectMathematical models
dc.subjectProbability density function
dc.subjectSpeech analysis
dc.subjectStatistical methods
dc.subjectVectors
dc.subjectCepstral features
dc.subjectHidden Markov model (HMM)
dc.subjectPhoneme recognition
dc.subjectSpeech recognition
dc.titleHybrid combination of knowledge- and cepstral-based features for phoneme recognition
dc.typeArticle
Files
Collections
Stellenbosch University - Scopus Publications