Sie befinden Sich nicht im Netzwerk der Universität Paderborn. Der Zugriff auf elektronische Ressourcen ist gegebenenfalls nur via VPN oder Shibboleth (DFN-AAI) möglich. mehr Informationen...
Noisy Speech Segmentation/Enhancement with Multiband Analysis and Neural Fuzzy Networks
Ist Teil von
Advances in Soft Computing — AFSS 2002, 2002, p.301-309
Ort / Verlag
Berlin, Heidelberg: Springer Berlin Heidelberg
Erscheinungsjahr
2002
Quelle
Alma/SFX Local Collection
Beschreibungen/Notizen
Background noise added to speech can decrease the performance of speech segmentation and enhancement. To solve this problem, new methods have been developed in this thesis. First, a new speech segmentation method (ATF-based SONFIN algorithm) is proposed in fixed noise-level environment. This method contains the multiband analysis and a neural fuzzy network, and it achieves higher recognition rate than the TF-based robust algorithm by 5%. In addition, a new speech segmentation method called RTF-based RSONFIN algorithm is proposed for variable noise-level environment. The RTF-based RSONFIN algorithm contains a recurrent neural fuzzy network. This method contains the multiband analysis and achieve higher recognition rate than the TFbased robust algorithm by 12%.