GENERAL ENQUIRIES: Tel: + 27 12 841 2911 | Email: callcentre@csir.co.za

Show simple item record

dc.contributor.author Feld, M
dc.contributor.author Barnard, E
dc.contributor.author Van Heerden, C
dc.contributor.author Muller, C
dc.date.accessioned 2012-01-18T13:38:09Z
dc.date.available 2012-01-18T13:38:09Z
dc.date.issued 2009-12
dc.identifier.citation Feld, M, Barnard, E, Van Heerden, C and Muller, C. 2009. Multilingual speaker age recognition: regression analyses on the Lwazi corpus. 2009 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU-09), Merano, Italy, 13-17 December 2009 en_US
dc.identifier.isbn 978-1-4244-5479-2
dc.identifier.uri http://hdl.handle.net/10204/5506
dc.description 2009 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU-09), Merano, Italy, 13-17 December 2009 en_US
dc.description.abstract Multilinguality represents an area of significant opportunities for automatic speech-processing systems: whereas multilingual societies are commonplace, the majority of speechprocessing systems are developed with a single language in mind. As a step towards improved understanding of multilingual speech processing, the current contribution investigates how an important para-linguistic aspect of speech, namely speaker age, depends on the language spoken. In particular, the authors study how certain speech features affect the performance of an age recognition system for different South African languages in the Lwazi corpus. By optimizing our feature set and performing language-specific tuning, we are working towards true multilingual classifiers. As they are closely related, ASR and dialog systems are likely to benefit from an improved classification of the speaker. In a comprehensive corpus analysis on long-term features, we have identified features that exhibit characteristic behaviors for particular languages. In a follow-up regression experiment, we confirm the suitability of our feature selection for age recognition and present cross-language error rates. The mean absolute error ranges between 7.7 and 12.8 years for same-language predictors and rises to 14.5 years for cross-language predictors. en_US
dc.language.iso en en_US
dc.publisher IEEE en_US
dc.subject Automatic speech recognition system en_US
dc.subject ASR en_US
dc.subject Lwazi ASR en_US
dc.subject Lwazi corpus en_US
dc.subject Germanic languages en_US
dc.subject Bantu languages en_US
dc.subject Automatic speech-processing systems en_US
dc.subject Speaker age en_US
dc.subject South Africa en_US
dc.subject Multilingual corpus en_US
dc.subject Speaker classification en_US
dc.title Multilingual speaker age recognition: regression analyses on the Lwazi corpus en_US
dc.type Article en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record

Search ResearchSpace


Advanced Search

Browse

My Account