Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Constrained subword units for speaker recognition

: Baum, D.; Schneider, D.; Mertens, T.; Köhler, J.

Postprint urn:nbn:de:0011-n-1513673 (103 KByte PDF)
MD5 Fingerprint: e0d9a519bfc60ce4151c32abf0bed2b9
Erstellt am: 27.1.2011

International Speech Communication Association -ISCA-:
Odyssey 2010 - The Speaker and Language Recognition Workshop : Brno, Czech Republic, 28 June-1 July 2010
Raleigh: ISCA, 2010
Paper 02, 6 S.
Odyssey - The Speaker and Language Recognition Workshop <2010, Brno>
Konferenzbeitrag, Elektronische Publikation
Fraunhofer IAIS ()

Phonetic features have been proposed to overcome performance degradation in spectral speaker recognition in difficult acoustic conditions. The harmful effect of those conditions, however, is not restricted to spectral systems but also affects the performance of the open-loop phone recognisers on which phonetic systems are based. In automatic speech recognition, larger subword units and the use of additional constraints from language models have been employed to improve robustness against adverse acoustic conditions. This paper evaluates the performance of more constrained phone recognition and different subword units for speaker recognition on heterogeneous broadcast data from German parliamentary speeches. Using phone clusters and a strong language model instead of phones obtained from unconstrained recognition improves the equal error rate from 14.3% to 8.6% on the given data.