Recognition of unvoiced stops from their time-frequency representation

Rangoussi, Maria; Delopoulos, Anastasios

dc.contributor.author	Rangoussi, Maria	en
dc.contributor.author	Delopoulos, Anastasios	en
dc.date.accessioned	2014-03-01T02:41:07Z
dc.date.available	2014-03-01T02:41:07Z
dc.date.issued	1995	en
dc.identifier.issn	07367791	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/30373
dc.subject	Learning Vector Quantization	en
dc.subject	Signal Extraction	en
dc.subject	Time Frequency Analysis	en
dc.subject	Time Frequency Representation	en
dc.subject	Wigner Distribution	en
dc.subject	Speaker Independent	en
dc.subject.other	Database systems	en
dc.subject.other	Feature extraction	en
dc.subject.other	Pattern recognition systems	en
dc.subject.other	Speech	en
dc.subject.other	Speech analysis	en
dc.subject.other	Classification (of information)	en
dc.subject.other	Frequency domain analysis	en
dc.subject.other	Speech processing	en
dc.subject.other	Time domain analysis	en
dc.subject.other	Vector quantization	en
dc.subject.other	Learning vector quantization	en
dc.subject.other	Speech database	en
dc.subject.other	Time frequency representation	en
dc.subject.other	Unvoiced stop signals	en
dc.subject.other	Learning vector quantization classifier	en
dc.subject.other	Phonemes	en
dc.subject.other	Unvoiced stops recognition	en
dc.subject.other	Wigner distribution	en
dc.subject.other	Speech recognition	en
dc.title	Recognition of unvoiced stops from their time-frequency representation	en
heal.type	conferenceItem	en
heal.identifier.primary	10.1109/ICASSP.1995.479813	en
heal.identifier.secondary	http://dx.doi.org/10.1109/ICASSP.1995.479813	en
heal.publicationDate	1995	en
heal.abstract	Recognition of the unvoiced stop sounds /k/, /p/ and /t/ in a speech signal is an interesting problem, due to the irregular, aperiodic, nonstationary nature of the corresponding signals. Their spotting is much easier, however, thanks to the characteristic silence interval they include. Classification of these three phonemes is therefore proposed in the present paper, based on patterns extracted from their time - frequency representation. This is possible because the different articulation points of /k/, /p/ and /t/ are reflected into distinct patterns of evolution of their spectral contents with time. These patterns can be obtained by suitable time - frequency analysis, and then used for classification. The Wigner distribution of the unvoiced stop signals, appropriately smoothed and subsampled, is proposed here as the basic classification pattern. Finally, for the classification step, the Learning Vector Quantization (LVQ) classifier of Kohonen is employed on a set of unvoiced stop signals extracted from the TIMIT speech database, with encouraging results under context- and speaker- independent testing conditions.	en
heal.publisher	IEEE, Piscataway, NJ, United States	en
heal.journalName	ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings	en
dc.identifier.doi	10.1109/ICASSP.1995.479813	en
dc.identifier.volume	1	en
dc.identifier.spage	792	en
dc.identifier.epage	795	en