Search by:
Year of publication
Author name
Paper title
Speech Signal Segmentation Based on the Speaker Change
Full text (PDF)
UDC: 004.8
Publication Language: Ukrainian
Stuc. intelekt. 2011; 16(3):167-173
Abstract: An approach to the segmentation of speech signals based on the speaker change, as well as to the detection of the speaker change positions in a speech signal is suggested. Speaker change positions are determined by analyzing the sets of characteristic vectors at the pause within the signal based on the Bayesian information criterion. Improvement in quality of the characteristic vectors is achieved by taking into account only the segments with the log energy above the given threshold. It is also suggested an approach for adaptive automatic pause detection in speech signal.
Keywords:
References:
- Krivonos Ju.G. Shtuchnij іntelekt. № 3. 2009. S. 228-233.
- Kotti M., Benetos E., Kotropoulos C. Automatic Speaker Change Detection with the Bayesian InformationCriterion using MPEG-7 Features and a Fusion Scheme. Proc. of ISCAS 2006. http://poseidon.csd.auth.gr/papers/PUBLISHED/CONFERENCE/pdf/Kotti06b.pdf
- L. Lu. Proceedings of the tenth ACM international conference on Multimediaju. 2002. P. 602-610.
- T.Y. Wu. Microsoft Research. http://research.microsoft.com/users/llu/publications/mmm03_ubmforspkseg.pdf.
- S. Kwon, S. Proc. of International conference on spoken language processing. 2002. Vol. 4. P. 2537-2540.
- Krivonos Ju.G. Shtuchnij іntelekt. № 3. 2010. S. 220-226.
- Ajmera J. IEEE Signal Processing Letters. № 8, vol. 11. 2004. P. 649-651.
- Rabiner L. IEEE Transaction on Acoustics, Speech and Signal Processing. №7. Vol. 25. 1977. P. 338-343.
- Tanyer G.S. IEEE Transactions on Speech and Audio Processing. №4. Vol. 8. 2000. P. 478-482.
- Zagvazdіn O.S. Zhurnal obchysljuval'noi ta prikladnoi matematyky. № 2. 2010. S. 35-43.
- Rabilner L. IEEE Transactions