Steidl S, Stemmer G, Hacker C, Nöth E, Niemann H (2003)
Publication Language: English
Publication Type: Conference contribution, Conference Contribution
Publication year: 2003
Original Authors: Steidl Stefan, Stemmer Georg, Hacker Christian, Nöth Elmar, Niemann Heinrich
Publisher: Springer-Verlag
City/Town: Berlin, Heidelberg, New York
Pages Range: 600-607
Conference Proceedings Title: Pattern Recognition, 25th DAGM Symposium, Magdeburg, Germany, September 2003, Proceedings
URI: http://www5.informatik.uni-erlangen.de/Forschung/Publikationen/2003/Steidl03-ICS.pdf
In this paper we address the problem of building a good speech recognizer if there is only a small amount of training data available. The acoustic models can be improved by interpolation with the well-trained models of a second recognizer from a different application scenario. In our case, we interpolate a children’s speech recognizer with a recognizer for adults’ speech. Each hidden Markov model has its ownset of interpolation partners; experiments were conducted with up to 50 partners. The interpolation weights are estimated automatically on a validation set using the EM algorithm. The word accuracy of the children’s speech recognizer could be improved from 74.6 % to 81.5 %. This is a relative improvement of almost 10 %.
APA:
Steidl, S., Stemmer, G., Hacker, C., Nöth, E., & Niemann, H. (2003). Improving Children's Speech Recognition by HMM Interpolation with an Adults' Speech Recognizer. In Michaelis Bernd, Krell Gerald (Eds.), Pattern Recognition, 25th DAGM Symposium, Magdeburg, Germany, September 2003, Proceedings (pp. 600-607). Magdeburg, DE: Berlin, Heidelberg, New York: Springer-Verlag.
MLA:
Steidl, Stefan, et al. "Improving Children's Speech Recognition by HMM Interpolation with an Adults' Speech Recognizer." Proceedings of the DAGM '03, 25th Pattern Recognition Symposium, Magdeburg Ed. Michaelis Bernd, Krell Gerald, Berlin, Heidelberg, New York: Springer-Verlag, 2003. 600-607.
BibTeX: Download