Publication Details
Some like it Gaussian...
Schwarz Petr, Ing., Ph.D. (DCGM FIT BUT)
Karafiát Martin, Ing., Ph.D. (DCGM FIT BUT)
Černocký Jan, prof. Dr. Ing. (DCGM FIT BUT)
speech recognition, feature extraction, Gaussianization, non-linear transform
Gaussianization of speech features to improve recognition accuracy
In Hidden Markov models, speech features are modeled by Gaussian distributions. In this paper, we propose to gaussianize the features to better fit to this modeling. A distribution of the data is estimated and a transform function is derived. We have tested two methods of the transform estimation (global and speaker based). The results are reported on recognition of isolated Czech words (SpeechDat-E) with CI and CD models and on medium vocabulary continuous speech recognition task (SPINE). Gaussianized data provided in all three cases results superior to standard MFC coefficients proving, that the gaussianization is a cheap way to increase the recognition accuracy
@INPROCEEDINGS{FITPUB7025, author = "Pavel Mat\v{e}jka and Petr Schwarz and Martin Karafi\'{a}t and Jan \v{C}ernock\'{y}", title = "Some like it Gaussian...", pages = "321--324", booktitle = "Proc. 5th International Conference Text, Speech and Dialogue, TSD2002", series = "Lecture notes in artificial intelligence 2448", year = 2002, location = "Berlin, DE", publisher = "Springer Verlag", ISBN = "3-540-44129-8", language = "english", url = "https://www.fit.vut.cz/research/publication/7025" }