Selection of the Best Wavelet Packet Nodes Based on Mutual Information for Speaker Identification
The analysis of the speech signal using wavelet packet trees (WPT) is a very flexible tool, capable of effectively manipulate the frequency subbands thanks to the orthonormal bases it provides. Here, dimension reduction becomes very important since the number of subbands grows exponentially with the level of decomposition, and their discriminative relevancy is different, which leads to different resolution for each one. A method based on mutual information is proposed in order to keep as much discriminative information as possible and the less amount of redundant information.
KeywordsWPT mutual information feature selection speaker identification
- 2.Kinnunen, T.: Spectral Features for Automatic Text-Independent Speaker Recognition. Licentiate’s thesis, University of Joensuu, Department of Computer Science, Joensuu, Finland (2004)Google Scholar
- 4.Farooq, O., Datta, S.: Mel-scaled wavelet filter based features for noisy unvoiced phoneme recognition. In: International Conference on Spoken Language Processing ICSLP, pp. 1017–1020 (2002)Google Scholar
- 9.Siafarikas, M., Ganchev, T., Fakotakis, N.: Wavelet Packet Based Speaker Verification. In: Ortega-Garcia, J., et al. (eds.) The Speaker and Language Recognition Workshop ODYSSEY (2004)Google Scholar
- 12.Lu, X., Dang, J.: Dimension reduction for speaker identification based on mutual information. In: Interspeech, pp. 2021–2024 (2007)Google Scholar