Kernel eigenvoice

Speaker adaptation is an important technology to fine-tune either features or speech models for mis-match due to inter-speaker variation. In the last decade, eigenvoice (EV) speaker adaptation has been developed. It makes use of the prior knowledge of training speakers to provide a fast adaptation algorithm (in other words, only a small amount of adaptation data is needed). Inspired by the kernel eigenface idea in face recognition, kernel eigenvoice (KEV) is proposed.[1] KEV is a non-linear generalization to EV. This incorporates Kernel principal component analysis, a non-linear version of Principal Component Analysis, to capture higher order correlations in order to further explore the speaker space and enhance recognition performance.

See also

References

  • Kernel Eigenvoice Speaker Adaptation, ScientificCommons
  • Mak, B.; Ho, S. (2005). "Various Reference Speakers Determination Methods for Embedded Kernel Eigenvoice Speaker Adaptation". IEEE International Conference on Acoustics, Speech, and Signal Processing, 2005. Proceedings. ICASSP '05. 1. pp. 981–984. doi:10.1109/ICASSP.2005.1415280.
  • Mak, B.; Kwok, J. T.; Ho, S. (September 2005). "Kernel Eigenvoice Speaker Adaptation". IEEE Transactions on Speech and Audio Processing. 13 (5): 984–992. doi:10.1109/TSA.2005.851971. ISSN 1063-6676. Retrieved 2017-11-15.
  • Speedup of Kernel Eigenvoice Speaker Adaptation by Embedded Kernel PCA, ICSLP 2004.
  • Speaker Adaptation via Composite Kernel PCA, NIPS 2003.
  • Mak, Brian Kan-Wing; Hsiao, Roger Wend-Huu; Ho, Simon Ka-Lung; Kwok, J. T. (July 2006). "Embedded kernel eigenvoice speaker adaptation and its implication to reference speaker weighting". IEEE Transactions on Audio, Speech, and Language Processing. 14 (4): 1267–1280. doi:10.1109/TSA.2005.860836.
This article is issued from Wikipedia. The text is licensed under Creative Commons - Attribution - Sharealike. Additional terms may apply for the media files.