A Method on the Improvement of Speaker Enrolling Speed for a Multilayer Perceptron Based Speaker Verification System through Reducing Learning Data

다층신경망 기반 화자증명 시스템에서 학습 데이터 감축을 통한 화자등록속도 향상방법

  • 이백영 (한국항공대학교 항공전자공학과) ;
  • 황병원 (한국항공대학교 항공전자공학과) ;
  • 이태승 (한국항공대학교 항공전자공학과)
  • Published : 2002.08.01

Abstract

While the multilayer perceptron(MLP) provides several advantages against the existing pattern recognition methods, it requires relatively long time in learning. This results in prolonging speaker enrollment time with a speaker verification system that uses the MLP as a classifier. This paper proposes a method that shortens the enrollment time through adopting the cohort speakers method used in the existing parametric systems and reducing the number of background speakers required to learn the MLP, and confirms the effect of the method by showing the result of an experiment that applies the method to a continuant and MLP-based speaker verification system.

다층 신경망 (MLP: multilayer perceptron)은 기존의 패턴인식 방법에 비해 몇 가지 이점을 제공하지만 학습에 비교적 많은 시간을 요구한다. 이 점은 화자증명 시스템의 인식방법으로서 다층 신경망을 사용할 경우 등록시간이 길어지는 문제를 발생시킨다. 본 논문에서는 기존의 시스템에서 채택한 화자군집 방법을 응용하여 다층 신경망 학습에 필요한 배경화자 수를 줄임으로써 화자등록 시간을 단축하는 방법을 제안하고, 지속음을 인식단위로 하는 다층 신경망 화자증명 시스템에 이 방법을 적용한 실험결과를 통해 그 효과를 확인한다.

Keywords

References

  1. Automatic Speech and Speaker Recognition An Overview of Speaker Recognition Technology S.Furui
  2. Automatic Speech and Speaker Recognition Hybrid Connectionist Models for Continuous Speech Recognition N.Morgan;H.Bourlard
  3. Neural Networks S.Haykin
  4. Proceedings of the International Joint Conference on Neural Networks v.2 A modular connectionist architecture for text-independent talker identification Y.Bennani;P.Gallinari
  5. Proceeding of the IEEE International Conference on Acoustics, Speech and Signal Processing v.2 A high performance text independent speaker recognition system based on vowel Spotting and Neural Nets N.Fakotakis;J.Sirigos
  6. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing v.1 Speaker background models for connected digit password speaker verification A.E.Rosenberg;S.Parthasarathy https://doi.org/10.1109/ICASSP.1996.540295
  7. Speech Communication v.17 Likelihood normalization for speaker verification using a phoneme- and speaker-independent model T.Matsui;S.Furui https://doi.org/10.1016/0167-6393(95)00011-C
  8. Digital Signal Processing v.1 Speaker verification using randomized phrase prompting A.L.Higgins(et al) https://doi.org/10.1016/1051-2004(91)90098-6
  9. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing v.3 A probabilistic approach to the understanding and training of neural network classifiers H.Gish
  10. Speech Recognition C.Becchetti;L.P.Ricotti
  11. Proceedings of the IEEE International Conference on Electronics, Circuits and Systems v.3 New cepstrum frequency scale for neural network speaker verification P.Cristea;Z.Valsan
  12. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing v.2 Phoneme based speaker verification M.Savic;J.Sorensen
  13. IEEE Acoustics, Speech, and Signal Processing Magazine v.4 An Introduction to computing with neural nets R.P.Lippmann
  14. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing v.2 Text-prompted speaker verification experiments with phoneme specific MLPs D.P.Delacretaz;J.Hennebert