Stereo Vision Neural Networks with Competition and Cooperation for Phoneme Recognition

  • Kim, Sung-Ill (Division of Electrical and Electronic Engineering, Kyungnam University) ;
  • Chung, Hyun-Yeol (School of Electrical Engineering and Computer Science, Yeungnam University)
  • Published : 2003.03.01

Abstract

This paper describes two kinds of neural networks for stereoscopic vision, which have been applied to an identification of human speech. In speech recognition based on the stereoscopic vision neural networks (SVNN), the similarities are first obtained by comparing input vocal signals with standard models. They are then given to a dynamic process in which both competitive and cooperative processes are conducted among neighboring similarities. Through the dynamic processes, only one winner neuron is finally detected. In a comparative study, with, the average phoneme recognition accuracy on the two-layered SVNN was 7.7% higher than the Hidden Markov Model (HMM) recognizer with the structure of a single mixture and three states, and the three-layered was 6.6% higher. Therefore, it was noticed that SVNN outperformed the existing HMM recognizer in phoneme recognition.

Keywords

References

  1. P. C. Woodland, C. J. Leggestter, J. J. Odell, et.al., 'The 1994 HTK Large Vocabulary Speech Recognition System,' Proc. IEEE Infernational Conference on Acoustics, Speech, and Signal Processing, 1, 73-76, 1995
  2. X. D. Huang, Y. Ariki, M. A. Jack, Hidden Makov Models for Speech Recognition, Edinburgh University Press, Edinburgh, U.K., 1990
  3. K. F. Lee, H. W. Hon, 'Speaker-lndependent Phone Reconition Using Hidden Markov Models,' IEEE Transacfion on Acoustic, Speech and Signal Processing. 37 (11), 641-1648, 1989
  4. H. Bourlard and C. J. Wellekens, 'Links between Markov Models and Multi-layer Perceptrons,' IEEE Transaction Patt. Anal. Machine Intell., 12, 1167-1178, 1990 https://doi.org/10.1109/34.62605
  5. J. Lang, A. Waibel and G. E. Hinton, 'A Time-DeIay Neural Network Architecture for Isolated Word Recognition,' Artificial Neural Networks, Paradigms, Applications and Hardware Implementations, IEEE press, New York, 388-408, 1992
  6. G. Martinelli, 'Hidden Control Neural Network,' IEEE Transaction on Circuits and Systems, Analog and Signal Processing, 41 (3), 245-247, 1994
  7. D. Reimann, T. Ditzinger, E. Fischer and H. Haken, 'Vergence eye movement control and multivalent perception of Autostereograms,' Biol. Cybern., 73, 123-128, 1995 https://doi.org/10.1007/BF00204050
  8. D. Reinmann and H. Haken, 'Stereo Vision by Self-organization,' Biol. Cybern., 71, 17-26, 1994 https://doi.org/10.1007/BF00198908
  9. S. Amari and M. A. Arbib, 'Competition and Cooperation in Neural Nets,' Systems Neuroscience, Academic Press, 119-165, 1977
  10. Y. Yoshitomi, T. Kanda, T. Kitazoe, 'Neural Nets Pattern Recognition Equation for Stereo Vision,' Trans. IPS, 29-38, 1998
  11. Y. Yoshitomi, T. Kitazoe, J. Tomiyama, Y. Tatebe, 'Sequential stereo Vision and Phase Transition,' Proc. Third International Symposium on Artificial Life, and Robotics, 318-323, 1998
  12. T. Kitazoe, J. Tomiyama, Y. Yoshitomi et al., 'Sequential Stereoscopic Vision and Hysteresis,' Proc. Neural Information Processing, 391-396, 1998