论文信息 - Robust model based methods for speech enhancement

Robust model based methods for speech enhancement

The paper compares the performance of Kalman filter and hidden Markov model approaches for the recovery of speech from FM signals within impulsive noise environments. Two applications of speech recovery are investigated, namely within demodulation and post demodulation. Simulation studies are presented which illustrate the best case and worst case performance trends.

Langford B. White | G. A. Einicke

[1] James W. Modestino,et al. Simulated Performance of an Adaptive Multielement Integrated Receiver Structure in Impulse or Burst Noise , 1987, IEEE Trans. Commun..

[2] S. Kay. Fundamentals of statistical signal processing: estimation theory , 1993 .

[3] Seong Rag Kim,et al. Adaptive robust impulse noise filtering , 1995, IEEE Trans. Signal Process..

[4] John B. Moore,et al. On-line estimation of hidden Markov model parameters based on the Kullback-Leibler information measure , 1993, IEEE Trans. Signal Process..

[5] Brian D. O. Anderson,et al. Characterization of threshold for single tone maximum likelihood frequency estimation , 1995, IEEE Trans. Signal Process..

[6] Kenneth J. Kerpez. Minimum mean squared error impulse noise estimation and cancellation , 1995, IEEE Trans. Signal Process..