The problems of efficacy estimation in noise cancelling by a neural net (LIN-Lateral Inhibition Net [5]) and the use of this information in weighting matching algorithms are focused. Since the effect of noise on the speech signal is variable and the backpropagation training algorithm is essentially stochastic (most common patterns have more influence in the weights re-estimation process), it is reasonable to suppose that the LIN efficacy depends on the input and each noisy frame could be associated to a reliability coefficient that attempts to measure how reliable is the result of the neural net processing. Isolated word recognition experiments have shown that reliability weighting can result in a mean error rate reduction as high as 96, 80, 58 and 36 % at SNR=12, 6, 3 and 0dB, respectively, when the noise is white Gaussian.
Cite as: Yoma, N.B., McInnes, F.R., Jack, M.A. (1996) Use of a reliability coefficient in noise cancelling by neural net and weighted matching algorithms. Proc. 4th International Conference on Spoken Language Processing (ICSLP 1996), 2297-2300, doi: 10.21437/ICSLP.1996-581
@inproceedings{yoma96_icslp,
author={Nestor Becérra Yoma and Fergus R. McInnes and Mervyn A. Jack},
title={{Use of a reliability coefficient in noise cancelling by neural net and weighted matching algorithms}},
year=1996,
booktitle={Proc. 4th International Conference on Spoken Language Processing (ICSLP 1996)},
pages={2297--2300},
doi={10.21437/ICSLP.1996-581}
}