Robust speech pulse detection using adaptive noise modelling and non-stationarity measure
Néstor Becerra Yoma, Fergus R. McInnes, M.A. Jack · 2002
The problem of speech pulse detection with additive noise at SNR as low as 0 and -6 dB is addressed. The noise is assumed to be reasonably stationary and correlated. Three techniques have been examined: firstly, the autoregressive analysis of noise; secondly, spectral density comparison; and finally nonstationarity measure.