Abstract: This method comprises steps of: a) partitioning (10, 16) the spectrum of the noisy signal into a HF part and a LF part; b) operating denoising processes in a differentiated manner for each of the two parts of the spectrum with, for the HF part, a denoising by prediction of the useful signal from one sensor to the other between sensors of a first sub-array (R1), by means of a first adaptive algorithm estimator (14), and, for the LF part, a denoising by prediction of the noise from one sensor to the other between sensors of a second sub-array (R2), by means of a second adaptive algorithm estimator (18); c) reconstructing the spectrum by combining together (22) the signals delivered after denoising of the two parts of the spectrum, respectively; and d) selectively reducing the noise (24) by an Optimized Modified Log-Spectral Amplitude gain, OM-LSA, process.

Abstract: This method comprises steps of: a) partitioning (10, 16) the spectrum of the noisy signal into a HF part and a LF part; b) operating denoising processes in a differentiated manner for each of the two parts of the spectrum with, for the HF part, a denoising by prediction of the useful signal from one sensor to the other between sensors of a first sub-array (R1), by means of a first adaptive algorithm estimator (14), and, for the LF part, a denoising by prediction of the noise from one sensor to the other between sensors of a second sub-array (R2), by means of a second adaptive algorithm estimator (18); c) reconstructing the spectrum by combining together (22) the signals delivered after denoising of the two parts of the spectrum, respectively; and d) selectively reducing the noise (24) by an Optimized Modified Log-Spectral Amplitude gain, OM-LSA, process.

Abstract: The forward signal is submitted to a finite-impulse-response linear-prediction filter in order to determine a first residual signal of minimal energy. The return signal is submitted to a finite-impulse-response filter the coefficients of which are the same as those of the linear-prediction filter of the forward signal in order to determine a second residual signal. The ratio of the energies contained in the first and second residual signals is calculated. It is determined whether the return signal contains components other than the echo components from the forward signal by comparing the calculated ratio to a detection threshold.

Abstract: An electroencephalographic signal processing method allows the stages of sleep to be determined in real time as an EEG signal is being picked up. The EEG signal is divided into time segments of the same length; each signal segment is sampled digitally to obtain successive blocks of N samples; for each block i, M standardized correlation coefficients R.sub.i,k are determined, M being much far smaller than N; for each block i, the distances D(i,j) between the coefficients R.sub.i,k of the current block and the coefficients R.sub.j,k of the centers of gravity of existing classes j are calculated; the minimum D(i,p) of distances D(i,j) is selected; block i is assigned to class p and the center of gravity of the class is updated if D(i,p) is below a given threshold, while a new class is created if D(i,p) is above the given threshold; and each class is assigned to the closest stage from the standpoint of energy distribution beween frequency bands.