Mohsine Karrakchou

Wavelet packets-based high-resolution spectral estimation

Abstract In this paper, a new scheme of sinusoidal components frequency estimation is proposed. One of the innovations presented is that the estimation is performed on a subband decomposition of the signal. To minimize some of the drawbacks of a rigid filter bank, the decomposition is made adaptive using wavelet packets optimizing a new criterion. Thereafter, a high-resolution estimation technique is applied in the subbands. The new criterion consists in counting the number of modes contained in the subbands using the minimum description length (MDL) criterion, derived from the Akaike information criterion. The optimal subband decomposition is found by maximizing the number of modes over the decomposition tree. This causes the decomposition to stop when a mode will be aliased. The adaptive subband decomposition guarantees the benefits of estimation from a subband decomposition without the inconveniences of the aliasing effects. Simulations performed on synthetic signals confirmed the gain in performance of the proposed method.

Improving cardiovascular monitoring through modern techniques

This article presents two applications of modern processing techniques for cardiovascular monitoring. These two applications are quite different in terms of theory as well as application, but they nevertheless present similarities. In both cases, the incentive for using new techniques came from the need for higher performance in order to achieve desired goals. The results obtained are not limited to the particular cases presented here and may lead to original developments in pure signal processing. The first application deals with automatic in-vivo estimation of microvascular pulmonary pressure (P/sub mv/). New structures of multirate adaptive filtering based on wavelet packets have been developed to remove the respiratory artifact present. The other application investigates the possibility of using chaos and fractal theory to analyze heart rate variability. The use of chaos might help predict cardiac events, such as ventricular tachycardia or fibrillation, with better accuracy than conventional methods.

Subband adaptive filtering: The mutual wavelet packets approach

The authors present a scheme for subband adaptive filtering, where the subband decomposition is performed using a novel mutual wavelet packets decomposition scheme. They contrast the performance and algorithmic complexity of this new structure to that of regular subband adaptive filtering and point out its usefulness in achieving better filtering performance. The use of this mutual wavelet packets decomposition allows the partitioning of the frequency axis into nonuniform subbands adapted to the spectral content of the signals to be filtered. A criterion maximizing the cross-correlation between the signals is proposed for this purpose. Simulations on both synthetic and real signals confirmed the improvement achieved in performance.<<ETX>>

Analysis of heart rate variability: Comparison between spectra obtained from ECG and finger blood pressure

In this paper two techniques for the measurement of Heart Rate Variability are briefly described. These techniques are applied to the ECG signal and the finger blood pressure signal. Comparison of the spectra obtained from the same patient shows that there is a great similarity for low frequencies. However the spectra differ for high frequencies.

Analyzing pulmonary capillary pressure

A new mutual wavelet packets scheme is proposed for subband adaptive filtering. Its relation to the existing literature on the subject is outlined. The proposed scheme performs a decomposition of the two signals used in the adaptation process onto the same optimal basis. That basis is chosen so that it favors the adaptation performance. A new criterion, based on a functional of the two signals, is introduced. This proposed criterion consists in the maximization of the magnitude of the cross-correlation samples of the two signals considered. This maximizes the cross-correlation vector of the well known normal equations. At the expense of a slightly higher computational complexity, the mutual wavelet packets scheme reduces the aliasing drawback of regular subband adaptive filtering, preserving its main advantages. Simulations have been performed on pulmonary capillary pressure transients for respiratory interference cancelling. The method proved to be more efficient than classical methods. >

Multiscale analysis for singularity detection in pulmonary microvascular pressure transients

This paper addresses the problem of pulmonary microvascular pressure estimation. The “why” and “how” of multiscale approaches in the context of singularity detection are discussed. From a linear viewpoint, the scalogram technique and the multiresolution representation for singularity detection are discussed in general under the wavelet framework. A technique as well as a criterion for the segmentation and the extraction of the region of interest or the optimal scale is proposed. A new nonlinear multiscale technique for singularity detection is also proposed. This multiscale representation is based on rank order filters and on mathematical morphology operators in particular. A scaling parameter is introduced for such operators as in the linear continuous wavelet transform leading to what we call morpholograms. Experiments performed on pulmonary artery pressure transients illustrate how the proposed technique can give a synopsis of the signal characteristics through the scales.

Parameter estimation of decaying exponentials by projection on the parameter space

Abstract We propose in this paper a new estimation method for the parameters of a linear combination of exponentials flcorrupted by some noise. It is shown that this method does not require any a priori information about the nature of the noise probability density. The algorithm is used to estimate parameters of synthetic signals and to estimate effective pulmonary ncapillary pressure. This algorithm is compared to the method of Kumaresan and Tufts which is based on linear prediction and singular value decomposition. Our experiments demonstrate increased accuracy over the method of Kumaresan and Tufts.

Optimal ramp edge detection by orthogonal wavelet transform

A method for multiresolution edge detection is described which is based on wavelet representation in which the edges are detected from the local maxima of the detail images in the orthogonal wavelet transform. The new idea is to compute an optimal edge detection filter for ramp edges. It is shown that the approximation of an edge at each scale in its orthogonal wavelet decomposition can be modeled by a ramp contour whose equation is given. Good edge detection results are obtained. Such a multiscale edge detector is important not only for the usual applications such as pattern recognition and image segmentation, but also for object-based image coding where the reconstruction of the original image from the edges at different scales is made possible by wavelet analysis-synthesis.<<ETX>>

Nonlinear Interference Cancelling In Biomedical Systems

While the linear FIR adaptive filtering is very useful and simple for interference cancelling, there are many practical situations where the use of a nonlinear model is more realistic. In this paper a method for nonlinear system identification is briefly described. A new interference canceller scheme based on this method is also discussed. This scheme is applied to two different biomedical signals: Respiratory artifacts cancelling in pulmonary pressure transients and ECG interference cancelling in diaphragmatic muscle signals. Good results are obtained making this scheme attractive also for other applications.

Digital measurement of blood vessels diameters on chest radiographs

To detect congestive left heart failure radiologists must evaluate the diameter of blood vessels on chest radiographs. An automatic digital method to detect the vessels and to measure their diameter is presented in this paper. The vessels are detected by searching the areas which are conform to an ideal vessel model. Then the false vessels detected are eliminated.