Daoud Boutana

Multiresolution wavelet-based QRS complex detection algorithm suited to several abnormal morphologies

The electrocardiogram (ECG) signal is considered as one of the most important tools in clinical practice in order to assess the cardiac status of patients. In this study, an improved QRS (Q wave, R wave, S wave) complex detection algorithm is proposed based on the multiresolution wavelet analysis. In the first step, high frequency noise and baseline wander can be distinguished from ECG data based on their specific frequency contents. Hence, removing corresponding detail coefficients leads to enhance the performance of the detection algorithm. After this, the authors method is based on the power spectrum of decomposition signals for selecting detail coefficient corresponding to the frequency band of the QRS complex. Hence, the authors have proposed a function g as the combination of the selected detail coefficients using two parameters λ 1 and λ 2, which correspond to the proportion of the frequency ranges of the selected detail compared with the frequency range of the QRS complex. The proposed algorithm is evaluated using the whole arrhythmia database. It presents considerable capability in cases of low signal-to-noise ratio, high baseline wander and abnormal morphologies. The results of evaluation show the good detection performance; they have obtained a global sensitivity of 99.87%, a positive predectivity of 99.79% and a percentage error of 0.34%.

On the selection of Intrinsic Mode Function in EMD method: Application on heart sound signal

Empirical mode decomposition (EMD) allows decomposing an observed multicomponent signal into a set of monocomponent signals called Intrinsic Mode Functions (IMFs). EMD provides a large number of IMFs and it is important to select the fundamental IMFs and eliminate the redundant ones. This paper proposes a new criterion, based simultaneously on the Minkowski distance and the Jensen Rényi divergence of order α (α-JRD), to automatically select the appropriate IMFs in a set of the extracted ones. Examples, using synthetic and real-life heart sound signals, are presented in order to validate the performance of the proposed technique.

Denoising and characterization of heart sound signals using optimal intrinsic mode functions

Empirical mode decomposition (EMD) allows decomposing an observed multicomponent signal into a set of monocomponent signals, called Intrinsic Mode Functions (IMFs). The aim of this paper is to characterize some heart sound (HS) signals embedded in noise using the EMD approach. In particular, the proposed technique automatically selects the most appropriate IMFs achieving the denoising based on EMD and Euclidean measure. Synthetic and real-life signals are used in the various examples to validate, and demonstrate the effectiveness, of the proposed method. Furthermore, this technique is compared to the commonly known approach based on the noise model.

Identification of Aortic Stenosis and Mitral Regurgitation By Heart Sound Segmentation On Time-Frequency Domain

Heart sounds are multicomponent non-stationary signals which characterize the normal phonocardiogram signals (PCGs) and more significantly the pathological PCGs. The time-frequency distributions (TFDs) are a useful tool for local analysis of non-stationary and fast transient wideband signals especially for (PCG) signals. This tool provides a noninvasive probe to detect and to characterize the presence of abnormal murmur in the diagnosis of heart disease. In this paper, we introduce a method for the segmentation and the analysis the PCG signal for detecting the murmur based on time frequency analysis in conjunction with a threshold based on Renyi entropy. The method was applied to differents sets of PCGs: Early Aortic Stenosis (EAS), Late systolic Aortic Stenosis (LAS), and finely the Mitral Regurgitation (MR.). The analysis has been conducted on data which have been collected from [1] . Test performed on these real biomedical data proves the ability of the method for segmentation between the main components of the PCG signal and the pathological murmurs. Also, the method permits to elucidate and extract useful features for diagnosis and pathological recognition.

Benefits of prior speech segmentation for best time-frequency visualisation using Renyi's entropy

In this paper, a new approach that operates in the joint time-frequency domain for speech segmentation is presented. Segmentation is an important application in speech and audio processing. The segmentation in time domain is based on Renyi entropy especially on Renyi marginal entropy (RME) properties. Experiments were conducted using real-life speech signal as consonant-vowel (CV) transition that consists of two different events. They demonstrated the ability of the method for segmentation of speech signal made of CV transition. This technique is also useful for best time-frequency visualization with appropriate parameters. Because of the simplicity and effectiveness of proposed segmentation technique, it can be applied in many applications such as speaker identification/verification, estimation of the duration of the plosives, feature extraction, and classification.

Automatic detection method of R-wave positions in electrocardiographic signals

In this study, we have presented a robust method of R-wave detection in electrocardiographic signals (ECG) using multiresolution wavelet transform. Daubechies wavelets (db4 and db6) were used for this detection. The presented method requires a pre-processing by using a Median filter which is an effective tool to remove the baseline of the signal and correct it. Then, the ECG signals under test have been decomposed to the required level using the selected wavelets. Finally, the algorithm has been tested and evaluated on MIT-BIH arrhythmia database. Signals from lead II have been only analyzed. In this validation, the QRS detection algorithm achieved the very good detection performance using db6 wavelet, because db6 wavelet gives the higher values of the overall sensitivity and predictivity than to db4 wavelet.

Wavelet based segmentation and time-frequency caracterisation of some abnormal heart sound signals

Biomedical signal recordings are so complex and nonstationnary, they are also affected by different kinds of noise that make their interpretation difficult. The major goal of the paper consists of two ideas. In the first one, we present the results of segmentation method followed by the time frequency caracterisation of some phonocardiogram (PCG) signals. The paper using the Discrete Wavelet Transform (DWT) in conjunction with Shannon entropy provided a useful tool for phonocardiogram (PCG) segmentation. In the segmentation technique, we calculate the entropy of the details coefficients at each level and threshold it in order to detect the murmur of heart sound signals. Several real-life signals were used: Early Aortic Stenosis (EAS), Late Aortic stenosis (LAS), Mitral Regurgitation (MR), Aortic Regurgitation (AR) .The results of the method illustrate clearly the detection of the main components S1, S2, S3, pathological murmurs and the identification of the valves disease.

EEG Signals Classification Based on Time Frequency Analysis

This paper presents a method to characterize, identify and classify some pathological Electroencephalogram (EEG) signals. We use some Time Frequency Distributions (TFDs) to analyze its nonstationarity. The analysis is conducted by the spectrogram (SP), the Choi–Williams Distribution (CWD) and the Smoothed Pseudo Wigner Ville Distribution (SPWVD). The studies are carried on some real EEG signals collected from a known database. The estimation of the best value of parameters for each distribution is achieved using the Renyi entropy (RE). The time-frequency results have permitted to characterize some pathological EEG signals. In addition, the Renyi Marginal Entropy (RME) is used for the purpose of detecting the peak seizures and discriminates between normal and pathological EEG signals. The frequency bands are evaluated using the Marginal Frequency (MF). The EEG signal classification of two sets A and E containing normal and pathologic EEG signals, respectively, is performed using our proposed method based on energy extraction of signals from time-frequency plane. Also, the Moving Average (MA) is used as a tool to obtain better classification results. The results conducted on real-life EEG signals illustrate the effectiveness of the proposed method.

Comparative Study of Time Frequency Analysis Application on Abnormal EEG Signals

This paper presents a time-frequency analysis for some pathological Electroencephalogram (EEG) signals. The proposed method is to characterize some pathological EEG signals using some time-frequency distributions (TFD). TFDs are useful tools for analyzing the non-stationary signals such as EEG signals. We have used spectrogram (SP), Choi-Williams Distribution (CWD) and Smoothed Pseudo Wigner Ville Distribution (SPWVD) in conjunction with Renyi entropy (RE) to calculate the best value of their parameters. The study is conducted on some case of epileptic seizure of EEG signals collected on a known database. The best values of the analysis parameters are extracted by the evaluation of the minimization of the RE values. The results have permit to visualize in time domain some pathological EEG signals. Also, the Renyi marginal entropy (RME) has been used in order to identify the peak seizure. The characterization is achieved by evaluating the frequency bands using the marginal frequency (MF).

Spectral analysis of Arabic speech signals: Cases of children with normal and impaired hearing

L’objectif de ce travail est de mener une étude spectrale comparative du signal de parole arabe des enfants normo-entendants et des enfants malentendants portant sur la discrimination phonémique et plus particulièrement sur les caractères acoustiques de voisement au niveau perceptif ainsi que l’extraction de caractères spectraux en relation avec les différents types de déficiences auditives. On étudie, à l’aide de l’analyse spectrale par prédiction linéaire (LPC), les positionnements des formants des signaux de parole, plus particulièrement des voyelles des enfants normo-entendants et des enfants malentendants. On note l’utilisation du troisième formant lié à la qualité sonore. Par ailleurs, on étudie l’utilisation de la distance spectrale d’Itakura-Saito comme paramètre de discrimination entre les voyelles des enfants normo-entendants et des enfants malentendants.