Binwei Weng

ECG Denoising Based on the Empirical Mode Decomposition

The electrocardiogram (ECG) has been widely used for diagnosis purposes of heart diseases. Good quality ECG are utilized by the physicians for interpretation and identification of physiological and pathological phenomena. However, in real situations, ECG recordings are often corrupted by artifacts. One prominent artifact is the high frequency noise caused by electromyogram induced noise, power line interferences, or mechanical forces acting on the electrodes. Noise severely limits the utility of the recorded ECG and thus need to be removed for better clinical evaluation. Several methods have been developed for ECG denoising. In this paper, we proposed a new ECG denoising method based on the recently developed Empirical Mode Decomposition (EMD). The proposed EMD-based method is able to remove high frequency noise with minimum signal distortion. The method is validated through experiments on the MIT-BIH database. Both quantitative and qualitative results are given. The results show that the proposed method provides very good results for denoising

Nonlinear system identification in impulsive environments

Nonlinear system identification has been studied under the assumption that the noise has finite second and higher order statistics. In many practical applications, impulsive measurement noise severely weakens the effectiveness of conventional methods. In this paper, /spl alpha/-stable noise is used as a noise model. In such case, the minimum mean square error (MMSE) criterion is no longer an appropriate metric for estimation error due to the lack of finite second-order statistics of the noise. Therefore, we adopt minimum dispersion criterion, which in turn leads to the adaptive least mean pth power (LMP) algorithm. It is shown that the LMP algorithm under the /spl alpha/-stable noise model converges as long as the step size satisfies certain conditions. The effect of p on the performance is also investigated. Compared with conventional methods, the proposed method is more robust to impulsive noise and has better performance.

Baseline Wander Correction in ECG by the Empirical Mode Decomposition

The electrocardiogram (ECG) has been widely used for diagnosis purposes of heart diseases. A good quality ECG may help the physicians to easily interpret any physiological or pathological phenomena. However, in real situations, ECG recordings are often affected by several factors that result in the baseline wander. Baseline wander is a low frequency artifact that may be due to respiration or the motion of the patients or the electrodes. A large baseline wander severely limits the utility of the recorded ECG and thus need to be corrected to enable better clinical evaluation. In this paper, we propose a new baseline wander correction method based on the recently developed tool-Empirical Mode Decomposition (EMD). We validate our method by experiments from the MIT-BIH databases and also compare our method with the highpass filtering method. Both qualitative and quantitative results show that the proposed EMD-based method provides very good results.

Optimal Signal Reconstruction Using the Empirical Mode Decomposition

The empirical mode decomposition (EMD) was recently proposed as a new time-frequency analysis tool for nonstationary and nonlinear signals. Although the EMD is able to find the intrinsic modes of a signal and is completely self-adaptive, it does not have any implication on reconstruction optimality. In some situations, when a specified optimality is desired for signal reconstruction, a more flexible scheme is required. We propose a modified method for signal reconstruction based on the EMD that enhances the capability of the EMD to meet a specified optimality criterion. The proposed reconstruction algorithm gives the best estimate of a given signal in the minimum mean square error sense. Two different formulations are proposed. The first formulation utilizes a linear weighting for the intrinsic mode functions (IMF). The second algorithm adopts a bidirectional weighting, namely, it not only uses weighting for IMF modes, but also exploits the correlations between samples in a specific window and carries out filtering of these samples. These two new EMD reconstruction methods enhance the capability of the traditional EMD reconstruction and are well suited for optimal signal recovery. Examples are given to show the applications of the proposed optimal EMD algorithms to simulated and real signals.

Optimal and Bidirectional optimal Empirical Mode Decomposition

The empirical mode decomposition (EMD) was recently proposed as a new time-frequency analysis tool for nonstationary and nonlinear signals. Although the EMD is able to find the intrinsic modes of the signal and is completely self-adaptive, it does not have any implication on optimality. In some situation, when certain optimality is considered, we need a more flexible signal decomposition and reconstruction scheme. We propose a modified version of the EMD, which enhances the capability of the EMD. The proposed modified EMD algorithm gives the best estimate to a given signal in the minimum mean square error sense. Two different formulations are proposed. The first one utilizes a linear weighting for the intrinsic mode functions (IMF). The second algorithm adopts a bidirectional weighting, namely, it not only uses weighting for IMF modes, but also exploits the correlations between samples in a specific window and carries out filtering in the window. These two new EMD methods extend the capability of the traditional EMD and is well suited for optimal signal recovery. Simulation studies are performed to show the application of the proposed optimal EMD algorithms to denoising problem.

Atrial fibrillation detection using stationary wavelet transform analysis

Atrial fibrillation (AF) is a common cardiac arrythmia that is usually developed for elder people with aging. AF may result in complications such as chest pain or even heart failure in later stage. Based on the characteristics of surface ECG, AF can be detected by several methods. A particular investigation on the fibrillatory waveform reveals the inherent structure of AF signals. As opposed to traditional frequency domain methods, we utilize the stationary wavelet transform to extract the information from ECG signal which differentiates AF and non-AF cases based on some feature extraction and selection processes. A linear classifier is then designed for computational efficiency. The proposed method eliminates the need for QRST cancellation step which is required for frequency domain methods and provides a more systematic approach for AF detection. Extensive experiments are tested on signals from the MIT-BIH Atrial Fibrillation Database to show the superior performance of the proposed algorithm.

Empirical mode decomposition as a tool for DNA sequence analysis from terahertz spectroscopy measurements

DNA sequence analysis has been widely studied by gene-expression microarray techniques. Few results, however, have been provided by Terahertz spectroscopy which reveals the absorbtion or reflectance percentage from different DNA sequences. Previous Terahertz methods have lacked a quantitative analysis of the spectroscopy features, and no definitive conclusion regarding the data can be easily drawn. In this paper, we use a signal processing approach which gives a quantitative interpretation of the DNA spectroscopy. Due to the presence of physical noise, the data can be contaminated by both random fluctuations and impulsive noise. A new signal processing tool called empirical mode decomposition (EMD) is employed to remove the noise and extract the trend of the signal. The data is subsequently partitioned by clustering methods. Experimental results of Terahertz spectroscopy of several different DNA samples show that the EMD aids the clustering process and yields clustering of higher validity than that obtained from the raw data.

Time-Varying Volterra System Identification Using Kalman Filtering

Most nonlinear system identification methods based on Volterra model assume that the underlying system is time-invariant. In this paper, a novel identification method for time-varying Volterra systems (TVVS) is proposed. We view this problem from a different perspective in the sense that the system identification problem is converted to a state estimation problem of a dynamic system. The time-varying Volterra kernels are governed by a Gauss-Markov stochastic difference equation upon which a state-space representation of time-varying Volterra systems is built. The state transition matrix and noise covariance of the underlying state equations are usually unknown. Therefore, we develop a method to estimate these unknown quantities. Finally, a Kalman filtering scheme is utilized to identify and track the time-varying Volterra system. Simulation examples are given to illustrate the better performance of the proposed method as compared with other adaptive identification methods such as the LMS and RLS algorithms.

TR-MUSIC: a robust frequency estimation method in impulsive noise

Sinusoidal frequency estimation has been studied for many years. The MUSIC method represents a class of super-resolution methods based on subspace decomposition. However, the MUSIC method has poor performance in impulsive noise environments due to the prevalence of outliers and very large noise variance. A more robust method called trimmed correlation based-MUSIC (TR-MUSIC) method is proposed in this paper. Through a trimming operation, outliers in the samples participating in the correlation calculation are discarded, yielding a correlation sequence that is closer to the true underlying correlation. The amount of trimming is determined by the Mahalanobis distance in which robust estimates of location and scale are utilized to compensate for outlier effects. Frequency estimation results from the eigendecomposition of the trimmed correlation matrix. In the simulations, we take α-stable noise (α > 1) as an example of impulsive noise. The proposed method is very robust and performs better than the conventional MUSIC and other robust methods. Furthermore, it can be applied to real signals as well as complex signals.

Discriminating DNA Sequences from Terahertz Spectroscopy - A Wavelet Domain Analysis

DNA sequence analysis has been widely studied by geneexpression microarray techniques. Few results, however, have been provided by Terahertz spectroscopy which reveals the absorbtion or reflectance percentage from different DNA sequences. Previous Terahertz methods have lacked a quantitative analysis of the spectroscopy features, and no definitive conclusion regarding the data can be easily drawn. In this paper, we use a data clustering approach which gives a quantitative interpretation of the DNA spectroscopy. The wavelet transform is applied to extract features from the original measurements. The data is subsequently partitioned by a hierarchical clustering method. Experimental results of Terahertz spectroscopy of several different DNA samples show that the wavelet domain analysis aids the clustering process and yields clustering of higher validity than that obtained from the raw data.