Extension Of Empirical Mode Decomposition Research Articles

MEMD-Based Hybrid Modal Identification for High-Rise Structures with Multi-Sensor Vibration Measurements

Although widely used in various fields due to its powerful capability of signal processing, empirical mode decomposition has to decompose signals separately, which limits its application for multivariate data such as the structural monitoring data recorded by multiple sensors. In order to avoid this shortcoming, a multivariate extension of empirical mode decomposition is proposed to deal with the multidimensional signals synchronously by employing a real-valued projection on hyperspheres. This study presents a hybrid modal identification method combining the multivariate empirical mode decomposition with stochastic subspace identification and fast Bayesian FFT methods to more conveniently and accurately identify structural dynamic parameters from multi-sensor vibration measurements. Deployed as a preprocessing tool, the multivariate signals are decomposed into several aligned intrinsic mode functions, which contain only a dominant component in the frequency domain. Then, the modal parameters can be identified by advanced fast Bayesian FFT and stochastic subspace identification directly. The combined method is first validated by a numerical illustration of a frame structure and then is applied in a shaking table test and a full-scale measurement under nonstationary earthquake excitation. Compared with the finite element method, the peak–pick, the half-power bandwidth methods, and Hilbert–Huang transform method, the results show that this hybrid method is more robust and reliable in the modal parameters identification. The main contribution of this paper is to develop a more effective integrated approach for accurate modal identification with the output-only multi-dimensional nonstationary signal.

A Strategy for Classification of "Vaginal vs. Cesarean Section" Delivery: Bivariate Empirical Mode Decomposition of Cardiotocographic Recordings.

We propose objective and robust measures for the purpose of classification of “vaginal vs. cesarean section” delivery by investigating temporal dynamics and complex interactions between fetal heart rate (FHR) and maternal uterine contraction (UC) recordings from cardiotocographic (CTG) traces. Multivariate extension of empirical mode decomposition (EMD) yields intrinsic scales embedded in UC-FHR recordings while also retaining inter-channel (UC-FHR) coupling at multiple scales. The mode alignment property of EMD results in the matched signal decomposition, in terms of frequency content, which paves the way for the selection of robust and objective time-frequency features for the problem at hand. Specifically, instantaneous amplitude and instantaneous frequency of multivariate intrinsic mode functions are utilized to construct a class of features which capture nonlinear and nonstationary interactions from UC-FHR recordings. The proposed features are fed to a variety of modern machine learning classifiers (decision tree, support vector machine, AdaBoost) to delineate vaginal and cesarean dynamics. We evaluate the performance of different classifiers on a real world dataset by investigating the following classifying measures: sensitivity, specificity, area under the ROC curve (AUC) and mean squared error (MSE). It is observed that under the application of all proposed 40 features AdaBoost classifier provides the best accuracy of 91.8% sensitivity, 95.5% specificity, 98% AUC, and 5% MSE. To conclude, the utilization of all proposed time-frequency features as input to machine learning classifiers can benefit clinical obstetric practitioners through a robust and automatic approach for the classification of fetus dynamics.

Seizure detection from EEG signals using Multivariate Empirical Mode Decomposition

We present a data driven approach to classify ictal (epileptic seizure) and non-ictal EEG signals using the multivariate empirical mode decomposition (MEMD) algorithm. MEMD is a multivariate extension of empirical mode decomposition (EMD), which is an established method to perform the decomposition and time-frequency (T−F) analysis of non-stationary data sets. We select suitable feature sets based on the multiscale T−F representation of the EEG data via MEMD for the classification purposes. The classification is achieved using the artificial neural networks. The efficacy of the proposed method is verified on extensive publicly available EEG datasets.

Robust multitask learning with three-dimensional empirical mode decomposition-based features for hyperspectral classification

Empirical mode decomposition (EMD) and its variants have recently been applied for hyperspectral image (HSI) classification due to their ability to extract useful features from the original HSI. However, it remains a challenging task to effectively exploit the spectral-spatial information by the traditional vector or image-based methods. In this paper, a three-dimensional (3D) extension of EMD (3D-EMD) is proposed to naturally treat the HSI as a cube and decompose the HSI into varying oscillations (i.e. 3D intrinsic mode functions (3D-IMFs)). To achieve fast 3D-EMD implementation, 3D Delaunay triangulation (3D-DT) is utilized to determine the distances of extrema, while separable filters are adopted to generate the envelopes. Taking the extracted 3D-IMFs as features of different tasks, robust multitask learning (RMTL) is further proposed for HSI classification. In RMTL, pairs of low-rank and sparse structures are formulated by trace-norm and l1,2-norm to capture task relatedness and specificity, respectively. Moreover, the optimization problems of RMTL can be efficiently solved by the inexact augmented Lagrangian method (IALM). Compared with several state-of-the-art feature extraction and classification methods, the experimental results conducted on three benchmark data sets demonstrate the superiority of the proposed methods.

A novel quadrature clutter rejection approach based on the multivariate empirical mode decomposition for bidirectional Doppler ultrasound signals

A novel quadrature clutter rejection approach based on multivariate empirical mode decomposition (MEMD), which is an extension of empirical mode decomposition (EMD) to multivariate for processing multichannel signals, is proposed in this paper to suppress the quadrature clutter signals induced by the vascular wall and the surrounding stationary or slowly moving tissues in composite Doppler ultrasound signals, and extract more blood flow components with low velocities. In this approach, the MEMD algorithms, which include the bivariate empirical mode decomposition with a nonuniform sampling scheme for adaptive selection of projection directions (NS_BEMD) and the trivariate empirical mode decomposition with noise assistance (NA_TEMD), are directly employed to adaptively decompose the complex-valued quadrature composite signals echoed from both bidirectional blood flow and moving wall into a small number of zero-mean rotation components, which are defined as complex intrinsic mode functions (CIMFs). Then the relevant CIMFs contributed to blood flow components are automatically distinguished in terms of the break of the CIMFs’ power, and then directly added up to give the quadrature blood flow signal. Specific simulation and human subject experiments are taken up to demonstrate the advantages and limitations of this novel method for quadrature clutter rejection in bidirectional Doppler ultrasound signals. Due to eliminating the extra errors induced by the Hilbert transform or complex FIR filter algorithms used in the traditional clutter rejection approaches based on the directional separation process, the proposed method provides improved accuracy for clutter rejection, and preserve more slow blood blow components, which could be helpful to early diagnose arterial diseases.

Myoelectric pattern identification of stroke survivors using multivariate empirical mode decomposition.

This study presents a novel feature extraction method for myoelectric pattern recognition using a multivariate extension of empirical mode decomposition (EMD), namely multivariate EMD (MEMD). The method processes multiple surface electromyogram (EMG) channels simultaneously rather than in a channel-by-channel manner. From mode-aligned intrinsic mode functions (IMFs, representing signal components over multiple scales) derived from the MEMD analysis, normalized amplitude distributions of the same-mode/scale IMFs across different channels were calculated as features, which serve to reveal the underlying relationship in the aligned intrinsic scales across multiple muscles. The proposed method was assessed for identification of 18 different functional movement patterns via 27-channel surface EMG signals recorded from the paretic forearm muscles of 12 subjects with hemiparetic stroke. With a linear discriminant classifier, the proposed MEMD based feature set resulted in an average error rate of 4.61 ± 4.70% for classification of all the different movements, significantly lower than that of the conventional time-domain feature set (7.14 ± 6.15%, p < 0.05). The results indicate that the MEMD based feature extraction of multi-channel surface EMG data provides a promising approach to modeling of muscle couplings and identification of different myoelectric patterns.

Classification of Motor Imagery BCI Using Multivariate Empirical Mode Decomposition

Brain electrical activity recorded via electroencephalogram (EEG) is the most convenient means for brain-computer interface (BCI), and is notoriously noisy. The information of interest is located in well defined frequency bands, and a number of standard frequency estimation algorithms have been used for feature extraction. To deal with data nonstationarity, low signal-to-noise ratio, and closely spaced frequency bands of interest, we investigate the effectiveness of recently introduced multivariate extensions of empirical mode decomposition (MEMD) in motor imagery BCI. We show that direct multichannel processing via MEMD allows for enhanced localization of the frequency information in EEG, and, in particular, its noise-assisted mode of operation (NA-MEMD) provides a highly localized time-frequency representation. Comparative analysis with other state of the art methods on both synthetic benchmark examples and a well established BCI motor imagery dataset support the analysis.

Identification of Defective Areas in Composite Materials by Bivariate EMD Analysis of Ultrasound

In recent years, many alternative methodologies and techniques have been proposed to perform nondestructive inspection and maintenance operations of moving structures. In particular, ultrasonic techniques have shown to be very promising for automatic inspection systems. From the literature, it is evident that the neural paradigms are considered, by now, the best choice to automatically classify ultrasound data. At the same time, the most appropriate preprocessing technique is still undecided. The aim of this paper is to propose a new and innovative data preprocessing technique that converts real-valued ultrasonic data into complex-valued signals. This allows analysis using phase synchrony, a robust tool that has been previously employed in brain science for establishing robust features in noisy data. Synchrony estimation is achieved using complex extensions of empirical mode decomposition, a data-driven algorithm for detecting temporal scales, thus facilitating the modeling of nonlinear and nonstationary signal dynamics. Experimental tests aiming to detect defective areas in composite materials are reported, and the effectiveness of the proposed methodology is illustrated.

Multiresolution Information Mining for Pavement Crack Image Analysis

Empirical mode decomposition (EMD) is a multiresolution data analysis method recently developed to cater to the inherent nonstationarity in real-world signals. A two-dimensional (2D) extension of EMD is used in this paper as a pavement distress image analytical tool. The algorithm decomposes an image into a set of narrow band components (called bidimensional intrinsic mode function, or BIMF) that uniquely reflect the variations in the image. Although some components could have good image edge characteristics, others might hold fidelity to the shape and size of objects or trends in the image. Therefore, the complete spatial and frequency characteristic of a desired image feature might also be divided into the different components, indicating that all attributes of a desired feature might not be found in a single component. An optimal solution requires image mining from the different component resolutions to accurately extract those specific attributes without compromising certain spatial and frequency characteristics. Two major contributions to pavement image analysis are achieved. First, the paper explores pavement image denoising or enhancement by combining the EMD and a weighted reconstruction technique as a tool for background standardization of images acquired under different types of illumination effects. Second, using principal component pursuit (PCP), the authors reconstruct a composite image by selecting salient information from coarse and fine resolution BIMFs useful for accurate extraction of linear patterns in a pavement distress image. Compared with conventional image reconstruction or approximation techniques, the methodology used in this paper yields better results. DOI:10.1061/(ASCE)CP.1943-5487.0000178. (C) 2012 American Society of Civil Engineers.

Empirical Mode Decomposition for Trivariate Signals

<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> An extension of empirical mode decomposition (EMD) is proposed in order to make it suitable for operation on trivariate signals. Estimation of local mean envelope of the input signal, a critical step in EMD, is performed by taking projections along multiple directions in three-dimensional spaces using the rotation property of quaternions. The proposed algorithm thus extracts rotating components embedded within the signal and performs accurate time-frequency analysis, via the Hilbert–Huang transform. Simulations on synthetic trivariate point processes and real-world three-dimensional signals support the analysis. </para>