Blind Source Extraction Algorithm Research Articles

Auxiliary function-based algorithm for blind extraction of a moving speaker

In this paper, we propose a novel algorithm for blind source extraction (BSE) of a moving acoustic source recorded by multiple microphones. The algorithm is based on independent vector extraction (IVE) where the contrast function is optimized using the auxiliary function-based technique and where the recently proposed constant separating vector (CSV) mixing model is assumed. CSV allows for movements of the extracted source within the analyzed batch of recordings. We provide a practical explanation of how the CSV model works when extracting a moving acoustic source. Then, the proposed algorithm is experimentally verified on the task of blind extraction of a moving speaker. The algorithm is compared with state-of-the-art blind methods and with an adaptive BSE algorithm which processes data in a sequential manner. The results confirm that the proposed algorithm can extract the moving speaker better than the BSE methods based on the conventional mixing model and that it achieves improved extraction accuracy than the adaptive method.

Using Joint Generalized Eigenvectors of a Set of Covariance Matrix Pencils for Deflationary Blind Source Extraction

In this paper, we develop a new deflationary blind source extraction (BSE) algorithm that extracts source signals in a sequential fashion via the joint generalized eigenvectors of a set of covariance matrix pencils. The new concept of joint generalized eigenvector is defined. We prove that these vectors can be made unique and identical to the source extraction vectors with properly selected matrix pencils. To resolve the open problem of estimating joint generalized eigenvectors, we develop an approach based on the deflation operation and the proportional property of the joint generalized eigenvectors. Specifically, with the proportional property, we show that the estimation problem can be formulated as an optimization involving a quadratic cost function and a unit-rank matrix constraint. An efficient iterative algorithm is then developed by applying the gradient search, matrix shrinkage, deflation, and symmetry-preserving vectorization techniques. This algorithm estimates the joint generalized eigenvectors and conducts BSE sequentially. Its computational complexity and convergence are analyzed. Simulations demonstrate that this algorithm outperforms many typical BSE or blind source separation algorithms. In particular, the new algorithm is more robust to both heavy noise and ill-conditioned mixing matrices.

Stationary Points of a Kurtosis Maximization Criterion for Noisy Blind Source Extraction

Blind source extraction (BSE) is often posed as the maximization of a statistical criterion under a unitary constraint. This paper addresses the convergence problem of a kurtosis-based criterion in the presence of noise. We present the stationary points of such criterion, and show that these extrema are simplified as the minimum mean square error (MMSE) solutions with some approximations. Moreover, we introduce a robust preprocessing approach, which allows one to find the MMSE separation matrix up to an orthogonal factor. The excellent performance of the BSE algorithm based on this preprocessing approach shows that the analysis of the stationary point is reliable.

A Framework of Target Detection in Hyperspectral Imagery Based on Blind Source Extraction

A framework based on the blind source extraction (BSE) algorithm is proposed to detect targets in remotely sensed hyperspectral images. The mean square cross prediction error (MSCPE)-based BSE method is used as the kernel algorithm where the autoregressive (AR) parameters of the targets' spectra are utilized as priors. Numerical simulations show that the proposed framework highlights the desired signal, suppresses the backgrounds, and is able to detect the distribution of the target. In the experiments, the data from the Rochester Institute of Technology (RIT) were used to evaluate the framework. The proposed method achieved a better performance in the tradeoff between the PD and the PFA with subpixel target detection compared with the constrained energy minimization (CEM), the adaptive cosine estimator (ACE), the matched filter (MF), the generalized likelihood ratio test (GLRT), the adaptive matched subspace detector (AMSD), and the orthogonal subspace projection (OSP).

Blind source extraction based on time-frequency characteristics for underwater object acoustic scattering

The physical mechanism and signal characteristics of acoustic scattering are the vital basis for target recognition. But underwater target acoustic scattering components are aliasing in time-frequency (TF) domain, for which the target elastic acoustic scattering characteristics are difficult to detect. Additionally, the existing blind source separation methods are effective only on condition that the number of array elements is equal to or greater than the number of the source signals. To address these problems, a novel TF domain blind source extraction method of separating target acoustic scattering components is proposed in this paper. The method only uses the TF energy characteristic differences among the target acoustic scattering components, and special limitations on target echo structures are unnecessary. Image morphology filter is used to remove the cross-term interference from time-frequency distribution (TFD) of the received array signals. Then, the single source which shows maximum energy concentration at the corresponding auto-term TF points is extracted through three operations: i) selecting the single source auto-term TF points from the auto-term ones; ii) constructing the spatial TFD matrix according to the selected single source auto-term TF points; iii) obtaining the single source by decomposing the eigenvalue of their spatial TFD matrix. Finally, the extracted single signal is excluded by the tightening process from the received array signals, and each single signal is separated successively by repeating the above steps. In addition, a signal processing model which can describe the physical characteristics of the target echoes is established based on the separated signal components. Simulations illustrate that the image morphological filter can remove the cross-term interference and improve the TF resolution of the Wigner-Ville distribution. Anechoic pool experimental results show that the TF domain blind source extraction algorithm can well separate each target acoustic scattering component, it can also achieve a higher output signal-to-noise ratio. Furthermore, the separated elastic acoustic scattering components are in good agreement with the results computed by the surface wave generating theory, so the method can provide the robust and reliable feature for underwater target recognition.

Blind Source Separation Based Analysis of Correlated Harmonic Sources in a Micro Grid Environment

Harmonics are a persistent problem in power systems starting from the simple to any complex natured power system network. Blind source separation using Independent Component Analysis (ICA) is applied to estimate the harmonic sources with least prior knowledge on the topology of power systems. The limitation of the method is that it assumes that the harmonic sources are statistically independent. However, in simple microgrid system powering a small region, it becomes very likely that the harmonic sources like two housing colonies,etc to have some degree of dependence which makes them correlated to a minimal extent. Under such circumstances, it is observed that the traditional ICA algorithms like FastICA breakdown. In the work, two algorithms suitable for statistical dependence namely the eigBSE(Eigen value based Blind source Extraction) algorithm and the maxNG (maximum non Gaussianity based) algorithm discussed in literature are applied to a simple four bus microgrid model. The separation quality and performance indices of the two algorithms are explored for various correlation coefficient values and it is found that in majority of the cases maxNG algorithm is more precise. More details of the theoretical issues of statistical dependence and the algorithm behaviour needs to be workedout.

Research on Blind Signal Extraction Based on Normalized Kurtosis

Blind source extraction (BSE) is a promising technique to solve signal mixture problems while only one or a few source signals are desired. In biomedical applications, one often knows certain prior information about a desired source signal in advance. In this paper, we explore specific prior information as a constrained condition so as to develop a flexible BSE algorithm. One can extract a desired source signal while its normalized kurtosis range is known in advance. Computer simulations on biomedical signals confirm the validity of the proposed algorithm.

FECG Extraction Algorithm Based on BSS Using Temporal Structure and DWT

Fetal electrocardiogram (FECG) blind source extraction (BSE) algorithm based on temporal structure and discrete wavelet transformation (DWT) in noise is proposed in this paper. After building the basic blind source separation (BSS) and BSE models for FECG, some preprocessing procedures based on the temporal structure of the FECG are constructed. Using DWT we can move the conventional time-domain signals to the wavelet-domain, and then the source number is detected and the robust noise reduction technique in FECG can be deduced too. According this preprocessing and second-order statistics (SOS), the proposed robust FECG extraction algorithm is derived.

An Improved Blind Source Extraction Algorithm

One-unit independent component analysis with reference (ICA-R) is an efficient method capable of extracting a desired source signal by using reference signal. In this paper, a new fast one-unit ICA-R algorithm is derived by using kurtosis contrast function based on new constrained independent component analysis (cICA) theory. The proposed algorithm has lower computational complexity and accurate extraction. Experiments with synthetic signals demonstrate the efficacy and accuracy of the proposed algorithm.

A Method for Extracting Signals with Specific Normalized Kurtosis Range

The famous FastICA algorithm has been widely used for blind signal separation. For every process, it only converges to an original source which has the maximum negentropy of the underlying signals. To ensure the first output is the desired signal, we incorporate a priori knowledge as a constraint into the FastICA algorithm to construct a robust blind source extraction algorithm. One can extract the desired signal if its normalized kurtosis is known to lie in a specific range, whereas other unwanted signals do not belong to this range. Experimental results on biomedical signals illustrate the validity and reliability of the proposed method.

A single stage approach to blind source extraction based on second order statistics

A single stage blind source extraction (BSE) approach to extract one desired source signal from an instantaneous mixture of several signals is studied in this paper. In contrast to methods based on blind source separation (BSS) followed by a classifier we extract directly the desired source signal. In recent papers it has been shown that prior information about the desired source's mixing column or autocorrelation function can be incorporated in a single stage BSE algorithm. The focus of this paper is to give insight in the BSE problem for scenarios where joint diagonalization of correlation matrices is applicable, i.e., to exploit the full potential of the SOS of observed signals.In this work we unify and expand these previously presented BSE methods, which lead to the unified instantaneous blind source extraction (UIBSE) algorithm. We show that UIBSE is even more flexible than previously presented since both types of prior information may be combined to obtain better results in selecting the desired source. Above that, the same algorithm may be used for different objective functions. Finally, a performance comparison with BSS algorithms that are followed by a classifier shows that UIBSE is more robust to noisy observations.

Fetal electrocardiogram extraction algorithm in noise: using BSE

Compared with the ultrasonic measurement of fetal heart motion, the Fetal Electrocardiogram (FECG) provides clinically significant information. In this paper, novel FECG Blind Source Extraction (BSE) algorithm based on Blind Source Separation (BSS) in noise is presented. We first build the basic BSE model for FECG extraction and then based on the time structure of the FECG and maternal ECG, the FECG extraction algorithm is constructed using Second-Order Statistics (SOS). To reduce the noise influence, the algorithm using Discrete Wavelet Transformation (DWT) moves the conventional time-domain signals to the wavelet-domain and then the approximate source number is detected and the noise in FECG can be reduced. According this noise reduction procedure and SOS, the proposed robust FECG extraction algorithm is derived. The experimental outcomes demonstrate the performance of the proposed method and , in turn, verify the validity of the model in biomedical applications. The advantages and limitations of the proposed algorithm and its applications are analysed in the discussions and conclusions.

A robust extraction algorithm for biomedical signals from noisy mixtures

Blind source extraction (BSE) is widely used to solve signal mixture problems where there are only a few desired signals. To improve signal extraction performance and expand its application, we develop an adaptive BSE algorithm with an additive noise model. We first present an improved normalized kurtosis as an objective function, which caters for the effect of noise. By combining the objective function and Lagrange multiplier method, we further propose a robust algorithm that can extract the desired signal as the first output signal. Simulations on both synthetic and real biomedical signals demonstrate that such combination improves the extraction performance and has better robustness to the estimation error of normalized kurtosis value in the presence of noise.

An Intelligent Atrial Fibrillation Wireless Monitoring System Based on Nonlinear Bind Source Extraction Algorithm

The proposed system is designed to monitor patients with atrial fibrillation (AF) in family. This system mainly consists of wireless sensor networks (WSNs), which contains several mobile sensor nodes and coordinator for acquisition of bio-signals, and an embedded computer (EC) for signal processing. The WSNs are responsible to acquire and transmit Electrocardiogram (ECG). The EC is to extract the AF signal using nonlinear blind source extraction (BSE) algorithm. The extracted AF signal is then utilized to intelligently judge whether or not AF is on, based on which the system will send alert information to related doctors via Ethernet. In the meantime, the extracted AF signal is displayed on liquid crystal display (LCD), and then is also sent to relate doctors. The system aims to be low-cost, low-power consumption, small size and long-distance (up to thirty meters) transmission, can be further integrated into other healthcare monitoring system, and is expected to have great potential in family monitoring.

Blind Signal Processing models and methods for Foetal Electrocardiogram signals extraction

Blind Signal Processing (BSP) has found great applications for foetal electrocardiogram (FECG) extraction. The purpose of this paper is to brief review the Blind Source Separation (BSS) and Blind Source Extraction (BSE) models and algorithms for FECG signal processing. After the perspective of linear BSS and BSE models for FECG extraction are presented, the relationship between BSS and BSE is given; also the advantages of BSE are shown. Then, the state-of-the-art BSE algorithms, such as BCBSE algorithm, semiBSE algorithm, GABSE algorithm, reference semiBSE algorithm and robustBSE algorithm, are constructed and overviewed. The advantages and limitations of these algorithms are analysed in the conclusion.

Kurtosis-Based Blind Source Extraction of Complex Non-Circular Signals with Application in EEG Artifact Removal in Real-Time

A new class of complex domain blind source extraction algorithms suitable for the extraction of both circular and non-circular complex signals is proposed. This is achieved through sequential extraction based on the degree of kurtosis and in the presence of non-circular measurement noise. The existence and uniqueness analysis of the solution is followed by a study of fast converging variants of the algorithm. The performance is first assessed through simulations on well understood benchmark signals, followed by a case study on real-time artifact removal from EEG signals, verified using both qualitative and quantitative metrics. The results illustrate the power of the proposed approach in real-time blind extraction of general complex-valued sources.

Noise Estimation Using Mean Square Cross Prediction Error for Speech Enhancement

This paper shows the feasibility of noise extraction from noisy speech and presents a two-stage approach for speech enhancement. The preproposed mean square cross prediction error (MSCPE) based blind source extraction algorithm is utilized to extract the additive noise from the noisy speech signal in the first stage. After that a modified spectral subtraction and a modified Wiener filter approach are proposed to extract the speech signal for speech enhancement in the second stage, where all the frequency spectra of the extracted noise are utilized. Theoretical justification shows that the MSCPE-based algorithm can extract desired signal from mixed sources. Experimental results show that the averaged correlation coefficient between the extracted noise and the original additive noise are beyond 85% for Gaussian noise and beyond 75% for real-world noise at SNR = 0 dB, and the proposed speech enhancement approaches perform better than conventional methods, such as spectral subtraction and Wiener filter.

Complex Blind Source Extraction From Noisy Mixtures Using Second-Order Statistics

A class of second-order complex domain blind source extraction algorithms is introduced to cater for signals with noncircular probability distributions, which is a typical case in real-world scenarios. This is achieved by employing the so-called augmented complex statistics and based on the temporal structures of the sources, thus permitting widely linear (WL) predictability to be the extraction criterion. For rigor, the analysis of the existence and uniqueness of the solution is provided based on both the covariance and the pseudocovariance and for both noise-free and noisy cases, and serves as a platform for the derivation of the algorithms. Both direct solutions and those requiring prewhitening are provided based on a WL predictor, thus making the methodology suitable for the generality of complex signals (both circular and noncircular). Simulations on synthetic noncircular sources support the uniqueness and convergence study, followed by a real-world example of electrooculogram artifact removal from electroencephalogram recordings in real time.

A reference-based blind source extraction algorithm

By utilizing a priori information available as reference, constrained independent component analysis (cICA) or independent component analysis with reference (ICA-R) achieves some advantages over other methods. However, ICA-R is very time-consuming; moreover, it is very difficult to determine its threshold parameter, once the value is improperly chosen the algorithm will fail to converge. In order to overcome these drawbacks, a very simple blind source extraction method, whose optimization function is simply the closeness measure between the desired output and its corresponding reference in ICA-R, is proposed in this paper. Experiments with synthesized data and real-world electrocardiograph data confirm its validity and superiority.

Blind source extraction based on generalized autocorrelations and complexity pursuit

Blind source extraction (BSE) is a special class of blind source separation (BSS) method. Due to its low computation load and fast processing speed, BSE has become one of the promising methods in signal processing and analysis. This paper addresses BSE problem when a desired source signal has temporal structures. Based on the generalized autocorrelations of the desired signal and the non-Gaussianity of its innovations, we develop an objective function. Maximizing this objective function, we present a BSE algorithm and further give its stability analysis in this paper. Simulations on image data and electrocardiogram (ECG) data indicate its better performance and the better property of tolerance to the estimation error of the time delay.