Second-order statistics based blind source separation using a bank of subband filters

Gharieb, R.R.; Cichocki, Andrzej

doi:10.1016/s1051-2004(02)00034-9

Cited by 30 publications

(16 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These algorithms are exact and fast but sensitive to noise. There are several proposals to improve efficiency and robustness of these algorithms when noise is present [2,8]. They mostly rely on an approximative joint diagonalization of a set of correlation or cumulant matrices like the algorithm second order blind identification (SOBI) [1].…”

Section: Article In Pressmentioning

confidence: 99%

Denoising using local projective subspace methods

et al. 2006

View full text Add to dashboard Cite

In this paper we present denoising algorithms for enhancing noisy signals based on Local ICA (LICA), Delayed AMUSE (dAMUSE) and Kernel PCA (KPCA). The algorithm LICA relies on applying ICA locally to clusters of signals embedded in a high-dimensional feature space of delayed coordinates. The components resembling the signals can be detected by various criteria like estimators of kurtosis or the variance of autocorrelations depending on the statistical nature of the signal. The algorithm proposed can be applied favorably to the problem of denoising multi-dimensional data. Another projective subspace denoising method using delayed coordinates has been proposed recently with the algorithm dAMUSE. It combines the solution of blind source separation problems with denoising efforts in an elegant way and proofs to be very efficient and fast. Finally, KPCA represents a non-linear projective subspace method that is well suited for denoising also. Besides illustrative applications to toy examples and images, we provide an application of all algorithms considered to the analysis of protein NMR spectra. r

show abstract

Section: Article In Pressmentioning

confidence: 99%

Denoising using local projective subspace methods

et al. 2006

View full text Add to dashboard Cite

show abstract

“…Assuming that each sensor signal is a linear combination X = HS of N underlying but unknown source signals (s i ), a source signal trajectory matrix S can be written in analogy to eqn(1) and eqn (2). Then the mixing matrix (H) is a block matrix with a diagonal matrix in each block…”

Section: Generalized Eigendecomposition Using Time-delayed Signalsmentioning

confidence: 99%

“…There are several proposals to improve efficiency and robustness of these algorithms when noise is present [1], [2] which mostly rely on an approximative joint diagonalization of a set of correlation or cumulant matrices. Also there exist local projective de-noising techniques which in a first step increase the dimension of the data by joining delayed versions of the signals [3], [4], [5] hence projecting them into a high-dimensional feature space.…”

Section: Introductionmentioning

confidence: 99%

Delayed AMUSE – A Tool for Blind Source Separation and Denoising

Teixeira

Tomé

Lang

et al. 2004

Independent Component Analysis and Blind Signal Separation

View full text Add to dashboard Cite

Abstract. In this work we propose a generalized eigendecomposition (GEVD) of a matrix pencil computed after embedding the data into a high-dim feature space of delayed coordinates. The matrix pencil is computed like in AMUSE but in the feature space of delayed coordinates. Its GEVD yields filtered versions of the source signals as output signals. The algorithm is implemented in two EVD steps. Numerical simulations study the influence of the number of delays and the noise level on the performance.

show abstract

“…The variations in the time-frequency characteristics of the EMG regardless of the induced artifacts are even more complex. The solutions to process ECG or EEG signals [24], [25], [33] recorded in the same conditions are consequently not adequate for EMG recordings. Therefore, the development of new analog and digital processing tools requires a better understanding of the induced potentials behavior.…”

Section: Introductionmentioning

confidence: 99%

An Experimental Setup to Characterize MR Switched Gradient-Induced Potentials

Fokapu

El-Tatar

2013

IEEE Trans. Biomed. Circuits Syst.

View full text Add to dashboard Cite

We have developed an experimental setup as an in vitro research tool for studying the contamination of electrophysiological signals (EPS) by MRI environment; particularly, when due to the switched gradient-induced potentials. The system is composed of: 1) a MRI compatible module for the transmission of the EPS into the MRI tunnel, 2) a gelatin-based tissue-mimicking phantom, placed inside the tunnel, in which EPS is injected, 3) a detection module composed of a five input channel MRI compatible transmitter placed inside the tunnel, allowing an on-site pre-amplification of the bio-potentials and their transmission, via an optical fiber cable, to a four filtered output per channel receiver (350 Hz, 160 Hz, 80 Hz, and 40 Hz, for a total of 20 channels) placed in the control room, and 4) a signal processing algorithm used to analyze the generated induced potentials. A set of tests were performed to validate the electronic performances of the setup. We also present in this work an interesting application of the setup, i.e., the acquisition and analysis of the induced potentials with respect of the slice orientation for a given MRI sequence. Significant modifications of the time and frequency characteristics were observed with respect to axial, coronal or sagittal orientations.

show abstract

Second-order statistics based blind source separation using a bank of subband filters

Cited by 30 publications

References 7 publications

Denoising using local projective subspace methods

Denoising using local projective subspace methods

Delayed AMUSE – A Tool for Blind Source Separation and Denoising

An Experimental Setup to Characterize MR Switched Gradient-Induced Potentials

Contact Info

Product

Resources

About