Complex ICA Using Nonlinear Functions
IEEE Transactions on Signal ProcessingPublished 14 August 2008
Tülay Adalı, Hualiang Li, M. Novey, J.-F. Cardoso
Citations112
SJR quartileQ1
SJR score2.05
SNIP2.36
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
A framework based on Wirtinger calculus for nonlinear complex-valued signal processing such that all computations can be directly carried out in the complex domain is introduced.
Abstract
We introduce a framework based on Wirtinger calculus for nonlinear complex-valued signal processing such that all computations can be directly carried out in the complex domain. The two main approaches for performing independent component analysis, maximum likelihood, and maximization of non-Gaussianity-which are intimately related to each other-are studied using this framework. The main update rules for the two approaches are derived, their properties and density matching strategies are discussed along with numerical examples to highlight their relationships.
Keywords
ChemistryComputer Science
Neural ComputationAn Information-Maximization Approach to Blind Separation and Blind Deconvolution
9,194 Citations1995Anthony J. Bell, Terrence J. Sejnowski
It is suggested that information maximization provides a unifying framework for problems in "blind" signal processing and dependencies of information transfer on time delays are derived.
IEEE Transactions on Neural NetworksFast and robust fixed-point algorithms for independent component analysis
6,275 Citations1999Aapo Hyvärinen
Using maximum entropy approximations of differential entropy, a family of new contrast (objective) functions for ICA enable both the estimation of the whole decomposition by minimizing mutual information, and estimation of individual independent components as projection pursuit directions.
IEE Proceedings F Radar and Signal ProcessingBlind beamforming for non-gaussian signals
2,729 Citations1993J.-F. Cardoso, Antoine Souloumiac
Signal ProcessingBlind separation of sources, part I: An adaptive algorithm based on neuromimetic architecture
2,522 Citations1991Christian Jutten, Jeanny Hérault
Neural ComputationIndependent Component Analysis Using an Extended Infomax Algorithm for Mixed Subgaussian and Supergaussian Sources
1,978 Citations1999Te-Won Lee, Mark Girolami +1 more
An extension of the infomax algorithm of Bell and Sejnowski (1995) is presented that is able blindly to separate mixed signals with sub- and supergaussian source distributions and is effective at separating artifacts such as eye blinks and line noise from weaker electrical signals that arise from sources in the brain.
A New Learning Algorithm for Blind Signal Separation
1,888 Citations1995Шун-ичи Амари, Andrzej Cichocki +1 more
A new on-line learning algorithm which minimizes a statistical dependency among outputs is derived for blind separation of mixed signals and has an equivariant property and is easily implemented on a neural network like model.
Proceedings of the IEEEBlind signal separation: statistical principles
1,852 Citations1998J.-F. Cardoso
The objectives of this paper are to review some of the approaches that have been developed to address blind signal separation and independent component analysis, to illustrate how they stem from basic principles, and to show how they relate to each other.
IEEE Transactions on Signal ProcessingEquivariant adaptive source separation
1,346 Citations1996J.-F. Cardoso, B.H. Laheld
A class of adaptive algorithms for source separation that implements an adaptive version of equivariant estimation and is henceforth called EASI, which yields algorithms with a simple structure for both real and complex mixtures.
IEEE Transactions on Information TheoryProper complex random processes with applications to information theory
914 Citations1993F.D. Neeser, James L. Massey
The covariance of complex random variables and processes, when defined consistently with the corresponding notion for real random variables, is shown to be determined by the usual complex covariance together with a quantity called the pseudo-covariance.
International Journal of Neural SystemsA FAST FIXED-POINT ALGORITHM FOR INDEPENDENT COMPONENT ANALYSIS OF COMPLEX VALUED SIGNALS
803 Citations2000Ella Bingham, Aapo Hyvärinen
In this article, a fast fixed-point type algorithm that is capable of separating complex valued, linearly mixed source signals is presented and its computational efficiency is shown by simulations and the local consistency of the estimator given by the algorithm is proved.
NeurocomputingBlind separation of convolved mixtures in the frequency domain
783 Citations1998Paris Smaragdis
It is observed that convolved Mixing in the time domain corresponds to instantaneous mixing in the frequency domain, and convolved mixing can be inverted using simpler and more robust algorithms than the ones recently developed.
IEEE Transactions on Signal ProcessingOn circularity
505 Citations1994B. Picinbono
The theory of linear mean square estimation for complex signals exhibits some connections with circularity, and it is shown that without this assumption, the estimation theory must be reformulated.
IEEE Transactions on Signal ProcessingThe complex backpropagation algorithm
489 Citations1991Henry Leung, S. Haykin
The backpropagation algorithm that provides a popular method for the design of a multilayer neural network to include complex coefficients and complex signals so that it can be applied to general radar signal processing and communications problems is generalized.
IEEE Transactions on Signal ProcessingBlind separation of mixture of independent sources through a quasi-maximum likelihood approach
464 Citations1997Dinh Tuan Pham, Philippe Garat
Two methods for separating mixture of independent sources without any precise knowledge of their probability distribution are proposed by considering a maximum likelihood (ML) solution corresponding to some given distributions of the sources and relaxing this assumption afterward.
IEEE Transactions on Signal ProcessingSecond-order analysis of improper complex random vectors and processes
379 Citations2003Peter J. Schreier, Louis L. Scharf
The concepts of propriety and joint propriety are linked to eigenanalysis and canonical correlation analysis and applied to the problem of rank reduction through principal components of complex random vectors and wide-sense stationary signals.
arXiv (Cornell University)The Complex Gradient Operator and the CR-Calculus
376 Citations2009Ken Kreutz-Delgado
International Journal of Neural SystemsA Fast Fixed-Point Algorithm for Independent Component Analysis of Complex Valued Signals
346 Citations2000Ella Bingham
IEE Proceedings H Microwaves Optics and AntennasA complex gradient operator and its application in adaptive array theory
335 Citations1983D.H. Brandwood
IEEE Transactions on Neural NetworksEfficient Variant of Algorithm FastICA for Independent Component Analysis Attaining the CramÉr-Rao Lower Bound
301 Citations2006Zbyněk Koldovský, Petr Tichavský +1 more
An improved version of the FastICA algorithm is proposed which is asymptotically efficient, i.e., its accuracy given by the residual error variance attains the Cramer-Rao lower bound (CRB).
Mathematics and Computers in SimulationTheory of complex functions
285 Citations1991Reinhold Remmert, R. B. Burckel +5 more
IEEE Transactions on Circuits and Systems I Fundamental Theory and ApplicationsRobust neural networks with on-line learning for blind identification and blind separation of sources
259 Citations1996Andrzej Cichocki, R. Unbehauen
Two unsupervised, self-normalizing, adaptive learning algorithms are developed for robust blind identification and/or blind separation of independent source signals from a linear mixture of them and are suitable for real-time implementations.
IEEE Transactions on Signal ProcessingBlind source separation-semiparametric statistical approach
246 Citations1997Шун-ичи Амари, J.-F. Cardoso
It is shown that an estimator of the mixing matrix or its learning version can be described in terms of an estimating function, which means that one can find a better or equally good estimator in the admissible part.
Neural ComputationApproximation by Fully Complex Multilayer Perceptrons
245 Citations2003Taehwan Kim, Tülay Adalı
Three proofs of the approximation capability of the fully complex MLP are provided based on the characteristics of singularity among ETFs, which shows the output of complex MLPs using ETFs with isolated and essential singularities uniformly converges to any nonlinear mapping in the deleted annulus of singularities nearest to the origin.
Neural NetworksComplex independent component analysis of frequency-domain electroencephalographic data
209 Citations2003Jörn Anemüller, Terrence J. Sejnowski +1 more
This work decomposes the recorded spectral-domain signals into independent components by a complex infomax ICA algorithm, leading to a model of convolutive signal superposition, in contrast with the commonly used instantaneous mixing model.
Magnetic Resonance ImagingPerformance of blind source separation algorithms for fMRI analysis using a group ICA method
193 Citations2006Nicolle Correa, Tülay Adalı +1 more
This paper studies the performance of four major classes of algorithms for spatial ICA, namely, information maximization, maximization of non-Gaussianity, joint diagonalization of cross-cumulant matrices and second-order correlation-based methods, when they are applied to fMRI data from subjects performing a visuo-motor task.
Complex random vectors and ICA models: Identifiability, uniqueness and separability
179 Citations2006Jenny Eriksson, Visa Koivunen +1 more
IEEE Transactions on Information TheoryComplex random vectors and ICA models: identifiability, uniqueness, and separability
167 Citations2006J. Eriksson, V. Koivunen
The conditions for identifiability, separability and uniqueness of linear complex valued independent component analysis (ICA) models are established and the Darmois-Skitovich theorem for complex-valued models is extended.
The Journal of VLSI Signal Processing Systems for Signal Image and Video TechnologyFlexible Independent Component Analysis
165 Citations2000Seungjin Choi, Andrzej Cichocki +1 more
One-unit contrast functions for independent component analysis: a statistical analysis
119 Citations2002Aapo Hyvärinen
The author's large family of one-unit contrast functions to be used in independent component analysis (ICA) is analyzed mathematically in the case of a finite sample, providing strong arguments in favor of the use of contrast functions based on slowly growing functions, and against theUse of kurtosis, which is the classical contrast function.
IEEE Transactions on Information TheoryDetection and Estimation of Improper Complex Random Signals
117 Citations2005Peter J. Schreier, Louis L. Scharf +1 more
A general result is derived comparing the performance of conventional processing, which ignores complementary covariances, with processing that takes these into account, and it is shown how this finding generalizes the result that coherent processing enjoys a 3-dB gain over noncoherent processing.
Signal ProcessingBlind separation methods based on Pearson system and its extensions
83 Citations2002Juha Karvanen, Visa Koivunen
A mutual lnformation-based method for blind separation ot statistically independent source signals and an extension of the Pearson system that can model multimodal distributions.
Complex-valued ICA using second order statistics
80 Citations2005Jenny Eriksson, Visa Koivunen
A novel transformation called strong-uncorrelating transform (SUT) is proposed that can be viewed as an ex- tension of the conventional whitening transform for complex ran- dom vectors.
IEEE Transactions on Neural NetworksEfficient source adaptivity in independent component analysis
73 Citations2001Nikos Vlassis, Yoichi Motomura
It is shown with a real and a synthetic experiment that this method can provide more accurate solutions than state-of-the-art methods when optimization is carried out in the vicinity of the global minimum of the contrast function.
Independent component analysis by complex nonlinearities
65 Citations2004Tülay Adalı, Tae-Hwan Kim +1 more
A number of complex nonlinear functions are proposed for the independent component analysis of complex-valued data and their efficiency in generating the higher order statistics needed for ICA is shown.
Zenodo (CERN European Organization for Nuclear Research)On The Blind Separation Of Non-Circular Sources
65 Citations2002Lieven De Lathauwer, Bart De Moor
It is proved that, if the sources are non-circular at order 2, by exploiting all the second-order information, the mixing matrix can be estimated up to a real orthogonal factor, based on a link with the Takagi factorization, for the computation of which a Jacobi-type algorithm is derived.
Pearson System based Method for Blind Separation
59 Citations2000Juha Karvanen, Jenny Eriksson +1 more
IEEE Transactions on Neural NetworksSelf-Adaptive Blind Source Separation Based on Activation Functions Adaptation
56 Citations2004Lenan Zhang, Andrzej Cichocki +1 more
A general framework of blind separation from a practical point of view with special emphasis on the activation function adaptation is developed and the exponential generative model for probability density functions is proposed.
Machine learning for signal processing ...Adaptable Nonlinearity for Complex Maximization of Nongaussianity and a Fixed-Point Algorithm
42 Citations2006M. Novey, Tülay Adalı
A fixed-point algorithm for blind separation of noncircular sources using CMN is derived and the adaptive CMN (A-CMN) algorithm that provides significant performance improvement by adapting the nonlinearity to the source distribution is introduced.
The Maximum Likelihood Approach to Complex ICA
41 Citations2006J.-F. Cardoso, T. Adalt
It is shown that both the form of nonlinearity and the relative gradient update equations for likelihood maximization naturally generalize to the complex case, and that they coincide with the real case.
A POLAR-COORDINATE BASED ACTIVATION FUNCTION FOR FREQUENCY DOMAIN BLIND SOURCE SEPARATION
30 Citations2001Hiroshi Sawada, Ryo Mukai +2 more
The results suggest that NAI did not provide these adults with significant defense against rapid re- Infection and suggest that this population-infection design could serve to demonstrate the efficacy of a drug or vaccine in preventing parasitemia.
ICA by Maximization of Nongaussianity using Complex Functions
28 Citations2005M. Novey, Tülay Adalı
The CMN algorithm is introduced and it is shown that CMN converges to the principal component of the source distribution and that the algorithm provides robust performance for both circular and non-circular sources.
Machine learning for signal processing ...Gradient and Fixed-Point Complex ICA Algorithms Based on Kurtosis Maximization
22 Citations2006Hualiang Li, Tülay Adalı
Two algorithms for independent component analysis of complex-valued signals based on the maximization of absolute value of kurtosis are presented and their properties are established.
Adaptive subband video coding using bivariate generalized Gaussian distribution model
21 Citations2002Muhammed Coban, R.M. Mersereau
It is shown that use of bivariate generalized Gaussian PDF models yields a significant improvement in rate-distortion performance over univariate GGD models.
IEEE Transactions on Information TheoryOn probability density functions for complex variables
20 Citations2006S.C. Olhed
The main purpose of this correspondence is to clarify the concept of a density function for a complex random variable, and to discuss its properties.
The Journal of VLSI Signal Processing Systems for Signal Image and Video TechnologyComplex Infomax: Convergence and Approximation of Infomax with Complex Nonlinearities
19 Citations2006Vince D. Calhoun, Tülay Adalı
Complex Fixed-Point ICA Algorithm for Separation of QAM Sources using Gaussian Mixture Model
16 Citations2007M. Novey, Tülay Adalı
A fixed-point algorithm for separation of quadrature amplitude modulated (QAM) sources through independent component analysis is introduced and it is noted that it provides improved performance over a wide range of operating conditions such as low signal-to-noise ratio, small sample sizes, and large number of sources.
2003 IEEE International Conference on Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03).Complex ICA for fMRI analysis: performance of several approaches
16 Citations2003Vince D. Calhoun, Tülay Adalı
The complex infomax algorithm that uses analytic non linearity has the ability to separate both sub- and super-Gaussian sources with a hyperbolic tangent nonlinearity and provides a potentially powerful method for exploratory analysis of fMRI data.