Vicente Zarzoso

Prof. Vicente Zarzoso

orcid

Research Focus & Metrics

Academic Career Length

25Years (2002-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
18 December 2009 in IEEE Transactions on Neural Networks
Verified
Journal Article
Closed access
VZVicente Zarzoso
Vicente Zarzoso
PCPierre Comon
Pierre Comon
Independent component analysis (ICA) aims at decomposing an observed random vector into statistically independent variables. Deflation-based implementations, such as the popular one-unit FastICA algorithm and its variants, extract the independent components one after another. A novel method for deflationary ICA, referred to as RobustICA, is put forward in this paper. This simple technique consists of performing exact line search optimization of the kurtosis contrast function. The step size leading to the global maximum of the contrast along the search direction is found among the roots of a fourth-degree polynomial. This polynomial rooting can be performed algebraically, and thus at low cost, at each iteration. Among other practical benefits, RobustICA can avoid prewhitening and deals with real- and complex-valued mixtures of possibly noncircular sources alike. The absence of prewhitening improves asymptotic performance. The algorithm is robust to local extrema and shows a very high convergence speed in terms of the computational cost required to reach a given source extraction quality, particularly for short data records. These features are demonstrated by a comparative numerical analysis on synthetic data. RobustICA's capabilities in processing real-world data involving noncircular complex strongly super-Gaussian sources are illustrated by the biomedical problem of atrial activity (AA) extraction in atrial fibrillation (AF) electrocardiograms (ECGs), where it outperforms an alternative ICA-based technique.

29 August 2018 in IEEE Journal of Biomedical and Health Informatics
Verified
Journal Article
Closed access
ASAilton Siqueira
Ailton Siqueira
ASAmanda Franco Spirandeli
Amanda Franco Spirandeli
RMRaimes Moraes
Raimes Moraes
VZVicente Zarzoso
Vicente Zarzoso
Respiratory patterns are commonly measured to monitor and diagnose cardiovascular, metabolic and sleep disorders. Electronic devices such as masks used to record respiratory waveforms usually require medical staff support and obstruct the patients breathing, causing discomfort. New techniques are being investigated to overcome such limitations. An emerging approach involves accelerometers to estimate the respiratory waveform based on chest motion. However, most existing techniques employ a single accelerometer placed on an arbitrary thorax position. The present work investigates the use and optimal placement of multiple accelerometers located on the thorax and the abdomen. The study population is composed of thirty healthy volunteers in three different postures. By means of a custom-made microcontrolled system, data are acquired from an array of ten accelerometers located on predefined positions and a pneumotachograph used as reference. The best sensor locations are identified by optimal linear reconstruction of the reference waveform from the accelerometer data in the minimum mean square error sense. The analysis shows that right-hand side locations contribute more often to optimal respiratory waveform estimates, a sound finding given that the right lung has a larger volume than the left lung. In addition, we show that the respiratory waveform can be blindly extracted from the recorded accelerometer data by means of independent component analysis. In conclusion, linear processing of multiple accelerometers in optimal positions can successfully recover respiratory information in clinical settings where the use of masks may be contraindicated.

22 June 2016 in IEEE Transactions on Pattern Analysis and Machine Intelligence
Verified
Journal Article
Open access
RMRubén Martín-Clemente
Rubén Martín-Clemente
VZVicente Zarzoso
Vicente Zarzoso
Principal component analysis (PCA) based on L1-norm maximization is an emerging technique that has drawn growing interest in the signal processing and machine learning research communities, especially due to its robustness to outliers. The present work proves that L1-norm PCA can perform independent component analysis (ICA) under the whitening assumption. However, when the source probability distributions fulfil certain conditions, the L1-norm criterion needs to be minimized rather than maximized, which can be accomplished by simple modifications on existing optimal algorithms for L1-PCA. If the sources have symmetric distributions, we show in addition that L1-PCA is linked to kurtosis optimization. A number of numerical experiments illustrate the theoretical results and analyze the comparative performance of different algorithms for ICA via L1-PCA. Although our analysis is asymptotic in the sample size, this equivalence opens interesting new perspectives for performing ICA using optimal algorithms for L1-PCA with guaranteed global convergence while inheriting the increased robustness to outliers of the L1-norm criterion.

1 January 2020 in IEEE Transactions on Signal Processing
Verified
Journal Article
Closed access
JGJose Henrique de Morais Goulart
Jose Henrique de Morais Goulart
POPedro Marinho Ramos de Oliveira
Pedro Marinho Ramos de Oliveira
RFRodrigo Cabral Farias
Rodrigo Cabral Farias
VZVicente Zarzoso
Vicente Zarzoso
PCPierre Comon
Pierre Comon
In some applications, blind source separation can be performed by computing an approximate block-term tensor decomposition (BTD), under much milder constraints than matrix-based techniques. However, choosing the BTD model structure (i.e., the number of blocks and their ranks) is a difficult problem, and the standard least-squares formulation can be ill-posed. This paper proposes an alternating group lasso algorithm to compute approximate low-rank BTDs. It solves, in a provably convergent manner, a well-posed mixed-norm regularized tensor approximation problem that allows estimating the model parameters and its structure jointly. A variant is also put forward for dealing with linearly constrained blocks, motivated by the problem of blind separation of sums of complex exponentials, which can be cast as a low-rank Hankel-structured block-term tensor approximation problem. An experimental comparison with a standard nonlinear least-squares algorithm on synthetic tensor data indicates that the proposed algorithm is much more robust with respect to initialization. We also apply the constrained variant to the extraction of atrial activity from semi-synthetic and real-world electrocardiogram recordings during atrial fibrillation episodes. Our results show its ability to consistently select an adequate structure and to extract multiple signals which can be physiologically interpreted as atrial fibrillation patterns.

1 December 2016 in Archives of Cardiovascular Diseases
Verified
Journal Article
Closed access
VZVicente Zarzoso
Vicente Zarzoso
DLDecebal G. Latcu
Decebal G. Latcu
Antonio R. Hidalgo-MuñozAntonio R. Hidalgo-Muñoz
Antonio R. Hidalgo-Muñoz
MMMarianna Meo
Marianna Meo
OMOlivier Meste
Olivier Meste
NSNadir Saoudi
Nadir Saoudi

11 September 2020 in IEEE Transactions on Signal Processing
Verified
Journal Article
Closed access
VKVaclav Kautsky
Vaclav Kautsky
ZKZbynek Koldovsky
Zbynek Koldovsky
PTPetr Tichavsky
Petr Tichavsky
VZVicente Zarzoso
Vicente Zarzoso
Blind source extraction (BSE) aims at recovering an unknown source signal of interest from the observation of instantaneous linear mixtures of the sources. This paper presents Cramer-Rao Lower Bound (CRLB) for the complex-valued BSE problem based on the assumption that the target signal is independent of the other signals. The target source is assumed to be non-Gaussian or non-circular Gaussian while the other signals (background) are circular Gaussian or non-Gaussian. The results confirm some previous observations known for the real domain and yield new results for the complex domain. Also, the CRLB for ICE is shown to coincide with that for ICA when the non-Gaussianity of background is taken into account. Second, we extend the CRLB analysis to piecewise determined mixing models, where the observed signals are assumed to obey the determined mixing model within short blocks where the mixing matrices can be varying from block to block. This model has applications, for instance, when separating dynamic mixtures. Either the mixing vector or the separating vector corresponding to the target source is assumed to be constant across the blocks. The CRLBs for the parameters of these models bring new performance limits for the BSE problem.