Catálogo de publicaciones - libros

The use of Blind Signal Separation methods (ICA and other approaches) for the analysis of astrophysical data remains quite unexplored. In this paper, we present a new approach for analyzing the infrared emission spectra of interstellar dust, obtained with NASA’s Spitzer Space Telescope, using and Non-negative Matrix Factorization (NMF). Using these two methods, we were able to unveil the spectra of three different types of carbonaceous nanoparticles present in interstellar space. These spectra can then constitute a basis for the interpretation of the mid-infrared emission spectra of interstellar dust in the Milky Way and nearby galaxies. We also show how to use these extracted spectra to derive the spatial distribution of these nanoparticles.

Blind Signal Separation (BSS) based on Independent Component Analysis (ICA) is an emerging approach which application is not limited to the signal processing research, where its application principle is rather straight forward. For an increasing amount of information processing fields, ICA has meaningful application which are still undiscovered. The aim of this paper is to investigate the ability of linguistic feature extraction based on word context preprocessing by ICA. The work refers to a first brief analysis in which ICA was applied to an English corpus. We continue this analysis depending on the number of components and the amount of syntactical information that we take into account. Furthermore we discuss to which extent the results deliver general linguistic features, or linguistic features giving us information about the text.

In modelling nonstationary sources, one possible strategy is to define a latent process of strictly positive variables to model variations in second order statistics of the underlying process. This can be achieved, for example, by passing a Gaussian process through a positive nonlinearity or defining a discrete state Markov chain where each state encodes a certain regime. However, models with such constructs turn out to be either not very flexible or non-conjugate, making inference somewhat harder. In this paper, we introduce a conjugate (inverse-) gamma Markov Random field model that allows random fluctuations on variances which are useful as priors for nonstationary time-frequency energy distributions. The main idea is to introduce auxiliary variables such that full conditional distributions and sufficient statistics are readily available as closed form expressions. This allows straightforward implementation of a Gibbs sampler or a variational algorithm. We illustrate our approach on denoising and single channel source separation.

Blind source separation (BSS) and Quantum Information Processing QIP) are two recent and rapidly evolving fields. No connection has ever been made between them to our knowledge. However, future practical QIP systems will probably involve ”observed mixtures”, in the BSS sense, of quantum states (qubits), e.g. associated to coupled spins. We here investigate how individual qubits may be retrieved from Heisenberg-coupled versions of them, and we show the relationship between this problem and classical BSS. We thus introduce new nonlinear mixture models for qubits, motivated by actual quantum physical devices. We analyze the invertibility and ambiguities of these models. We propose practical data processing methods for performing inversions.

This paper deals with the simulation of a ship-containers’ gantry crane cabin behavior, during an operation of load releasing and the BSS via ICA de-noising and movements separation. The goal consists of obtaining a reliable model of the cabin, with the aim of reducing the non-desired cabin vibrations. We present the -based simulation results and the result of the signal separation algorithms when the load is released by the crane in the containers’ ship. We conclude that the mass center position of the cabin affects dramatically to the vibrations of the crane. A set of graphs are presented involving displacements and rotations of the cabin to illustrate the effect of the mass center position’s bias and the results of the ICA action.

In recent works, we presented a blind image separation method based on a maximum likelihood approach, where we supposed the sources to be stationary, spatially autocorrelated and following Markov models. To make this method more adapted to real-world images, we here propose to extend it to non-stationary image separation. Two approaches, respectively based on blocking and kernel smoothing, are then used for the estimation of source score functions required for implementing the maximum likelihood approach, in order to allow them to vary within images. The performance of the proposed algorithm, tested on both artificial and real images, is compared to the stationary Markovian approach, and then to some classical blind source separation methods.

In this paper, a variant of the well known algorithm FastICA is proposed to be used for blind source separation in off-line (block processing) setup and a noisy environment. The algorithm combines a symmetric FastICA with test of saddle points to achieve fast global convergence and a one-unit refinement to obtain high noise rejection ability. A novel test of saddle points is designed for separation of complex-valued signals. The bias of the proposed algorithm due to additive noise can be shown to be asymptotically proportional to for small , where is the variance of the additive noise. Since the bias of the other methods (namely the bias of all methods using the orthogonality constraint, and even of recently proposed algorithm EFICA) is asymptotically proportional to , the proposed method has usually a lower bias, and consequently it exhibits a lower symbol-error rate, when applied to blind separation of finite alphabet signals, typical for communication systems.

Independent Component Analysis (ICA) is a statistical method for expressing an observed set of random vectors as a linear combination of statistically independent components. This paper tackles the task of comparing two ICA algorithms, in terms of their efficiency for compact representation of market securities. A recently developed sequential blind signal extraction algorithm, SmoothICA, is contrasted to a classical implementation of ICA, FastICA. SmoothICA uses an additional 2nd order constraint aiming at identifying temporally smooth components in the data set. This paper demonstrates the superiority of this novel smooth component extraction algorithm in terms of global and local approximation capability, applied to a portfolio of 60 NASDAQ securities, by utilizing common ordering algorithms for financial signals.

In this paper, we propose a Bayesian estimation approach to extend independent subspace analysis (ISA) for an overcomplete representation without imposing the orthogonal constraint. Our method is based on a synthesis of ISA [1] and overcomplete independent component analysis [2] developed by Hyvärinen et al. By introducing the variables of dot products (between basis vectors and whitened observed data vectors), we investigate the energy correlations of dot products in each subspace. Based on the prior probability of quasi-orthogonal basis vectors, the MAP (maximum a posteriori) estimation method is used for learning overcomplete independent feature subspaces. A gradient ascent algorithm is derived to maximize the posterior probability of the mixing matrix. Simulation results on natural images demonstrate that the proposed model can yield overcomplete independent feature subspaces and the emergence of phase- and limited shift-invariant features—the principal properties of visual complex cells.

In this paper, we develop an ICA-based obstacle detection and 3D-environment understanding for a mobile robot navigation. From a camera mounted on a mobile robot, the robot observes a sequence of images. This sequence of images allows the robot to compute optical flow, which is the apparent motion of each point on the image. We apply ICA to the optical flow field computed from images captured by the camera mounted on the robot. ICA-based separation of optical flow derives a obstacle region and a ground plane region in a space. For these applications, we also introduce an ordering criterion of independent components using its variances.