An informed source separation of astrophysical ice analogs

doi:10.1016/j.dsp.2007.02.003

Digital Signal Processing

Volume 17, Issue 5, September 2007, Pages 947-964

https://doi.org/10.1016/j.dsp.2007.02.003 Get rights and content

Abstract

The analysis of astrophysical ices and the determination of the compounds that are present in the molecular clouds play a fundamental role in order to predict the future evolution of the cloud, e.g., its transformation to protostellar bodies or the appearance of new radicals and molecules. Because of the difficulties of obtaining satellite data, the process is simulated first in the laboratory generating ice analogs under well controlled variables. In this case, the ice mixture is carried on allocating the different components in the appropriate concentrations and recording the spectrum of the aggregated ice. This process tries to simulate the real process of forming ice mantles under the environmental conditions of the interstellar medium. The spectrum of each ice can be modeled as the linear instantaneous superposition of the spectrum of the different compounds, so a source separation approach is proper. In addition, some priors about the sources and the mixing matrix entries can be assumed, obtaining an informed Bayesian approach to the problem. We present the results obtained with the variational Bayesian approach for simulated and real mixtures, showing the good performance of the algorithm.

Section snippets

Jorge Igual was born in Valencia (Spain) in 1969. He received the Ingeniero de Telecomunicación and the Doctor Ingeniero de Telecomunicación degrees from the Universidad Politécnica de Valencia (UPV) in 1993 and 2000, respectively. Since 1993 he is working at the Departamento de Comunicaciones (UPV) as an Associate Professor.

His research concentrates in the machine learning and statistical signal processing area, where he has published in different journals. He has also participated in projects

References (45)

P. Comon
Independent component analysis: A new concept?
Signal Process.
(1994)
C. Jutten et al.
Blind separation of sources. I: An adaptive algorithm based on neuromimetic architecture
Signal Process.
(1991)
E.E. Kuruoglu et al.
Source separation in astrophysical maps using independent factor analysis
Neural Networks
(2003)
E. Oja et al.
Independent Component Analysis
(2001)
J.F. Cardoso
Blind signal separation: Statistical principles
Proc. IEEE
(1998)
S. Amari et al.
Adaptive blind signal processing, neural network approaches
Proc. IEEE
(1998)
J.F. Cardoso et al.
Blind beamforming for nonGaussian signals
Proc. Inst. Elec. Eng. F
(1993)
A.J. Bell et al.
An information–maximization approach to blind separation and blind deconvolution
Neural Comput.
(1995)
D.T. Pham, P. Garat, C. Jutten, Separation of a mixture of independent sources through a maximum likelihood approach,...

C. Jutten, Calcul neuromimétique et traitement du signal. Analyse en composantes indépendentes, Ph.D. thesis, INP-USM,...

H. Lou et al.

Ensemble learning for independent component analysis of normal galaxy spectra

Astron. J.

(2006)

D. Nuzillard et al.

Blind source separation and analysis of multispectral astronomical images

Astron. Astrophys. Suppl. Ser.

(2000)

G. Patanchon, J. Delabrouille, F.F. Cardoso, Source separation on astrophysical data sets from the WMAP satellite, in:...

D. Maino et al.

All-sky astrophysical component separation with (FastICA)

Mon. Not. R. Astron. Soc.

(2002)

M. Costagli, E.E. Kuruoglu, A. Ahmed, Astrophysical source separation using particle filters, in: 5th Int. Conference...

H. Snoussi, G. Patanchon, J. Macias, A. Mohammad-Djafari, J. Delabrouille, Bayesian blind component separation for...

O. Forni et al.

Component separation of OMEGA spectra with ICA

Lunar Planet. Sci.

(2005)

L. Parra et al.

Unmixing Hyperspectral Data, Advances in Neural Information Processing Systems

(2000)

J. Bayliss, J. Gualtieri, R. Cromp, Analysing hyperspectral data with ICA, in: Proc. SPIE Applied Image and Pattern...

S.L. Miller

Ann. N.Y. Acad. Sci.

(1957)

P. Gerakines et al.

The infrared band strengths of H₂O, CO and CO₂ in laboratory simulations of astrophysical ice mixtures

Astron. Astrophys.

(1995)

Cited by (4)

Separation of galaxy spectra measured with slitless spectroscopy
2020, Digital Signal Processing: A Review Journal
Citation Excerpt :
Blind Source Separation (BSS), where little information about the mixing system is available, has largely been considered in several application domains [1,2]. In spectroscopy, the BSS methods have e.g. been used to separate source spectra in Earth observation [3,4], chemistry [5,6], nuclear magnetic resonance [7,8] and astronomy [9–11]. Here, we address a new application concerning the separation of spectra measured with slitless spectroscopy.
This paper investigates the problem of separating galaxy spectra resulting from the slitless spectroscopy system, which can e.g. be used in spatial missions like the WISP survey and the Euclid project. We first derive a physical mixing model linking observed data to source spectra, then simplify it to obtain two different approximate but realistic models. The first simplified model being entirely defined by a few parameters, we propose a semi-blind source separation method which estimates these model parameters together with the source spectra. The second simplified model is linear instantaneous, and has a special form which is used to propose a new blind source separation method exploiting the non-negativity and spatial sparsity of data, as well as the correlation of the target source spectra in different light dispersion directions. Both methods are tested on realistic simulated data and lead to very encouraging results, confirming their effectiveness.
Application of regularized Alternating Least Squares to an astrophysical problem
2012, Applied Mathematics and Computation
Citation Excerpt :
The laboratory spectrum does not correspond exactly to the astrophysical ice spectrum, due to measurement problems such as the inclination of ice film with respect to the infrared beam of the spectrometer and the polarization of the beam [11], the purity of the gases, small changes in the temperature, sequential mixing of gases, or the influence of the molecular mass in the deposition rate [12]. The review of some physical considerations about the absorption spectra and the molecules to justify the conditions of suitability of the BSS model (2) for the problem is carried out in [3]. The data corresponds to the public ice analogs database of the University of Leiden [18].
Determination of the compounds that are present in molecular clouds is carried out from the study of the infrared spectrum of astrophysical ices. This analysis plays a fundamental role in the prediction of the future evolution of the cloud under study. The process is simulated in the laboratory under similar conditions of thermal and energetic processing, recording the infrared absorption spectrum of the resultant ice. The spectrum of each ice can be modeled as the linear instantaneous superposition of the spectrum of the different compounds, so a Source Separation approach is appropriate. We propose the use of Alternating Least Squares and a Regularized version to identify the molecules that are present in the ice mixtures. Since the spectra and abundances are non-negative, a non-negativity constraint can be applied to obtain solutions with physical meaning. We perform several simulations of synthetic and real mixtures of ices in order to compare both solutions and to show how the proposed approach provides an efficient way to recover underlying spectral patterns that are physically meaningful.
Analysis of astrophysical ice analogs using regularized alternating least squares
2010, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Nonnegative matrix factorization of laboratory astrophysical ice mixtures
2008, IEEE Journal on Selected Topics in Signal Processing

His research concentrates in the machine learning and statistical signal processing area, where he has published in different journals. He has also participated in projects under contract with the industry in communications, computer vision, and other fields. Currently he is involved in applications of independent component analysis in ultrasonic, images, and astrophysics.

Raúl Llinares was born in Alcoy (Spain) in 1977. He received the Ingeniero de Telecomunicación degree from the Universidad Politécnica de Valencia (UPV) in 2001. Since 2001 he is working at the Departamento de Comunicaciones (UPV) as an Associate Professor.

His research concentrates in the statistical signal processing area, where he has worked and published on different theoretical and applied problems. Currently he is involved in independent component analysis and non-negative matrix factorization with applications in physics and biomedicine.

View full text

An informed source separation of astrophysical ice analogs

Abstract

Section snippets

Signal Process.

Signal Process.

Neural Networks

Independent Component Analysis

Blind signal separation: Statistical principles

Proc. IEEE

Adaptive blind signal processing, neural network approaches