Deux méthodes de déconvolution et séparation simultanées; application à la reconstruction des amas de galaxies
Two approaches for the simultaneous separation and deblurring ; application to astrophysical data
OPEN ACCESS
Two approaches are presented to solve the problem of simultaneously deconvolving and separating mixtures of components. The first one uses a statistical description of the wavelet coefficients of the signals. The second one consists in minimizing a variational functional. Both methods are applied to the reconstruction of Sunyaev-Zel’dovich galaxy clusters from Cosmic Microwave Background experiments such as ACT. We find that both methods, when tuned, yield similar results and that the reconstruction of intense clusters is substantially improved when their non-gaussianity is taken into account.
Résumé
Nous présentons deux approches pour résoudre le problème de séparation et de déconvolution simultanées de mélanges de composantes. La première est basée sur une description statistique des coefficients d’ondelettes des signaux. La seconde consiste à minimiser une fonctionnelle variationnelle. Nous appliquons ces deux méthodes à la reconstruction des amas de galaxies par l’effet Sunyaev-Zel’dovich dans le cadre de la mission d’observation du fond diffus cosmique par ACT. Nous trouvons que pour des paramètres appropriés, les deux méthodes donnent des résultats comparables et que prendre en compte le caractère nongaussien des amas très intenses améliore nettement leur reconstruction.
Signal estimation/separation/deconvolution, statistical approach, variational approach, wavelets, astrophysics
Mots clés
Estimation/séparation/déconvolution de signaux, approche statistique, approche variationnelle, ondelettes, astrophysique
[1] ANTHOINE S., «Wavelet-based approaches for the deconvolution of blurred mixtures of components; Application to astrophysical data», Ph.D thesis, Princeton University, 2005.
[2] BENNETT C. L., HALPERN M., HINSHAW G. et al., «First Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Preliminary Maps and Basic Results», Astrophys.J.Suppl. Vol.148, #1, 2003.
[3] CHAMBOLLE A., LIONS P. L., «Image recovery via total variation minimization and related problems », Numerische Mathematik, Vol.76, #3, 1997, p.167-188.
[4] CROUSE M. S., NOWAK R. D., BARANIUK R. G., «Wavelet-based statistical signal processing using hidden Markov models», IEEE Trans. Signal Processing, Vol. 46, 1998, p. 886-902.
[5] DAUBECHIES I., DEFRISE M., DE MOL C., «An iterative thresholding algorithm for linear inverse problems with a sparsity constraint », Comm. Pure Appl. Math., Vol. 57, # 11, 2004, p.1413-1457.
[6] KINGSBURY N.G., «Complex wavelets for shift invariant analysis and filtering of signals», Journal of Applied and Computational Harmonic Analysis, vol 10, #3, 2001, p. 234-253.
[7] PIERPAOLI E., ANTHOINE S., HUFFENBERGER K., DAUBECHIES I., «Reconstructing Sunyaev-Zeldovich clusters in future CMB experiments». Mon. Not. Roy. Astron. Soc, vol.359, #1, 2005, p.261-271.
[8] PIERPAOLI E.,ANTHOINE S., «Finding SZ clusters in the ACBAR maps», Advances in Space Research, Vol.36, p.757-761, 2005.
[9] PORTILLA J.,STRELA V.,WAINWRIGHT, M.,SIMONCELLI E., «Image denoising using a scale mixture of Gaussians in the wavelet domain », IEEE Trans. Image Proc., vol. 12, #11, 2003, p.1338-1351.
[10] PORTILLA J.,SIMONCELLI E., «Image Restoration using Gaussian Scale Mixtures in the Wavelet Domain», 9th IEEE Int’l Conf on Image Processing, vol. II, 2003, Barcelona, Spain. p. 965-968.
[11] ROMBERG J., CHOI H., BARANIUK R., « Bayesian TreeStructured Image Modeling using Wavelet Domain Hidden Markov Models», IEEE Transactions on Image Processing, Vol. 10, #7, p. 1056-68, 2001.
[12] RUDIN L., OSHER S., FATEMI E., «Nonlinear total variation based noise removal algorithms», Physica D, Vol.60, 1992, p. 259-268.
[13] SENDUR L., SELESNICK I. W., «Bivariate shrinkage functions for wavelet-based denoising exploiting interscale dependency», IEEE Trans. on Signal Processing. Vol.50, #11, 2002, p. 2744-2756.
[14] SHAPIRO I. W., «Embedded image coding using zerotrees of wavelet coefficients», IEEE Trans. Signal Processing, vol. 41, 1993, p. 3445-3462.
[15] ZHANG P, PEN U., WANG B., «The Sunyaev-Zeldovich Effect: Simulations and Observations», ApJ, Vol.577, 2002, p.555-568.