EPFL | Biomedical Imaging Group | Signal Reconstruction

Splines and Optimal Signal Reconstruction

M. Unser

Invited tutorial, Meeting on "Information, Signal, Images et ViSion: Thème A—Traitement Statistique de l'Information [Échantillonnage Irrégulier]" (ISIS-A-EI'08), Paris, French Republic, April 18, 2008.

We consider the generic problem of reconstructing a signal from its noisy samples. We argue that an “optimal” solution can be specified through the minimization of a hybrid cost function that is the sum of a discrete-domain data term, and a continuous-domain regularization functional that forces the reconstruction to be well behaved.

In order to derive a practical algorithm, we propose to represent the solution using compactly-supported B-spline basis functions, which has a number of computational advantages. In the case where the data is uniformly sampled and the regularization function quadratic (Tikhonov criterion), this yields a smoothing spline estimator that can be implemented efficiently by digital filtering. We also propose an alternative stochastic formulation (hybrid Wiener filter) that leads to the same type of algorithm. We prove that both solutions are globally-optimal and computationally equivalent, provided that the reconstruction space is matched to the regularization operator (deterministic signal) or, alternatively, to the whitening operator of the process (stochastic modeling). This suggests a unifying interpretation of the optimal reconstruction process in terms of generalized splines. Finally, we address the problem of the reconstruction of a multidimensional signal from non-uniform samples. We review the classical thin-plate spline solution, and present an alternative B-spline-based algorithm that is computationally much more efficient.

References

M. Unser, "Splines: A Perfect Fit for Signal and Image Processing," IEEE Signal Processing Magazine, vol. 16, no. 6, pp. 22-38, November 1999.
M. Unser, T. Blu, "Generalized Smoothing Splines and the Optimal Discretization of the Wiener Filter," IEEE Transactions on Signal Processing, vol. 53, no. 6, pp. 2146-2159, June 2005.
M. Arigovindan, M. Sühling, P. Hunziker, M. Unser, "Variational Image Reconstruction from Arbitrarily Spaced Samples: A Fast Multiresolution Spline Solution," IEEE Transactions on Image Processing, vol. 14, no. 4, pp. 450-460, April 2005.

@INPROCEEDINGS(http://bigwww.epfl.ch/publications/unser0807.html,
AUTHOR="Unser, M.",
TITLE="Splines and Optimal Signal Reconstruction",
BOOKTITLE="Meeting on ``Information, Signal, Images et ViSion:
	Th{\`{e}}me A---Traitement Statistique de l'Information
	[{\'{E}}chantillonnage Irr{\'{e}}gulier]'' ({ISIS-A-EI'08})",
YEAR="2008",
editor="",
volume="",
series="",
pages="",
address="Paris, French Republic",
month="April 18,",
organization="",
publisher="",
note="Invited tutorial")

© 2008 ISIS. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from ISIS. This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.