Share Email Print
cover

Proceedings Paper

Estimation algorithms with noisy frame coefficients
Author(s): Alexander M. Powell
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

The Rangan-Goyal (RG) algorithm is a recursive method for constructing an estimate xNRd of a signal xRd, given Nd frame coefficient measurements of x that have been corrupted by uniform noise. Rangan and Goyal proved that the RG-algorithm is constrained by the Bayesian lower bound: lim infN→∞N2 E||xxN||2 > 0. As a positive counterpart to this, they also proved that for every p < 1 and xRd, the RG-algorithm satisfies limN→∞ Np||xxN|| = 0 almost surely. One consequence of the existing results is that one "almost" has mean square error E||xxN||2 of order 1/N2 for random choices of frames. It is proven here that the RG-algorithm achieves mean square error of the optimal order 1/N2, and the applicability of such error estimates is also extended to deterministic frames where ordering issues play an important role. Approximation error estimates for consistent reconstruction are also proven.

Paper Details

Date Published: 20 September 2007
PDF: 12 pages
Proc. SPIE 6701, Wavelets XII, 67010U (20 September 2007); doi: 10.1117/12.732748
Show Author Affiliations
Alexander M. Powell, Vanderbilt Univ. (United States)


Published in SPIE Proceedings Vol. 6701:
Wavelets XII
Dimitri Van De Ville; Vivek K. Goyal; Manos Papadakis, Editor(s)

© SPIE. Terms of Use
Back to Top