DocumentCode :
3281771
Title :
Sparse reconstruction by convex relaxation: Fourier and Gaussian measurements
Author :
Rudelson, Mark ; Vershynin, Roman
Author_Institution :
Dept. of Math., Univ. of Missouri, Columbia, MO
fYear :
2006
fDate :
22-24 March 2006
Firstpage :
207
Lastpage :
212
Abstract :
This paper proves best known guarantees for exact reconstruction of a sparse signal f from few non-adaptive universal linear measurements. We consider Fourier measurements (random sample of frequencies of f) and random Gaussian measurements. The method for reconstruction that has recently gained momentum in the sparse approximation theory is to relax this highly non-convex problem to a convex problem, and then solve it as a linear program. What are best guarantees for the reconstruction problem to be equivalent to its convex relaxation is an open question. Recent work shows that the number of measurements k(r,n) needed to exactly reconstruct any r-sparse signal f of length n from its linear measurements with convex relaxation is usually O(r poly log (n)). However, known guarantees involve huge constants, in spite of very good performance of the algorithms in practice. In attempt to reconcile theory with practice, we prove the first guarantees for universal measurements (i.e. which work for all sparse functions) with reasonable constants. For Gaussian measurements, k(r,n) lsim 11.7 r [1.5 + log(n/r)], which is optimal up to constants. For Fourier measurements, we prove the best known bound k(r, n) = O(r log(n) middot log2(r) log(r log n)), which is optimal within the log log n and log3 r factors. Our arguments are based on the technique of geometric functional analysis and probability in Banach spaces.
Keywords :
Banach spaces; Gaussian processes; approximation theory; geometry; image reconstruction; linear programming; relaxation theory; Banach spaces; Fourier measurements; convex relaxation; geometric functional analysis; linear program; nonadaptive universal linear measurements; probability; random Gaussian measurements; sparse approximation theory; sparse reconstruction; sparse signal; Approximation methods; Collaborative work; Frequency measurement; Functional analysis; History; Length measurement; Linear programming; Mathematics; Relaxation methods; Time measurement;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Information Sciences and Systems, 2006 40th Annual Conference on
Conference_Location :
Princeton, NJ
Print_ISBN :
1-4244-0349-9
Electronic_ISBN :
1-4244-0350-2
Type :
conf
DOI :
10.1109/CISS.2006.286463
Filename :
4067804
Link To Document :
بازگشت