Abstract
In compressed sensing, we seek to gain information about a vector x∈ℝN from d ≪ N nonadaptive linear measurements. Candes, Donoho, Tao et al. (see, e.g., Candes, Proc. Intl. Congress Math., Madrid, 2006; Candes et al., Commun. Pure Appl. Math. 59:1207–1223, 2006; Donoho, IEEE Trans. Inf. Theory 52:1289–1306, 2006) proposed to seek a good approximation to x via ℓ 1 minimization. In this paper, we show that in the case of Gaussian measurements, ℓ 1 minimization recovers the signal well from inaccurate measurements, thus improving the result from Candes et al. (Commun. Pure Appl. Math. 59:1207–1223, 2006). We also show that this numerically friendly algorithm (see Candes et al., Commun. Pure Appl. Math. 59:1207–1223, 2006) with overwhelming probability recovers the signal with accuracy, comparable to the accuracy of the best k-term approximation in the Euclidean norm when k∼d/ln N.
Similar content being viewed by others
References
R. Baraniuk, M. Davenport, R. DeVore, M. Wakin, A simple proof of the restricted isometry property for random matrices, Constr. Approx. 28(3), 253–263 (2008).
E. Candes, Compressive sampling, in Proc. Intl. Congress Math., Madrid, 2006.
E. Candes, The restricted isometry property and its implications for compressed sensing, C. R. Acad. Sci. Paris, Ser. I 346, 589–592 (2008).
E. Candes, J. Romberg, T. Tao, Stable signal recovery from incomplete and inaccurate measurements, Commun. Pure Appl. Math. 59, 1207–1223 (2006).
E. Candes, T. Tao, Near optimal recovery from random projections: universal encoding strategies, IEEE Trans. Inf. Theory 52, 5406–5425 (2004).
E. Candes, T. Tao, Decoding by linear programing, IEEE Trans. Inf. Theory 51, 4203–4215 (2005).
A. Cohen, W. Dahmen, R. DeVore, Compressed sensing and best k-term approximation, J. Am. Math. Soc. 22(1), 211–231 (2009).
R. DeVore, G. Petrova, P. Wojtaszczyk, Instance-optimality in probability with an ℓ 1-minimization decoder. http://www.dsp.ece.rice.edu/cs/.
D. Donoho, Compressed sensing, IEEE Trans. Inf. Theory 52, 1289–1306 (2006).
A. Garnaev, E.D. Gluskin, The widths of Euclidean balls, Dokl. Akad. Nauk SSSR 277, 1048–1052 (1984).
B.S. Kashin, The widths of certain finite dimensional sets and classes of smooth functions, Izv. Akad. Nauk SSSR 41, 334–351 (1977).
A. Litvak, P. Mankiewicz, N. Tomczak-Jaegermann, Randomized isomorphic Dvoretzky theorem, C. R. Math. Acad. Sci. Paris 335(4), 345–350 (2002).
A. Litvak, A. Pajor, M. Rudelson, N. Tomczak-Jaegermann, Smallest singular value of random matrices and geometry of random polytopes, Adv. Math. 195, 491–523 (2005).
G.G. Lorentz, M.V. Golitschek, Y. Makovoz, Constructive Approximation; Advanced Problems (Springer, Berlin, 1996).
P. Mankiewicz, Compact groups of operators on subproportional quotients of ℓ m1 , Canad. J. Math. 52(5), 999–1017 (2000).
S. Mendelson, A. Pajor, N. Tomczak-Jaegermann, Reconstruction and subgaussian processes, C. R. Acad. Sci. Paris, Ser. I Math. 340, 885–888 (2005).
S. Mendelson, A. Pajor, N. Tomczak-Jaegermann, Uniform uncertainity principle for Bernoulli and subgaussian ensembles, Constr. Approx. 28(3), 277–289 (2008).
D. Needell, R. Vershynin, Uniform uncertainty principle and signal recovery via regularised orthogonal matching pursuit. http://www.dsp.ece.rice.edu/cs/.
D. Needell, R. Vershynin, Signal recovery from incomplete and inaccurate measurements via regularised orthogonal matching pursuit. http://www.dsp.ece.rice.edu/cs/.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Emmanuel Candes.
P. Wojtaszczyk would like to express his gratitude to Piotr Mankiewicz for answering questions about convex geometry and to Albert Cohen, Wolfgang Dahmen, and Ron DeVore for teaching him compressed sensing. Thanks are also due to Rachel Ward for an interesting discussion about stability results. The presentation of this paper was greatly improved by numerous and most helpful remarks of anonymous referees to whom the author would like to express his sincere thanks. This research was made possible by EC Marie Curie ToK program SPADE-2 at IMPAN. It was partially supported by MNiSW grant no. N201 269335.
Rights and permissions
About this article
Cite this article
Wojtaszczyk, P. Stability and Instance Optimality for Gaussian Measurements in Compressed Sensing. Found Comput Math 10, 1–13 (2010). https://doi.org/10.1007/s10208-009-9046-4
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10208-009-9046-4
Keywords
- Signal recovery algorithms
- Restricted isometry condition
- Compressed sensing
- ℓ1-minimization
- Gaussian measurement matrix
- Signal recovery
- Sparse approximation