Bayesian Cram\'{e}r-Rao Bound for Noisy Non-Blind and Blind Compressed Sensing
Information Theory
2010-05-25 v1 math.IT
Abstract
In this paper, we address the theoretical limitations in reconstructing sparse signals (in a known complete basis) using compressed sensing framework. We also divide the CS to non-blind and blind cases. Then, we compute the Bayesian Cramer-Rao bound for estimating the sparse coefficients while the measurement matrix elements are independent zero mean random variables. Simulation results show a large gap between the lower bound and the performance of the practical algorithms when the number of measurements are low.
Cite
@article{arxiv.1005.4316,
title = {Bayesian Cram\'{e}r-Rao Bound for Noisy Non-Blind and Blind Compressed Sensing},
author = {Hadi Zayyani and Massoud Babaie-Zadeh and Christian Jutten},
journal= {arXiv preprint arXiv:1005.4316},
year = {2010}
}
Comments
This paper was submitted at 2 June 2009 to IEEE Signal Processing Letters and was rejected at 21 August 2009