Abstract

We present an information-theoretic framework for measurement basis design in compressive imaging. Simulation results show that the reconstruction error obtained with information-optimal projections is nearly an order of magnitude lower than that for random projections.

© 2012 OSA

PDF Article