Infoscience

Conference paper

# Recipes on Hard Thresholding Methods

Compressive sensing (CS) is a data acquisition and recovery technique for finding sparse solutions to linear inverse problems from sub-Nyquist measurements. CS features a wide range of computationally efficient and robust signal recovery methods, based on sparsity seeking optimization. In this paper, we present and analyze a class of sparse recovery algorithms, known as hard thresholding methods. We provide optimal strategies on how to set up these algorithms via basic ingredients'' for different configurations to achieve complexity vs. accuracy tradeoffs. Simulation results demonstrate notable performance improvements compared to state-of-the-art algorithms both in terms of data reconstruction and computational complexity.