Output details

Iterative hard thresholding for compressed sensing

Significance of output:

In many applications signal and image acquisition can be expensive, time-consuming, physically challenging and dangerous, e.g. medical x-ray scanners expose patients to dangerous radiation. This paper presents techniques aimed at making these acquisitions significantly safer, cheaper, faster and more reliable. We show that a simple and computationally efficient algorithm developed by the author guarantees optimal performance under a wide range of applications. A variation of this algorithm was recently shown to significantly speed up a Magnetic Resonance based Brain Imaging method (patent US application number: 61/808696). This paper is the journal's second most cited paper in each year after publication.