Abstract
Concerning medical images, which are known to have sparsity in either the spatial (or its derivative), DFT, DCT or curvelet domain, we propose a new approach for reconstruction from sparse samples, based on Simultaneous Perturbation Stochastic Optimization (SPSA) to minimize a nonconvex l(p)-norm for 0 < p < 1. The value of p chosen is such as to achieve as close an approximation to l(0)-norm as is computationally feasible. This approach is distinct from the homotopy-theoretic and hard-thresholding techniques of recent literature for l(0)- and l(p)-norm minimization. For lack of space, our illustrations are limited to only one each of synthetic and real images.