Computing reconstructions from nonuniform Fourier samples: universality of stability barriers and stable sampling rates (Q1713634)

Let \(D \subset {\mathbb R}^d\) be a centered symmetric convex body, where a convex body is a compact convex set with non-empty interior. The authors reconstruct \(f\in L^2(D)\) from finitely many nonuniform samples of its Fourier transform taken on a closed countable subset \(\Omega \subset {\mathbb R}^d\) which fulfills certain gap condition. Such problems arise in various imaging applications, where Fourier samples are taken along radial lines or spirals. The authors apply an abstract theory of finite-dimensional approximation of continuous sampling problems presented by \textit{B. Adcock} et al. [SIAM J. Imaging Sci. 7, No. 3, 1690--1723 (2014; Zbl 1308.94045)]. They investigate the rate of convergence of approximate solutions and their numerical stability. It is shown that the proportion of Fourier samples that allow for stable approximations of a given accuracy is independent of the specific sampling geometry. For \(d=1\) and \(d=2\), the authors identify necessary and sufficient conditions for stable reconstruction of algebraic polynomials and wavelet coefficients from nonuniform Fourier data. Numerical experiments for stable recovery of wavelet coefficients from nonuniform Fourier samples are presented in [\textit{M. Gataric} and \textit{C. Poon}, SIAM J. Sci. Comput. 38, No. 2, A1075--A1099 (2016; Zbl 1343.65151)].