Model-adapted Fourier sampling for generative compressed sensing

TL;DR

This paper introduces a model-adapted Fourier sampling method for generative compressed sensing, reducing measurement requirements by optimizing sampling strategies based on neural network signal models.

Contribution

It develops a new theoretical framework for nonuniform sampling and proposes an optimized sampling distribution to improve measurement efficiency in generative compressed sensing.

Findings

Reduced sample complexity from O(kdn||α||∞^2) to O(kd||α||2^2)

Theoretical guarantees for nonuniform sampling distributions

Validated performance improvements on CelebA dataset

Abstract

We study generative compressed sensing when the measurement matrix is randomly subsampled from a unitary matrix (with the DFT as an important special case). It was recently shown that $\textit{O}(kdn\| \boldsymbol{\alpha}\|_{\infty}^{2})$ uniformly random Fourier measurements are sufficient to recover signals in the range of a neural network $G:\mathbb{R}^k \to \mathbb{R}^n$ of depth $d$, where each component of the so-called local coherence vector $\boldsymbol{\alpha}$ quantifies the alignment of a corresponding Fourier vector with the range of $G$. We construct a model-adapted sampling strategy with an improved sample complexity of $\textit{O}(kd\| \boldsymbol{\alpha}\|_{2}^{2})$ measurements. This is enabled by: (1) new theoretical recovery guarantees that we develop for nonuniformly random sampling distributions and then (2) optimizing the sampling distribution to minimize the number of measurements needed for these guarantees. This development offers a sample complexity applicable to natural signal classes, which are often almost maximally coherent with low Fourier frequencies. Finally, we consider a surrogate sampling scheme, and validate its performance in recovery experiments using the CelebA dataset.