Synthetic Aperture Imaging with Neural Signal Representations

Synthetic aperture imaging is a powerful computational imaging technique wherein a collection of low resolution measurements are computationally integrated into a single high-resolution image. This fusion step has been traditionally accomplished with pixel-based scene representations. This manuscript provides an overview of neural signal representations and describes how they can be used to enhance the synthetic aperture reconstruction pipeline.

1. Introduction

Almost all synthetic aperture imaging systems can be modeled by the equations

$$\begin{equation*} y_{i}=\mathcal{A}_{\theta_{i}}(x)+n_{i},\tag{1}\end{equation*}$$