Morphological variations to a ptychographic algorithm

TL;DR

This paper investigates how different illumination shapes, including convex and unconnected patterns like QR codes, affect the performance of ptychographic imaging algorithms, finding unconnected shapes often improve convergence and accuracy.

Contribution

It introduces a numerical study comparing the effects of various illumination shapes on ptychographic reconstruction performance.

Findings

Unconnected shapes outperform convex shapes in convergence.

Unconnected shapes can improve accuracy in some cases.

Shape choice impacts ptychographic algorithm efficiency.

Abstract

Ptychography is a technique widely used in microscopy for achieving high-resolution imaging. This method relies on computational processing of images gathered from diffraction patterns produced by several partial illuminations of a sample. In this work, we numerically studied the effect of using different shapes for illuminating the aforementioned sample: convex shapes, such as circles and regular polygons, and unconnected shapes that resemble a QR code. Our results suggest that the use of unconnected shapes seems to outperform convex shapes in terms of convergence and, in some cases, accuracy.