Physics-Inspired Unsupervised Classification for Region of Interest in X-Ray Ptychography

TL;DR

This paper introduces a physics-inspired unsupervised learning method to efficiently identify the region of interest in X-ray ptychography data, reducing computational load by filtering out irrelevant diffraction patterns before reconstruction.

Contribution

The work presents a novel unsupervised algorithm that leverages physics principles to automatically detect the RoI in ptychography data, improving efficiency without sacrificing image quality.

Findings

Effective filtering of non-RoI diffraction patterns

Reduced computational cost for ptychography reconstruction

Maintained high image quality after filtering

Abstract

X-ray ptychography allows for large fields to be imaged at high resolution at the cost of additional computational expense due to the large volume of data. Given limited information regarding the object, the acquired data often has an excessive amount of information that is outside the region of interest (RoI). In this work we propose a physics-inspired unsupervised learning algorithm to identify the RoI of an object using only diffraction patterns from a ptychography dataset before committing computational resources to reconstruction. Obtained diffraction patterns that are automatically identified as not within the RoI are filtered out, allowing efficient reconstruction by focusing only on important data within the RoI while preserving image quality.