Data-driven Simulations For Training AI-Based Segmentation of Neutron Images

Published

March 19, 2024

Author(s)

Pushkar Sathe, Caitlyn M. Wolf, Youngju Kim, Sarah M. Robinson, Michael Daugherty, Ryan Murphy, Jacob LaManna, Michael Huber, David Jacobson, Paul A. Kienzle, Kathleen Weigandt, Nikolai Klimov, Daniel Hussey, Peter Bajcsy

Abstract

Neutron interferometry is unique in its ability to measure atomic scale properties of materials that no other imaging modality can. However, building, operating, and using such neutron imaging instruments poses constraints on the acquisition time and on the number of measured images per sample. Experiment time-constraints yield small quantities of measured images that are insufficient for automating image analyses using supervised artificial intelligence (AI) models. One approach alleviates this problem by supplementing annotated measured images with synthetic images. To this end, we create a data-driven simulation framework that supplements training data beyond typical data-driven augmentations by leveraging statistical intensity models, such as the Johnson family of probability density functions (PDFs). We follow the simulation framework steps for an image segmentation task including Estimate PDFs $\,\to\,$ Validate PDFs $\,\to\,$ Design Image Masks $\,\to\,$ Generate Intensities $\,\to\,$ Train AI Model for Segmentation. Our goal is to minimize the manual labor to execute the steps and maximize our confidence in simulations and segmentation accuracy. We report results for a set of nine known materials (calibration phantoms) that were imaged using a neutron interferometry-based instrument acquiring four-dimensional images and segmented by AI models trained with synthetic and measured images and their masks.

Citation

Scientific Reports

Volume

14

Pub Type

Journals

Download Paper

https://doi.org/10.1038/s41598-024-56409-3

Local Download

Keywords

image segmentation, neutron image simulation, neural networks

Citation

Sathe, P. , Wolf, C. , Kim, Y. , Robinson, S. , Daugherty, M. , Murphy, R. , LaManna, J. , Huber, M. , Jacobson, D. , Kienzle, P. , Weigandt, K. , Klimov, N. , Hussey, D. and Bajcsy, P. (2024), Data-driven Simulations For Training AI-Based Segmentation of Neutron Images, Scientific Reports, [online], https://doi.org/10.1038/s41598-024-56409-3, https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=956578 (Accessed March 6, 2026)