Imaging

AI

AI technologies applied by Argonne National Laboratory are enabling cutting-edge imaging modalities that will help solve some of the most pressing scientific challenges of our era.

DOE-supported experimental facilities at Argonne, including the Advanced Photon Source (APS), provide access to world-class imaging and physical characterization capabilities for more than 14,000 visiting scientists and engineers annually. The advanced imaging tools and instruments at these facilities probe complex materials and processes, including those underpinning energy technologies and advanced manufacturing. Planned developments, such as the revolutionized high-energy synchrotron light source proposed by the APS Upgrade project that will increase brightness and coherence, will lead to multiple order-of-magnitude increases in data rates and experiment complexity. Because data of this scale is too large and complex for humans to analyze, work is underway to apply and develop advanced machine learning algorithms that can process data at this scale. AI technologies applied to imaging for DOE science will help solve some of our most pressing scientific challenges across the physical, materials, environmental, and life sciences.

MLExchange - A Collaborative Machine Learning Platform for Scientiﬁc Discovery

MLExchange is a software framework designed to facilitate collaboration among researchers, provide ML solutions, and reduce the barriers to use ML tools.

Photon Sciences

Applying Computer Vision to Simplify Use of X-Ray Microscopes

Argonne’s 3D X-ray microscope employs computer vision to detect features of interest and capture and visualize data.

Computing, Environment and Life Sciences

Calibrating Tomography Data Using Machine Learning

An Argonne team has developed a machine learning approach for calibrating the center of rotation in x-ray light source tomography data that provides better accuracy than conventional imaging processing-based methods.

Photon Sciences

A Sea Slug Provides Inspiration to Advance AI

From studying a sea slug, researchers have demonstrated a fundamental type of learning in an inorganic system that may serve as a building block for neuromorphic computing and AI applications.

Photon Sciences

AI-accelerated Nanoscale Imaging

Deep learning enables real-time 3D nanoscale imaging with X-rays.

Photon Sciences

Machine-learning Determines Interatomic Potentials from X-ray Diffraction Data

A new machine learning technique that uses data from high-energy X-ray diffraction experiments will significantly reduce model development time and human effort.

Photon Sciences

Machine-learning Predicts Potentials for Molten Salts from X-ray Pair Distribution Function (PDF) Data

A new machine learning approach significantly lowers the barrier to understanding and designing molten salts across the periodic table.

Photon Sciences

TomoGAN: Deep Learning for Robust X-ray Tomography

Scientists have developed a generative adversarial network-based deep learning model for robust tomographic imaging.

Photon Sciences

Low-Dose X-ray Tomography with Self-supervised Learning

Scientists employ computational techniques and machine learning to enhance computational tomography by using low-dose sampling

Photon Sciences

AI Identifies Spurious Data in X-Ray Diffraction

At the Advanced Photon Source, scientists developed an artificial intelligence approach that replaces human intuition in identifying spurious data in X-ray diffraction patterns.

Photon Sciences