Search NASA⌕ Search

SEARCH · Search NASA

Results for “Deep Material Network”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Bayesian SegNet for Semantic Segmentation with Improved Interpretation of Microstructural Evolution During Irradiation of Materials

Understanding the relationship between the evolution of microstructures of irradiated LiAlO2pellets and tritium diffusion, retention and release could improve predictions of tritium performance. Given expert-labeled segmented images of irradiated and unirradiated pellets, we trained Deep Convolutional Neural Networks to segment images into defect, grain, and boundary classes. Qualitative microstructural information was calculated from these segmented images to facilitate the comparison of unirradiated and irradiated pellets. We tested modifications to improve the sensitivity of the model, including incorporating meta-data into the model and utilizing uncertainty quantification. The predicted segmentation was similar to the expert-labeled segmentation for most methods of microstructural qualification, including pixel proportion, defect area, and defect density. Overall, the high performance metrics for the best models for both irradiated and unirradiated images shows that utilizing neural network models is a viable alternative to expert-labeled images.

Oostrom, Marjolein T.↗
NASA NTRSDate not supplied

Variance Decomposition of MEDLI2 Reconstructed Heating Using Neural Networks

The Mars Entry, Descent, and Landing Instrumentation (MEDLI2) sensor suite collected data during entry of the Mars 2020 Perseverance rover into Mars’ atmosphere. This suite included a network of MEDLI2 Instrumented Sensor Plugs (MISPs). Each MISP was comprised of a cylinder made of Thermal Protection System (TPS) material with 1-3 embedded thermocouples (TCs), and it was flush mounted into the heatshield or backshell. Data from these in-depth TCs were used to reconstruct the aeroheating environment of the vehicle throughout entry. Surface heating was posed as an inverse problem, with the goal of estimating the surface heating by minimizing an objective function of the difference between MISP temperature measurements during flight and the temperature predictions derived from the Fully Implicit Ablation and Thermal response (FIAT) program. Given an aerothermal environment, FIAT calculates the material response and provides in-depth temperatures throughout the TPS material. To achieve the reverse, an internal tool called FIAT_Opt runs through multiple different environments until the output temperature at the TC depth closely matches the flight data. 95% confidence intervals on the reconstructed surface heating were obtained using Monte Carlo analysis, in which uncertainties in the thermocouple depth and the TPS material properties (e.g., density, thermal conductivity, heat capacity, emissivity) based on flight-lot material testing were included. A variance decomposition method using Sobol indices was employed to assess the sensitivity of the reconstructed peak heating to the TC placement and material property uncertainties. Variance decomposition was found to require tens of thousands of FIAT_Opt runs in order for the Sobol indices to converge. With a single FIAT_Opt run taking on the order of 40 minutes, the required number of computations would take months to complete, even if using multiple CPUs. To mitigate this problem, three machine learning models (ridge regression with cross-validation, random forest regression, and a deep neural network) were trained and tested using the 2000 Monte Carlo runs that were already completed. A subset of 1600 runs were used to train the model (i.e., training set), while the remaining 400 runs were used as the test set. The predictions from the deep neural network (DNN) on the test set showed nearly perfect agreement to the actual values computed with FIAT_Opt (R2 > 0.99). Using the DNN as a surrogate model, the variance decomposition using 50,000 runs was completed within minutes. The resulting Sobol indices showed that the reconstructed peak surface heating was most sensitive to the uncertainties in the thermal conductivity (ST = 0.37) and heat capacity (ST = 0.26). This method can be leveraged to provide requirements for material property measurements needed to improve the accuracy of surface heating prediction and ultimately lead to the reduction of design margins in the future. This presentation will include background on the MEDLI2 suite; the method used for inverse heating estimation; the way that material property uncertainties were accounted for using Monte Carlo analysis; a brief background on variance decomposition; the motivation for using machine learning in this context; how a neural network was trained on the data to enable variance decomposition in a fraction of the time; and the variance decomposition results for one of the MISPs.

Hannah Alpert↗
NASA NTRS2022

Microstructure Segmentation With Deep Learning Encoders Pre-Trained on a Large Microscopy Dataset

This study examined the improvement of microscopy segmentation intersection over union accuracy by transfer learning from a large dataset of microscopy images called MicroNet. Many neural network encoder architectures were trained on over 100,000 labeled microscopy images from 54 material classes. These pre-trained encoders were then embedded into multiple segmentation architectures including UNet and DeepLabV3+ to evaluate segmentation performance on created benchmark microscopy datasets. Compared to ImageNet pre-training, models pre-trained on MicroNet generalized better to out-of-distribution micrographs taken under different imaging and sample conditions and were more accurate with less training data. When training with only a single Ni-superalloy image, pre-training on MicroNet produced a 72.2% reduction in relative intersection over union error. These results suggest that transfer learning from large in-domain datasets generate models with learned feature representations that are more useful for downstream tasks and will likely improve any microscopy image analysis technique that can leverage pre-trained encoders.

machine learning↗
DOE OSTI2025

Atomic‐Scale Mechanisms of Nucleation and Stabilization in CuCrO 2 and CuFeO 2 Delafossite Thin Films on Al 2 O 3

Abstract Delafossite thin films exhibit a range of intriguing physical properties derived from their layered structure, which consists of alternating noble metal ( A + ) and ( B O 2 − ) sublayers in an AB O 2 stoichiometry. The integration of these properties into functional devices requires the successful epitaxial growth of delafossites as thin films on appropriate substrates. Unfortunately, their unique lattice geometry complicates growth, as different delafossites display variable behavior on the same substrate, often unrelated to lattice mismatch. This suggests the presence of yet unidentified stabilization mechanisms that enable the selective growth of certain delafossites, allowing them to be grown either as films themselves or to be used as buffer layers for subsequent deposition. In this study, advanced scanning transmission electron microscopy (STEM) is employed to investigate the nucleation mechanisms governing the stable growth of Cu‐based delafossites on Al 2 O 3 , specifically CuCrO 2 and CuFeO 2 thin films synthesized via molecular‐beam epitaxy. These findings reveal that a combination of a misfit dislocation network at the interface and monolayer‐deep chemical intermixing effectively relieves lattice mismatch strain. This mechanism, which differs fundamentally from that observed in Pd‐based delafossites, provides key insights into the controlled epitaxy of delafossite materials.

Scheid, Anna [Max Planck Institute for Solid State↗
DOE OSTI2026

Machine learning surrogates for ion energy–angle distributions in thermal and RF plasma sheaths

Ion energy–angle distributions (IEADs) at material surfaces are a critical input for plasma–material interaction (PMI) studies in fusion devices, yet they are computationally expensive to obtain using particle-in-cell (PIC) simulations. In this work, we develop a machine learning surrogate based on a deep deconvolutional neural network (DDeCNN) trained on large databases generated with the hPIC2 code. The surrogate is capable of reconstructing IEADs from sheath parameters for both thermal and radio-frequency (RF) plasmas, including cases with multiple ion species. Across thousands of test cases, the model achieves high accuracy, with over 97 % of predictions classified as good or average based on standard error metrics (MAE, MSE, L2). Even in the more challenging RF and multi-species regimes, the surrogate reliably captures the multi-peak structure of PIC results. Once trained, the surrogate produces IEADs in milliseconds on a common workstation, yielding speedups of six to seven orders of magnitude compared with running a full PIC simulation. This computational gain enables dense parameter scans and direct coupling of IEAD predictions with PMI and erosion models on whole-device scales in fusion-relevant conditions.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

Learning nuclear cross sections across the chart of nuclides with graph neural networks

We explore the use of deep learning techniques to learn how nuclear cross sections change as we add or remove protons and neutrons. As a proof of principle, we focus on the neutron-induced reactions in the fast energy regime. Our approach follows a two-stage learning framework. First, we apply representation learning to encode cross section data into a latent space using either variational autoencoders (VAEs) or implicit neural representations (INRs). Then, we train graph neural networks (GNNs) on the resulting embeddings to predict missing values across the nuclear chart by leveraging the topological structure of neighboring isotopes. We demonstrate accurate cross section predictions within a 9 × 9 block of missing nuclei. We also find that the optimal GNN training strategy depends on the type of latent representation used, with VAE embeddings performing best under end-to-end optimization in the original space, while INR embeddings achieve better results when the GNN is trained only in the latent space. Furthermore, using clustering algorithms, we map groups of latent vectors into regions of the nuclear chart and show that VAEs and INRs can discover some of the neutron magic numbers. These findings suggest that deep-learning models based on the representation encoding of cross sections combined with graph neural networks hold significant potential in augmenting nuclear theory models, e.g., by providing reliable estimates of covariances of cross sections, including cross-material covariances.

Machine learning↗
DOE OSTI2026

Federated learning for 2D synchrotron x-ray diffractometry: a cross-institutional approach for phase quantification of Ti–6Al–4V alloy

High-energy Two dimensional (2D) synchrotron x-ray diffractometry provides important insights into the atomistic structure and phase evolution of materials, yet traditional analysis methods remain complex, knowledge-intensive, and computationally demanding. Deep-learning models offer a powerful alternative for automating their analysis. Institutions that hold these datasets may be unwilling to share their data due to privacy and security policies, as well as the challenges associated with large-scale data transfer. As a result, models trained on local datasets often perform well only on their own data but exhibit bias and poor generalization across different instruments or facilities. To overcome these limitations, we explore federated learning (FL) for 2D synchrotron diffractograms, enabling collaborative model training without exchanging raw data. In this study, 2D synchrotron diffractograms of Ti–6Al–4V alloy collected from two independent facilities are used to train convolutional neural networks for predicting the β-phase volume fraction. Experimental results show that federated global models significantly outperform locally trained models in terms of generalization and achieve accuracy comparable to centralized trained models. These findings demonstrate the potential of FL to enable secure, cross-institutional collaboration and enhance the scalability of deep-learning-based materials characterization.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Applications of Nickelate perovskites for neuromorphic computing from electronic structure and Machine Learning

While the limit of Moore's law is presently being reached with current microelectronic technologies, we need to develop new paradigms that overcome this limitation. In that respect, neuromorphic computing is a concept that emulates the neural behavior and response of the human brain, and it has been recognized as a promising alternative approach. In this research project, we will perform multi-fidelity scale bridging to explore the potential use of materials with metal to insulator transition for neuromorphic applications. In particular, rare earth nickelates are promising for such purposes, as the transition in these materials is quite sensitive to a broad set of different external stimuli. Our multi-fidelity approach will bridge the high-fidelity electronic structure calculations with classical potentials. We will bridge dynamical mean field theory with a classical atomistic representation via a deep learning force field. The neural network is trained with energies, charges, and forces obtained by accurate electronic structure theories based on Dynamical Mean Field Theory. The configurational space is generated from known crystal phases, ab initio molecular dynamics with exchange-correlation functionals corrected with the Hubbard model, disordered phases with different concentrations of oxygen vacancies, and nonsymmetrical positions and induced strain by grain interfaces or contact with a substrate. Strategies to train the model with a reduced number of training examples are obtained from active learning methods, and new structures for improving the learning process are generated by using machine learning autoencoders. This classical potential will be validated through a diversity of electronic structure methods and represents an important step to combine the flexibility and accuracy of first-principles with the speed of classical potentials. The generated multi-fidelity surrogate model will be used to understand the role of strain, oxygen vacancies, proton doping, the variation of the crystal phase, substrate effects, vibrational effects as the octahedral rotation, grain boundaries and defect effects on the response of a Metal to Insulator Transition (MIT) in correlated materials. Long time and large-scale simulations will help understand the role of different stimuli to control the hysteresis of the MIT, as it has been experimentally suggested. Selected configurations will be analyzed with higher-level theories to provide an accurate electronic description and to study how the orbitals and charges are rearranged under different conditions.

36 MATERIALS SCIENCE↗
NASA NTRS2008

Fast Electromechanical Switches Based on Carbon Nanotubes

Electrostatically actuated nanoelectromechanical switches based on carbon nanotubes have been fabricated and tested in a continuing effort to develop high-speed switches for a variety of stationary and portable electronic equipment. As explained below, these devices offer advantages over electrostatically actuated microelectromechanical switches, which, heretofore, have represented the state of the art of rapid, highly miniaturized electromechanical switches. Potential applications for these devices include computer memories, cellular telephones, communication networks, scientific instrumentation, and general radiation-hard electronic equipment. A representative device of the present type includes a single-wall carbon nanotube suspended over a trench about 130 nm wide and 20 nm deep in an electrically insulating material. The ends of the carbon nanotube are connected to metal electrodes, denoted the source and drain electrodes. At bottom of the trench is another metal electrode, denoted the pull electrode (see figure). In the off or open switch state, no voltage is applied, and the nanotube remains out of contact with the pull electrode. When a sufficiently large electric potential (switching potential) is applied between the pull electrode and either or both of the source and drain electrodes, the resulting electrostatic attraction bends and stretches the nanotube into contact with the pull electrode, thereby putting the switch into the "on" or "closed" state, in which substantial current (typically as much as hundreds of nanoamperes) is conducted. Devices of this type for use in initial experiments were fabricated on a thermally oxidized Si wafer, onto which Nb was sputter-deposited for use as the pull-electrode layer. Nb was chosen because its refractory nature would enable it to withstand the chemical and thermal conditions to be subsequently imposed for growing carbon nanotubes. A 200- nm-thick layer of SiO2 was formed on top of the Nb layer by plasma-enhanced chemical vapor deposition. In the device regions, the SiO2 layer was patterned to thin it to the 20-nm trench depth. The trenches were then patterned by electron- beam lithography and formed by reactive- ion etching of the pattern through the 20-nm-thick SiO2 to the Nb layer.

Kaul, Anupama↗
DOE OSTI2026

User Manual - HydraGNN v5.0: Distributed Implementation of Multi-Tasking Graph Neural Networks

This document serves as the user manual for HydraGNN v5.0, a scalable graph neural network (GNN) architecture for simultaneous prediction of multiple target properties using multi-task learning (MTL). This version of HydraGNN has been developed primarily to support the development, training, and deployment of predictive graph-based deep learning (DL) models for atomistic materials modeling. HydraGNN is templated over 13 message-passing policies, including invariant models (GIN, PNA, PNAPlus, GAT, MFC, CGCNN, SAGE, SchNet, DimeNet) and equivariant models (EGNN, PNAEq, PAINN, MACE), and supports distributed training via distributed data parallelism (DDP), DeepSpeed, and Fully Sharded Data Parallelism (FSDP) on leadership-class supercomputers. Although HydraGNN can be applied to problems beyond atomistic materials modeling, its current use is confined to homogeneous graphs. Additional capabilities include machine-learned interatomic potentials with energy-conserving forces, General, Powerful, and Scalable Graph Transformer (GraphGPS) global attention, periodic boundary conditions, hyperparameter optimization, mixed-precision training, and uncertainty quantification.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Large-scale deep learning for metastasis detection in pathology reports

Objectives No existing algorithm can reliably identify metastasis from pathology reports across multiple cancer types and the entire US population. In this study, we develop a deep learning model that automatically detects patients with metastatic cancer by using pathology reports from many laboratories and of multiple cancer types. Materials and Methods We use 60 471 unstructured pathology reports from 4 Surveillance, Epidemiology, and End Results (SEER) registries. The reports were coded into 1 of 3 labels: metastasis negative, metastases positive, or metastasis undetermined. We utilize a task-specific deep neural network trained from scratch and compare its performance with a widely used large language model (LLM). Results Our deep learning architecture trained on task-specific data outperforms a general-purpose LLM, with a recall of 0.894 compared to 0.824. We quantified model uncertainty and used it to defer reports for human review. We found that retaining 72.9% of reports increased recall from 0.894 to 0.969. Discussion A smaller deep learning architecture trained on task-specific data outperforms a general LLM. Equally critical to model performance is the incorporation of uncertainty quantification, achieved here through an abstention mechanism. Conclusions This study’s finding demonstrate the feasibility of developing algorithms to automatically identify metastatic cancer cases from unstructured pathology reports.

machine learning↗
DOE OSTI2025

Deep Learning with Reflection High-Energy Electron Diffraction Images to Predict Cation Ratio in Sr 2 x Ti 2(1– x ) O 3 Thin Films

Machine learning (ML) with in-situ diagnostics offers a transformative approach to accelerate, understand, and control thin film synthesis by uncovering relationships between synthesis conditions and material properties. In this study, we demonstrate the application of deep learning to predict the stoichiometry of Sr 2x Ti 2(1–x) O 3 thin films using reflection high-energy electron diffraction images acquired during pulsed laser deposition. A gated convolutional neural network trained for regression of the Sr atomic fraction achieved accurate predictions with a small dataset of 31 samples. Explainable AI techniques revealed a previously unknown correlation between diffraction streak features and cation stoichiometry in Sr 2x Ti 2(1–x) O 3 thin films. Here, our results demonstrate how ML can be used to transform a ubiquitous in-situ diagnostic tool, that is usually limited to qualitative assessments, into a quantitative surrogate measurement of continuously valued thin film properties. Such methods are critically needed to enable real-time control, autonomous workflows, and accelerate traditional synthesis approaches.

77 NANOSCIENCE AND NANOTECHNOLOGY↗
DOE OSTI2026

OPEN ALPHADIFFRACT

Open-source release of the AlphaDiffract data generation and training system. Includes only the public Materials Project dataset retrievers.AlphaDiffract is a deep learning framework that achieves state-of-the-art performance in predicting the crystal system, space group, and lattice parameters directly from PXRD patterns. AlphaDiffract utilizes a 1D adaptation of the ConvNeXt architecture, a modern convolutional neural network that integrates key design principles from transformers, coupledwith dedicated prediction heads for each crystallographic property.

Prince, Michael [Argonne National Laboratory (ANL)↗
DOE OSTI2025

Deep-learning atomistic semi-empirical pseudopotential model for nanomaterials

The semi-empirical pseudopotential method (SEPM) has been widely applied to provide computational insights into the electronic structure, photophysics, and charge carrier dynamics of nanoscale materials. We present “DeepPseudopot”, a machine-learned atomistic pseudopotential model that extends the SEPM framework by combining a flexible neural network representation of the local pseudopotential with parameterized non-local and spin-orbit coupling terms. Trained on bulk quasiparticle band structures and deformation potentials from GW calculations, the model captures many-body and relativistic effects with very high accuracy across diverse semiconducting materials, as illustrated for silicon and group III-V semiconductors. DeepPseudopot’s accuracy, efficiency, and transferability make it well-suited for data-driven in silico design and discovery of novel optoelectronic nanomaterials.

Lin, Kailai [University of California, Berkeley, C↗
NASA NTRS1997

The Solar-Stellar Connection (NAG5-6124: SOHO Guest Investigator Program)

The following is a final report from the SOHO Guest Investigator program to use the SUMER far-UV spectrometer to obtain imaging spectroscopy in support of the goals of the so-called "solar-stellar connection." In particular, a major emphasis was utilization of the long-slit time-resolved maps of the solar surface in bright far-UV emission lines to deduce how particular aspects of the temporally and spatially averaged line profiles trace back to individual structural features of the magnetically disturbed outer atmosphere; to help interpret the unresolved line profiles from high quality stellar observations (say, with the Space Telescope Imaging Spectrograph). The researchers served two tours of duty in the SOHO Operations Center as SUMER planners, during which time we conducted an extensive series of observing programs. These can be divided into three general categories: surface mapping, translimb spectroscopy, and active region diagnostics. We have analyzed some of the large volumes of data to the point where we have presented them in poster papers, and in invited papers at national and international meetings. Listed below are the titles of the preliminary publications we have written, including brief abstracts to indicate the main results. (1) Chromospheric structure and Dynamics-- Observations -- The chromosphere is a highly structured dynamic 'layer' of the solar outer atmosphere. Here, not only are the effects of mechanical heating first evident (moving upward in altitude from the deep photosphere), but also the amount of nonradiative energy deposited is far greater than in the albeit much hotter overlying transition region and corona. Further, the chromosphere is by far the thickest zone of the solar atmosphere with respect to the pressure scale height. A major goal of stellar astrophysics is to understand how the chromosphere is heated and why it adopts its peculiar structure. A cursory examination of solar filtergrams and high-resolution movies demonstrates that much of the chromospheric "action" must be occurring on fine spatial scales and short times; particularly in the cell interior transient brightenings, but also in the longer-lived network fragments. That regime of investigation is far removed from what one usually associates with "synoptic" measurements. Nevertheless, synoptic observations of chromospheric indices, filtergrams, and globally-averaged profile parameters not only can provide important insight concerning the crucial role of the cycle-variable part of the solar magnetic field; but they also can forge a key link with analogous measurements of the stars, where often the phenomena can be significantly exaggerated from the solar case, but high spatial resolution reconnaissance is not even a remote possibility. In addition to discussing the synoptic aspects of chromospheric structure and dynamics, I summarized new insights into the general problem of the solar chromosphere that have been obtained recently with the SUMER far-ultraviolet spectrometer on SOHO. (2) CO and the Temperature Structure of the Solar Atmosphere -- The surface layers of the Sun provide a crucial boundary condition for many of the processes that occur in the deep interior. The stratification of the outer solar atmosphere once was thought to be well understood. However, studies of thermally sensitive molecular absorptions in the infrared revealed puzzling anomalies. Strong lines of the CO fundamental vibration-rotation bands near 5 microns showed very cool temperatures at the extreme limb, and remarkable off-limb emissions extending well into the supposedly hot chromosphere. The conflicting pictures of the photosphere/chromosphere interface, from the widely separated wavelength regimes, has raised suspicions that those "layers" of the atmosphere are much more inhomogeneous than previously suspected. One proposal is that the low chromosphere is dominated by cool gas, the "COmosphere," which is threaded by a network of persistent small-scale hot magnetic filaments and occasionally disrupted by localized acoustic disturbances. The COmosphere is capped by the merged fields of the network elements in the chromospheric "canopy." (3) Translimb Spectroscopy with SOHO/SUMER -- We have used SUMER to obtain deep exposures of the 1300-1400 A spectrum, at the extreme limb and off-limb. Previous "translimb" studies in the thermal infrared had revealed remarkable extensions of cold material (T is approximately equal to 3000 K)-traced by carbon monoxide emission lines-into the heart of the hot chromosphere. A main objective of our program was to search for corresponding far-UV signatures of the "thermally-bifurcated" low chromosphere; for example, radiatively fluoresced emissions of the CO A-X 4th-positive system (collisional excitation would be negligible in cold gas). We conducted two separate observing programs with SUMER. Both made use of the 1 inch-diameter circular aperture, translated across the limb in the minimum motor step increments of 0."375, along the central meridian in the Northern polar coronal hole. The first program executed for nine hours beginning 19UT 25 Oct 1996. The full wavelength range was 1340-1400 A. It was recorded in two overlapping segments, placing key regions of the spectrum alternately on the KBr and bare parts of the detector, to help isolate 2nd-order features. Each segment was integrated for 500 s, and 32 pairs were obtained to span a 12" swath centered on the optical limb. The second program was conducted 00-09UT 01 Dec 1996. It consisted of a single wavelength setting (1300-1340 A) with exposure time 500 s, but twice the spatial coverage of the earlier series: 64 steps, for a total displacement of 24". The strong chromospheric resonance lines of atomic oxygen (1302-1305 A) and ionized carbon (1334-1335 A) were observed on the bare part of the MCP camera.

Ayres, Thomas R.↗
DOE OSTI2025

Impact of Domain Knowledge on the Property Prediction of Specialized Machine Learning Models

Developing transferable machine learning models is trending in data-driven materials research. However, how to apply such models to a specific research domain remains unclear. Here, in this work, we choose high-entropy materials as a platform with a specialized data set containing 145,323 DFT-relaxed materials. This data set is used to explore the role of domain-specific knowledge in training effective models. Our tests with three representative graph neural network architectures indicate the model complexity has much smaller influence on performance than the data itself. Specifically, the consideration of low-energy atomic ordering, structures with diverse elemental coverage, and high-order interactions significantly influences the model performance. We also find that domain knowledge-driven sampling can greatly enhance unsupervised learning techniques. This research highlights that developing specialized data sets is more beneficial than further complicating deep learning architectures. Additionally, physics-inspired sampling algorithms are crucially needed for better machine learning models for a specific materials research domain.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Sparse-Data Deep Learning Strategies for Radiographic Non-Destructive Testing

Radiography is an imaging technique used in a variety of applications, such as medical diagnosis, airport security, and nondestructive testing. We present a deep learning system for extracting information from radiographic images. We perform various prediction tasks using our system, including material classification and regression on the dimensions of a given object that is being radiographed. Our system is designed to address the sparse-data issue for radiographic nondestructive testing applications. It uses a radiographic simulation tool for synthetic data augmentation, and it uses transfer learning with a pre-trained convolutional neural network model. Using this system, our preliminary results indicate that the object geometry regression task saw an improvement of 70% in the R-squared value when using a multi-regime model. In addition, we increase the performance of the object material classification tasks by utilizing data from different imaging systems. In particular, using neutron imaging improved the material classification accuracy by 20% when compared to x-ray imaging.

convolutional neural networks↗
DOE OSTI2026

High-speed quantitative X-ray multi-contrast imaging with deep learning based modulated pattern analysis

The advent of X-ray multi-contrast imaging methods, providing absorption, phase, and dark-field images, holds tremendous promise for complementary and non-destructive visualization of inner structures within materials and bio-samples. However, the low efficiency in measuring and analyzing X-ray modulated patterns has hindered their application in high-resolution in situ imaging. In this work, the Enhanced Scanning Pattern-based Imaging Neural Network (ESPINNet) is introduced as a powerful tool for achieving high-speed, high-resolution quantitative imaging. ESPINNet is faster than correlation-based speckle tracking methods such as XSVT and UMPA, and provides a balanced performance in terms of resolution and speed for data collection by using fewer scanning images. In comparison with our previously developed neural network, ESPINNet introduces the capability to generate dark-field images, further enhancing its versatility. By leveraging scanning patterns, ESPINNet significantly improves resolution and measurement precision. Furthermore, its adaptability to various modulation patterns, including those produced by sandpaper, coded masks, or gratings, ensures broad applicability. These features enable real-time 2D and 3D multi-contrast imaging, positioning ESPINNet as a transformative solution for applications in materials science and biomedical research, particularly for high-speed and in situ measurements.

X-ray at-wavelength metrology↗