AI ENHANCED TECHNIQUE FOR HIGH QUALITY RECONSTRUCTION
SF-25-118 Aether is a Python package that offers AI-enhanced methods for high-quality, artifact-reduced ptychographic image reconstruction.
Du, Ming [Argonne National Laboratory (ANL), Argon↗
SEARCH · Engineering Papers
Results for “Image quality”
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.
A co-registered in-situ and ex-situ dataset from wire arc additive manufacturing process
Recent progress in sensing techniques and data analytics tools have significantly accelerated the development of Wire Arc Additive Manufacturing (WAAM) systems. This data-centric approach emphasizes leveraging sensor data available throughout the production process to optimize performance. Integration of extensive data analysis provides opportunities for improving precision, reducing waste, and enhancing the quality of produced parts. This method relies on AI/ML models and optimization techniques, which are developed using the data collected from various sources, including in-situ sensors, ex-situ imaging, and manufacturing process parameters. The quality and diversity of this data, along with the alignment between different data streams (achieved through spatiotemporal registration) are critical for the successful development of AI/ML and optimization models. In this work, we present a spatiotemporally registered dataset generated during the WAAM process of deposition of a rectangular block. The dataset includes a comprehensive description of the deposition process, process parameters, welding characteristics and acoustic data collected in-situ, and X-Ray Computed Tomography data of the build.
42 ENGINEERING↗
A Co-Registered In-Situ and Ex-Situ Dataset of Electrical, Acoustic, and CT Characteristics from Wire Arc Additive Manufacturing Process
Recent progress in sensing techniques and data analytics tools have significantly accelerated the development of Wire Arc Additive Manufacturing (WAAM) systems. This data centric approach emphasizes leveraging available data throughout the production process to optimize performance. Integration of extensive data analysis provides the opportunity to improve precision, reduce waste, and enhance the quality of produced parts. This method relies on AI/ML models and optimization techniques, which are developed using the data collected from various sources, including in-situ sensors, ex-situ imaging, and manufacturing process parameters. The quality and diversity of this data, along with the alignment between different data streams (achieved through spatiotemporal registration) are critical for the successful development of AI/ML and optimization models. In this work, we present a spatiotemporally registered dataset generated during the WAAM process of deposition of a rectangular block. The dataset includes the comprehensive description of deposition process, process parameters, in-situ collected welding characteristics, acoustic data, and X-Ray Computed Tomography analysis data for the build. Dataset A Co-Registered In-Situ and Ex-Situ Dataset of Electrical, Acoustic, and CT Characteristics from Wire Arc Additive Manufacturing Process has arisen under UT-Battelle, LLC’s Prime Contract No. DE-AC05-00OR22725 with the U.S. Department of Energy (DOE) to manage and operate the Oak Ridge National Laboratory. UT-Battelle, LLC will not assert any rights under United States law or under the Prime Contract it has in the dataset against any user of the dataset, including any copyrights or patent rights. UT-Battelle, LLC requests that attribution to the dataset is provided as academically appropriate.
42 ENGINEERING↗
Imaging mechanism for hyperspectral scanning probe microscopy via Gaussian process modelling
We investigate the ability to reconstruct and derive spatial structure from sparsely sampled 3D piezoresponse force microcopy data, captured using the band-excitation (BE) technique, via Gaussian Process (GP) methods. Even for weakly informative priors, GP methods allow unambiguous determination of the characteristic length scales of the imaging process both in spatial and frequency domains. We further show that BE data set tends to be oversampled in the spatial domains, with ~30% of original data set sufficient for high-quality reconstruction, potentially enabling faster BE imaging. At the same time, reliable reconstruction along the frequency domain requires the resonance peak to be within the measured band. This behavior suggests the optimal strategy for the BE imaging on unknown samples. Finally, we discuss how GP can be used for automated experimentation in SPM, by combining GP regression with non-rectangular scans.
47 OTHER INSTRUMENTATION↗
RECON Label Quality Report
The final quality of any AI/ML system is directly related to the quality of the input data used to train the system. In this case, we are trying to build a reliable image classifier that can correctly identify electrical components in x-ray images. The classification confidence is directly related to the quality of the labels in the training data, which are used in developing the AI/ML classifier. Incorrect or incomplete labels can substantially hinder the performance of the system during the training process, as it tries to compensate for variations that should not exist. Image labels are entered by subject matter experts, and in general can be assumed to be correct. However, this is not a guarantee, so developing ways to measure label quality and help identify or reject bad labels is important, especially as the database continues to grow. Given the current size of the database, a full manual review of each component is not feasible. This report will highlight the current state of the “RECON” x-ray image database and summarize several recent developments to try to help ensure high quality labeling both now and in the future. Questions that we hope to answer with this development include: 1) Are there any components with incorrect labels? 2) Can we suggest labels for components that are marked “Unknown”? 3) What kind of overall confidence do we have in the quality of the existing labels? 4) What systems or procedures can we put in place to maximize label quality?
42 ENGINEERING↗
Positron emission tomography harmonization in the Alzheimer's Disease Neuroimaging Initiative: A scalable and rigorous approach to multisite amyloid and tau quantification
Abstract INTRODUCTION A key goal of the Alzheimer's Disease NeuroImaging Initiative (ADNI) positron emission tomography (PET) Core is to harmonize quantification of β‐amyloid (Aβ) and tau PET image data across multiple scanners and tracers. METHODS We developed an analysis pipeline (Berkeley PET Imaging Pipeline, B‐PIP) for ADNI Aβ and tau PET images and applied it to PET data from other multisite studies. Steps include image pre‐processing, refacing, magnetic resonance imaging (MRI)/PET co‐registration, visual quality control (QC), quantification of tracer uptake, and standardization of Aβ and tau standardized uptake value ratios (SUVrs) across tracers. RESULTS Measurements from 10,105 cross‐sectional and longitudinal Aβ and tau PET scans acquired in several studies between 2010 and 2024 can be processed, harmonized, and directly merged across tracers and cohorts. DISCUSSION The B‐PIP developed in ADNI is a scalable image harmonization approach used in several observational studies and clinical trials that facilitates rigorous Aβ and tau PET quantification and data sharing. Highlights Quantitative results from ADNI Aβ and tau PET data are generated using a rigorous, scalable image processing pipeline This pipeline has been applied to PET data from several other large, multisite studies and trials Quantitative outcomes are harmonizable across studies and are shared with the scientific community
Neurosciences & Neurology↗
Deep generative learning of magnetic frustration in artificial spin ice from magnetic force microscopy images
Increasingly large datasets of microscopic images with nanoscale resolution facilitate the development of machine learning methods to identify and analyze subtle physical phenomena embedded within the images. In this work, microscopic images of honeycomb lattice spin-ice samples serve as datasets from which we automate the calculation of net magnetic moments and directional orientations of spin-ice configurations. In the first stage of our workflow, machine learning models are trained to accurately predict magnetic moments and directions within spin-ice structures. Variational Autoencoders (VAEs), an emergent unsupervised deep learning technique, are employed to generate high-quality synthetic magnetic force microscopy (MFM) images and extract latent feature representations, thereby reducing experimental and segmentation errors. The second stage of proposed methodology enables precise identification and prediction of frustrated vertices and nanomagnetic segments, effectively correlating structural and functional aspects of microscopic images. This facilitates the design of optimized spin-ice configurations with controlled frustration patterns, enabling potential on-demand synthesis.
36 MATERIALS SCIENCE↗
Interlaboratory Study of Digital Volume Correlation Error Due to X-Ray Computed Tomography Equipment and Scan Parameters: an Update from the DVC Challenge
Background: The quality of Digital Volume Correlation (DVC) full-field displacement measurements depends directly on the characteristics of the X-ray Computed Tomography (XCT) equipment, and scan procedures used to acquire the tomographic images. Objective: In this work, we seek to experimentally study the effects of XCT equipment and tomographic scan procedures on the quality of these images for DVC analysis, and to survey the level of DVC error that may be achieved using standard XCT operating procedures. Methods: Six participants in an interlaboratory study acquired high-quality XCT scans of a syntactic foam before and after rigid body motion. The resulting images were correlated using commercial DVC software to quantify error sources due to random image noise, reconstruction artifacts, as well as systematic spatial or temporal distortion. Results: In the absence of rigid body motion, the standard deviation of the displacement measurements ranged from 0.012 to 0.043 voxels using a moderate subvolume size, indicating that subvoxel measurement resolution could readily be achieved with a variety of XCT equipment and scan recipes. Comparison of consecutive scans without rigid body motion showed transient dilatational displacement gradients due to self-heating of the X-ray source and/or thermal expansion of the foam. Evaluation of the scans after rigid body motion showed significant, machine-specific spatial distortion in the displacement fields of up to 0.5 voxels; new approaches to remove this error need to be developed. Conclusions: Analysis of the scan protocols used in the interlaboratory study, as well as a complementary parametric sensitivity study, showed that the DVC error was strongly influenced by the XCT equipment, but could be mitigated by adjusting the total scan duration.
47 OTHER INSTRUMENTATION↗
Computer processing through distance-based quality score method in geospatial-temporal semantic graphs
A computer-implemented method of improving processing of overhead image data by a processor using a distance-based quality score in a geospatial-temporal semantic graph. An allowable range for each attribute in the subgraph search template is defined. For each match in a comparison, attribute values of each match element are compared against the preferred range and the allowable range to compute a corresponding distance of each match attribute from the subgraph search template. A corresponding overall match quality score is determined for each match from the subgraph search template, wherein determining the corresponding overall match qualities is performed using a corresponding required quality score and a corresponding optional quality score. All corresponding overall match quality scores are sorted into an ordered list and then displayed.
97 MATHEMATICS AND COMPUTING↗
Solution processed high aspect ratio ultra-long vertically well-aligned ZnO nano scintillators for potential X-ray imaging applications
We report the photon (PL), electron (CL) and X-ray (XEL) induced luminescence characteristics of high aspect ratio ultra-long (~ 50 µm) ZnO nanorods (NRs) and discuss the potential for fast X-ray detection based on the consistent and efficient visible emission (~ 580 nm) from ZnO NRs. Nanostructured ZnO scintillators were rearranged to form a vertically well-aligned NR design in order to help light absorption and coupling resulting in luminescent and fast scintillation properties. The design of the nanorod array combines the key advantages of a low-cost growth technique together with environmentally friendly and widely available materials. A low temperature hydrothermal method was adopted to grow ZnO NRs in one cycle growth and their structural, optical and X-ray scintillation properties were investigated. The relatively short (~ 10 µm) ZnO NRs emitting in the near-band-edge region were found to be almost insensitive to X-rays. On the other hand, the higher XEL response of long ZnO NRs, which is a key parameter for evaluation of materials to be used as scintillators for high quality X-ray detection and imaging, along with a decay time response in the order of ns confirmed promising scintillation properties for fast and high-resolution X-ray detector applications.
47 OTHER INSTRUMENTATION↗
A multimodal and integrated approach to interrogate human kidney biopsies with rigor and reproducibility: guidelines from the Kidney Precision Medicine Project
Comprehensive and spatially mapped molecular atlases of organs at a cellular level are a critical resource to gain insights into pathogenic mechanisms and personalized therapies for diseases. The Kidney Precision Medicine Project (KPMP) is an endeavor to generate three-dimensional (3-D) molecular atlases of healthy and diseased kidney biopsies by using multiple state-of-the-art omics and imaging technologies across several institutions. Obtaining rigorous and reproducible results from disparate methods and at different sites to interrogate biomolecules at a single-cell level or in 3-D space is a significant challenge that can be a futile exercise if not well controlled. Here we describe a “follow the tissue” pipeline for generating a reliable and authentic single-cell/region 3-D molecular atlas of human adult kidney. Our approach emphasizes quality assurance, quality control, validation, and harmonization across different omics and imaging technologies from sample procurement, processing, storage, shipping to data generation, analysis, and sharing. We established benchmarks for quality control, rigor, reproducibility, and feasibility across multiple technologies through a pilot experiment using common source tissue that was processed and analyzed at different institutions and different technologies. A peer review system was established to critically review quality control measures and the reproducibility of data generated by each technology before their being approved to interrogate clinical biopsy specimens. The process established economizes the use of valuable biopsy tissue for multiomics and imaging analysis with stringent quality control to ensure rigor and reproducibility of results and serves as a model for precision medicine projects across laboratories, institutions and consortia.
59 BASIC BIOLOGICAL SCIENCES↗
Improving and Assessing the Quality of Uncertainty Quantification in Deep Learning
Deep learning (DL) models have enjoyed increased attention in recent years because of their powerful predictive capabilities. While many successes have been achieved, standard deep learning methods suffer from a lack of uncertainty quantification (UQ). While the development of methods for producing UQ from DL models is an active area of current research, little attention has been given to the quality of the UQ produced by such methods. In order to deploy DL models to high-consequence applications, high-quality UQ is necessary. This report details the research and development conducted as part of a Laboratory Directed Research and Development (LDRD) project at Sandia National Laboratories. The focus of this project is to develop a framework of methods and metrics for the principled assessment of UQ quality in DL models. This report presents an overview of UQ quality assessment in traditional statistical modeling and describes why this approach is difficult to apply in DL contexts. An assessment on relatively simple simulated data is presented to demonstrate that UQ quality can differ greatly between DL models trained on the same data. A method for simulating image data that can then be used for UQ quality assessment is described. A general method for simulating realistic data for the purpose of assessing a model’s UQ quality is also presented. A Bayesian uncertainty framework for understanding uncertainty and existing metrics is described. Research that came out of collaborations with two university partners are discussed along with a software toolkit that is currently being developed to implement the UQ quality assessment framework as well as serve as a general guide to incorporating UQ into DL applications.
97 MATHEMATICS AND COMPUTING↗
How Low Can You Go? Using Synthetic 3D Imagery to Drastically Reduce Real-World Training Data for Object Detection
Deep convolutional neural networks (DCNNs) currently provide state-of-the-art performance on image classification and object detection tasks, and there are many global security mission areas where such models could be extremely useful. Crucially, the success of these models is driven in large part by the widespread availability of high-quality open source data sets such as Image Net, Common Objects in Context (COCO), and KITTI, which contain millions of images with thousands of unique labels. However, global security relevant objects-of-interest can be difficult to obtain: relevant events are low frequency and high consequence; the content of relevant images is sensitive; and adversaries and proliferators seek to obscure their activities. For these cases where exemplar data is hard to come-by, even fine-tuning an existing model with available data can be effectively impossible. Recent work demonstrated that models can be trained using a combination of real-world and synthetic images generated from 3D representations; that such models can exceed the performance of models trained using real-world data alone; and that the generated images need not be perfectly realistic (Tremblay, et al., 2018). However, this approach still required hundreds to thousands of real-world images for training and fine tuning, which for sparse, global security-relevant datasets can be an unrealistic hurdle. In this research, we validate the performance and behavior of DCNN models as we drive the number of real-world images used for training object detection tasks down to a minimal set. We perform multiple experiments to identify the best approach to train DCNNs from an extremely small set of real-world images. In doing so, we: Develop state-of-the-art, parameterized 3D models based on real-world images and sample from their parameters to increase the variance in synthetic image training data; Use machine learning explainability techniques to highlight and correct through targeted training the biases that result from training using completely synthetic images; and Validate our results by comparing the performance of the models trained on synthetic data to one another, and to a control model created by fine-tuning an existing ImageNet-trained model with a limited number (hundreds) of real-world images.
97 MATHEMATICS AND COMPUTING↗
Dimorphic Mechanisms of Fragility in Diabetes Mellitus: the Role of Reduced Collagen Fibril Deformation
ABSTRACT Diabetes mellitus (DM) is an emerging metabolic disease, and the management of diabetic bone disease poses a serious challenge worldwide. Understanding the underlying mechanisms leading to high fracture risk in DM is hence of particular interest and urgently needed to allow for diagnosis and treatment optimization. In a case–control postmortem study, the whole 12th thoracic vertebra and cortical bone from the mid-diaphysis of the femur from male individuals with type 1 diabetes mellitus (T1DM) (n = 6; 61.3 ± 14.6 years), type 2 diabetes mellitus (T2DM) (n = 11; 74.3 ± 7.9 years), and nondiabetic controls (n = 18; 69.3 ± 11.5) were analyzed with clinical and ex situ imaging techniques to explore various bone quality indices. Cortical collagen fibril deformation was measured in a synchrotron setup to assess changes at the nanoscale during tensile testing until failure. In addition, matrix composition was analyzed including determination of cross-linking and non-crosslinking advanced glycation end-products like pentosidine and carboxymethyl-lysine. In T1DM, lower fibril deformation was accompanied by lower mineralization and more mature crystalline apatite. In T2DM, lower fibril deformation concurred with a lower elastic modulus and tendency to higher accumulation of non-crosslinking advanced glycation end-products. The observed lower collagen fibril deformation in diabetic bone may be linked to altered patterns mineral characteristics in T1DM and higher advanced glycation end-product accumulation in T2DM. © 2022 The Authors. Journal of Bone and Mineral Research published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research (ASBMR).
59 BASIC BIOLOGICAL SCIENCES↗
Accommodating a hexagonal ζ- phase Mn 2 N film on a cubic MgO (001) substrate
The thin films of c -plane dominated hexagonal ζ- phase Mn 2 N were successfully grown on a cubic MgO (001) substrate directly using plasma-assisted molecular beam epitaxy. The surface was comprehensively studied through experimental and theoretical approaches. Reflection high energy electron diffraction revealed two pseudo-cubic domains along [100] MgO and [110] MgO , and one hexagonal domain, which is 30° apart from the [100] MgO direction. Scanning tunneling microscopy was used to image and resolve the high-quality surface, revealing a distorted hexagonal surface structure. Furthermore, atomic resolution of the hexagonal domain with a 2 × 2 reconstructed surface is presented. For c-plane surface, theoretical investigations were carried out using first principles studies to determine the surface formation energy for various reconstructed surfaces. The theoretical study shows that the nitrogen terminated 2 × 2 structure with a manganese adatom on the surface as the most stable reconstruction that reproduces the experimental results. A corresponding simulated scanning tunneling microscopy model is also presented, providing strong support for the experimentally observed in-plane lattice structures. Furthermore, the manganese: nitrogen stoichiometry within the bulk and surface is in good agreement with the expected ratio of 2:1.
75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
Machine learning-assisted upscaling analysis of reservoir rock core properties based on micro-computed tomography imagery
Optimum solutions for geologic modeling and reservoir simulation in industries such as oil and gas recovery and carbon capture and storage require accurate characterization of reservoir properties, which are often heterogeneous. In this study, high-quality micro-computed tomography (CT) images (1.475-μm/pixel resolution) of a sandstone core acquired from the Bell Creek oil field, USA, were used to provide nondestructive analysis of pore- and core-scale heterogeneity across measurement scales of 94–566 μm. In addition to characterizing the as-received sample, the core sample was flooded with brine to evaluate the capacity of the core sample to receive injected fluids. The micro-CT images were systematically segmented into pore spaces and grains via machine learning (ML) steps including image preprocessing, label creation using a traditional ML method based on limited manual image annotation, and finally U-Net segmentation. The segmented image stacks were reconstructed into digital cubes of various scales of voxel lengths. The 3D porosity values were calculated for all the digital cubes, and the fractal dimensions of the cubes were estimated using a box-counting method. The results showed that smaller cubes had greater heterogeneity and that the porosity values could be accurately estimated by fractal dimension and voxel lengths using ML models. For the core sample with brine flooding, the ratio of pores filled by brine to the total pore space was related to the porosity and could also be accurately estimated by porosity, fractal dimension, and voxel lengths using ML models. In conclusion, the results of this study demonstrate that the concept of fractal dimension can be a useful vector to perform upscaling analysis of sandstone rock heterogeneity from the pore to core scale and that fractal dimensions can be used to estimate porosity values and pore space-filling capacity across those scales.
58 GEOSCIENCES↗
Forecasting Battery Electrode Performance via Electrochemical Fluorescence Microscopy and Machine-Learning
Predicting lithium-ion battery performance is hindered by microscale electrode heterogeneities invisible to conventional diagnostics. Here, we combine electrochemical fluorescence microscopy (EFM), which maps electronic connectivity by visualizing an electrofluorophore reaction distribution, with a multitask ElasticNet regression to forecast discharge capacity from spatial heterogeneity. Analyzing 196 images from six pilot-scale LiNi 0.5 Mn 0.3 Co 0.2 O 2 cathodes with varying carbon loadings, we extract 62 descriptors that capture morphology and texture. A compact five-feature model predicts capacity across eight discharge rates, achieving a per-target R 2 of up to 0.63 and an overall R 2 of 0.92, with a mean absolute percentage error of less than 2%. This performance rivals impedance-based approaches while avoiding their reliance on postformation data and incomplete electronic network information. Our facile and rapid, image-driven method may enable electrode quality control upstream of costly cell assembly to offer a transformative tool for data-driven battery research and manufacturing.
battery electrodes↗
Storage Conditions of Human Kidney Tissue Sections Affect Spatial Lipidomics Analysis Reproducibility
Lipids often are labile, unstable, and tend to degrade overtime, so it is of the upmost importance to study these molecules in their most native state. In this study, we sought to understand the optimal storage conditions for spatial lipidomic analysis of human kidney tissue sections. Specifically, we evaluated human kidney tissue sections on several different days throughout the span of a week using our established protocol for elucidating lipids using high mass resolution matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI). We studied kidney tissue sections stored under five different conditions: open stored at –80 °C, vacuumed sealed and stored at –80 °C, with matrix preapplied before storage at –80 °C, under a nitrogen atmosphere and stored at –80 °C, and at room temperature in a desiccator. Results were compared to data obtained from kidney tissue sections that were prepared and analyzed immediately after cryosectioning. Data was processed using METASPACE. After a week of storage, the sections stored at room temperature showed the largest amount of lipid degradation, while sections stored under nitrogen and at –80 °C retained the greatest number of overlapping annotations in relation to freshly cut tissue. Overall, we found that molecular degradation of the tissue sections was unavoidable over time, regardless of storage conditions, but storing tissue sections in an inert gas at low temperatures can curtail molecular degradation within tissue sections.
60 APPLIED LIFE SCIENCES↗