arXiv subjects

N. Duane Loh

Publications and source records attributed to N. Duane Loh.

At least 19 recordsLinked to original sources

arXiv2026

Context-Aware Deep Learning for Defect Classification in Atomic-Resolution STEM

Artificial intelligence is rapidly advancing materials characterization, yet most applications in electron microscopy rely solely on image contrast, overlooking the chemical and experimental context that shapes image formation. This limitation makes defect classification inherently ambiguous, as similar contrasts can arise from different materials or imaging conditions. Here we develop a context-aware learning framework that integrates image-derived contrast with metadata describing composition, beam energy, and detector geometry. Using a systematically constructed dataset of ~55 million simulated patches spanning 576 cases across 96 doped monolayer transition-metal dichalcogenides, we show that conditioning on contextual variables transforms defect classification from an ill-posed image-only task into a well-posed, physically grounded problem. The framework achieves over 98% accuracy on simulations and near-human agreement on experimental data, with a 94% reduction in posterior entropy. By emphasizing contextual grounding over architectural complexity, this approach links experimental image contrast to the underlying chemical and imaging conditions, supporting physically grounded defect assignments and a general pathway toward multimodal AI models for autonomous materials characterization.

cond-mat.mtrl-sci↗

arXiv2026

HSG-12M: A Large-Scale Benchmark of Spatial Multigraphs from the Energy Spectra of Non-Hermitian Crystals

AI is transforming scientific research by revealing new ways to understand complex physical systems, but its impact remains constrained by the lack of large, high-quality domain-specific datasets. A rich, largely untapped resource lies in non-Hermitian quantum physics, where the energy spectra of crystals form intricate geometries on the complex plane -- termed as Hamiltonian spectral graphs. Despite their significance as fingerprints for electronic behavior, their systematic study has been intractable due to the reliance on manual extraction. To unlock this potential, we introduce Poly2Graph: a high-performance, open-source pipeline that automates the mapping of 1-D crystal Hamiltonians to spectral graphs. Using this tool, we present HSG-12M: a dataset containing 11.6 million static and 5.1 million dynamic Hamiltonian spectral graphs across 1401 characteristic-polynomial classes, distilled from 177 TB of spectral potential data. Crucially, HSG-12M is the first large-scale dataset of spatial multigraphs -- graphs embedded in a metric space where multiple geometrically distinct trajectories between two nodes are retained as separate edges. This simultaneously addresses a critical gap, as existing graph benchmarks overwhelmingly assume simple, non-spatial edges, discarding vital geometric information. Benchmarks with popular GNNs expose new challenges in learning spatial multi-edges at scale. Beyond its practical utility, we show that spectral graphs serve as universal topological fingerprints of polynomials, vectors, and matrices, forging a new algebra-to-graph link. HSG-12M lays the groundwork for data-driven scientific discovery in condensed matter physics, new opportunities in geometry-aware graph learning and beyond.

cs.LG↗

arXiv2025

Motifs in self-organising cells

In complex systems, groups of interacting objects may form prevalent and persistent spatiotemporal patterns, which we refer to as motifs. These motifs can exhibit features that reveal how individual objects interact with one another. Simultaneously, the motifs can also interact, causing new coarse-grained properties to emerge in the system. In this paper, we found motifs in a simulated system of Dynamically Self-Organising cells. We also found that quantifying these motifs with a set of physically interpretable structural and dynamic features efficiently captures the interaction dynamics of the motifs' underlying cells. Using these motif features, we revealed packing strain and defects in large compact aggregates, semi-periodicity in motif ensembles, and phase space classes with unsupervised machine learning. Additionally, we trained neural networks to infer the critical hidden microscopic interaction parameters within each motif from coarse-grained motif features extracted from snapshots of the system. Furthermore, we uncovered emergent features that can predict the movement of cell collectives by hierarchically coarse-graining smaller motifs into larger ones (e.g. motif clusters). We speculate that this concept of motif hierarchies may be applied broadly to many-body interacting systems that are otherwise too complex to understand.

physics.bio-ph↗

arXiv2025

Neural network enabled wide field-of-view imaging with hyperbolic metalenses

The ultrathin form factor of metalenses makes them highly appealing for novel sensing and imaging applications. Amongst the various phase profiles, the hyperbolic metalens stands out for being free from spherical aberrations and having one of the highest focusing efficiencies to date. For imaging, however, hyperbolic metalenses present significant off-axis aberrations, severely restricting the achievable field-of-view (FOV). Extending the FOV of hyperbolic metalenses is thus feasible only if these aberrations can be corrected. Here, we demonstrate that a Restormer neural network can be used to correct these severe off-axis aberrations, enabling wide FOV imaging with a hyperbolic metalens camera. Importantly, we demonstrate the feasibility of training the Restormer network purely on simulated datasets of spatially-varying blurred images generated by the eigen-point-spread function (eigenPSF) method, eliminating the need for time-intensive experimental data collection. This reference-free training ensures that Restormer learns solely to correct optical aberrations, resulting in reconstructions that are faithful to the original scene. Using this method, we show that a hyperbolic metalens camera can be used to obtain high-quality imaging over a wide FOV of 54° in experimentally captured scenes under diverse lighting conditions.

physics.optics↗

arXiv2025

Nanoscale cuticle mass density variations influenced by pigmentation in butterfly wing scales

How pigment distribution influences the cuticle density within a microscopic butterfly wing scale, and how both impact each scale's final reflected color, remains unknown. We use ptychographic X-ray computed tomography to quantitatively determine, at nanoscale resolutions, the three-dimensional mass density of scales with pigmentation differences. By comparing cuticle densities between two pairs of scales with pigmentation differences, we determine that the density of the lower lamina is inversely correlated with pigmentation. In the upper lamina structure of Junonia orithya and Bicyclus anynana, low pigment levels also correlate with sheet-like chitin structures as opposed to rod-like structures. Within each scale, we determine that the lower lamina in all scales has the highest density, and distinct layers within the lower lamina help explain reflected color. We hypothesize that pigments, in addition to absorbing specific wavelengths, can affect cuticle polymerization, density, and refractive index, thereby impacting reflected wavelengths that produce colors.

physics.bio-ph↗

arXiv2024

Symmetry quantification and segmentation in STEM imaging through Zernike moments

We present a method using Zernike moments for quantifying rotational and reflectional symmetries in scanning transmission electron microscopy (STEM) images, aimed at improving structural analysis of materials at the atomic scale. This technique is effective against common imaging noises and is potentially suited for low-dose imaging and identifying quantum defects. We showcase its utility in the unsupervised segmentation of polytypes in a twisted bilayer TaS$_2$, enabling accurate differentiation of structural phases and monitoring transitions caused by electron beam effects. This approach enhances the analysis of structural variations in crystalline materials, marking a notable advancement in the characterization of structures in materials science.

cond-mat.mtrl-sci↗

arXiv2024

Resolving non-equilibrium shape variations amongst millions of gold nanoparticles

Nanoparticles, exhibiting functionally relevant structural heterogeneity, are at the forefront of cutting-edge research. Now, high-throughput single-particle imaging (SPI) with x-ray free-electron lasers (XFELs) creates unprecedented opportunities for recovering the shape distributions of millions of particles that exhibit functionally relevant structural heterogeneity. To realize this potential, three challenges have to be overcome: (1) simultaneous parametrization of structural variability in real and reciprocal spaces; (2) efficiently inferring the latent parameters of each SPI measurement; (3) scaling up comparisons between $10^5$ structural models and $10^6$ XFEL-SPI measurements. Here, we describe how we overcame these three challenges to resolve the non-equilibrium shape distributions within millions of gold nanoparticles imaged at the European XFEL. These shape distributions allowed us to quantify the degree of asymmetry in these particles, discover a relatively stable `shape envelope' amongst nanoparticles, discern finite-size effects related to shape-controlling surfactants, and extrapolate nanoparticles' shapes to their idealized thermodynamic limit. Ultimately, these demonstrations show that XFEL SPI can help transform nanoparticle shape characterization from anecdotally interesting to statistically meaningful.

cond-mat.mes-hall↗

arXiv2024

Ghostbuster: a phase retrieval diffraction tomography algorithm for cryo-EM

Ewald sphere curvature correction, which extends beyond the projection approximation, stretches the shallow depth of field in cryo-EM reconstructions of thick particles. Here we show that even for previously assumed thin particles, reconstruction artifacts which we refer to as ghosts can appear. By retrieving the lost phases of the electron exitwaves and accounting for the first Born approximation scattering within the particle, we show that these ghosts can be effectively eliminated. Our simulations demonstrate how such ghostbusting can improve reconstructions as compared to existing state-of-the-art software. Like ptychographic cryo-EM, our Ghostbuster algorithm uses phase retrieval to improve reconstructions, but unlike the former, we do not need to modify the existing data acquisition pipelines.

physics.comp-ph↗

arXiv2023

Single-shot, coherent, pop-out 3D metrology

Three-dimensional (3D) imaging of thin, extended specimens at nanometer resolution is critical for applications in biology, materials science, advanced synthesis, and manufacturing. One route to 3D imaging is tomography, which requires a tilt series of a local region. Here we describe a coherent imaging alternative that recovers the 3D volume of a thin, homogeneously amorphous specimen with only a single, energy-filtered, bright-field image. We demonstrated this technique with a transmission electron microscope to fill a glaring gap for rapid, accessible, non-destructive 3D nanometrology. This technique is applicable, in general, to any coherent bright field imaging with electrons, photons, or any other wavelike particles.

physics.ins-det↗

arXiv2023

Large field-of-view and multi-color imaging with GaP quadratic metalenses

Metalenses, in order to compete with conventional bulk optics in commercial imaging systems, often require large field of view (FOV) and broadband operation simultaneously. However, strong chromatic and coma aberrations present in common metalens designs have so far limited their widespread use. Stacking of metalenses as one of the possible solutions increases the overall complexity of the optical system and hinders the main benefit of reduced thickness and light weight. To tackle both issues, here we propose a single-layer imaging system utilizing a recently developed class of metalenses providing large field of view. Using it, we demonstrate full-color imaging with a FOV of 100 degrees. This approach, empowered by computational imaging techniques, produce high quality images, both in terms of color reproduction and sharpness. Suitable for real-time unpolarized light operation with the standard color filters present in prevalent camera systems, our results might enable a pathway for consumer electronics applications of this emerging technology.

physics.optics↗

arXiv2023

A multiscale generative model to understand disorder in domain boundaries

A continuing challenge in atomic resolution microscopy is to identify significant structural motifs and their assembly rules in synthesized materials with limited observations. Here we propose and validate a simple and effective hybrid generative model capable of predicting unseen domain boundaries in a potassium sodium niobate thin film from only a small number of observations, without expensive first-principles calculation. Our results demonstrate that complicated domain boundary structures can arise from simple interpretable local rules, played out probabilistically. We also found new significant tileable boundary motifs and evidence that our system creates domain boundaries with the highest entropy. More broadly, our work shows that simple yet interpretable machine learning models can help us describe and understand the nature and origin of disorder in complex materials.

cond-mat.mtrl-sci↗

arXiv2021

Learning Motifs and their Hierarchies in Atomic Resolution Microscopy

Progress in functional materials discovery has been accelerated by advances in high throughput materials synthesis and by the development of high-throughput computation. However, a complementary robust and high throughput structural characterization framework is still lacking. New methods and tools in the field of machine learning suggest that a highly automated high-throughput structural characterization framework based on atomic-level imaging can establish the crucial statistical link between structure and macroscopic properties. Here we develop a machine learning framework towards this goal. Our framework captures local structural features in images with Zernike polynomials, which is demonstrably noise-robust, flexible, and accurate. These features are then classified into readily interpretable structural motifs with a hierarchical active learning scheme powered by a novel unsupervised two-stage relaxed clustering scheme. We have successfully demonstrated the accuracy and efficiency of the proposed methodology by mapping a full spectrum of structural defects, including point defects, line defects, and planar defects in scanning transmission electron microscopy (STEM) images of various 2D materials, with greatly improved separability over existing methods. Our techniques can be easily and flexibly applied to other types of microscopy data with complex features, providing a solid foundation for automatic, multiscale feature analysis with high veracity.

cond-mat.mtrl-sci↗

arXiv2021

Unsupervised learning approaches to characterize heterogeneous samples using X-ray single particle imaging

One of the outstanding analytical problems in X-ray single particle imaging (SPI) is the classification of structural heterogeneity, which is especially difficult given the low signal-to-noise ratios of individual patterns and that even identical objects can yield patterns that vary greatly when orientation is taken into consideration. We propose two methods which explicitly account for this orientation-induced variation and can robustly determine the structural landscape of a sample ensemble. The first, termed common-line principal component analysis (PCA) provides a rough classification which is essentially parameter-free and can be run automatically on any SPI dataset. The second method, utilizing variation auto-encoders (VAEs) can generate 3D structures of the objects at any point in the structural landscape. We implement both these methods in combination with the noise-tolerant expand-maximize-compress (EMC) algorithm and demonstrate its utility by applying it to an experimental dataset from gold nanoparticles with only a few thousand photons per pattern and recover both discrete structural classes as well as continuous deformations. These developments diverge from previous approaches of extracting reproducible subsets of patterns from a dataset and open up the possibility to move beyond studying homogeneous sample sets and study open questions on topics such as nanocrystal growth and dynamics as well as phase transitions which have not been externally triggered.

eess.IV↗

arXiv2020

Ptychographic wavefront characterisation for single-particle imaging at X-ray lasers

A well-characterised wavefront is important for many X-ray free-electron laser (XFEL) experiments, especially for single-particle imaging (SPI), where individual bio-molecules randomly sample a nanometer-region of highly-focused femtosecond pulses. We demonstrate high-resolution multiple-plane wavefront imaging of an ensemble of XFEL pulses, focused by Kirkpatrick-Baez (KB) mirrors, based on mixed-state ptychography, an approach letting us infer and reduce experimental sources of instability. From the recovered wavefront profiles, we show that while local photon fluence correction is crucial and possible for SPI, a small diversity of phase-tilts likely has no impact. Our detailed characterisation will aid interpretation of data from past and future SPI experiments, and provides a basis for further improvements to experimental design and reconstruction algorithms.

physics.optics↗

arXiv2020

ReCoDe: A Data Reduction and Compression Description for High Throughput Time-Resolved Electron Microscopy

Fast, direct electron detectors have significantly improved the spatio-temporal resolution of electron microscopy movies. Preserving both spatial and temporal resolution in extended observations, however, requires storing prohibitively large amounts of data. Here, we describe an efficient and flexible data reduction and compression scheme (ReCoDe) that retains both spatial and temporal resolution by preserving individual electron events. Running ReCoDe on a workstation we demonstrate on-the-fly reduction and compression of raw data streaming off a detector at 3 GB/s, for hours of uninterrupted data collection. The output was 100-fold smaller than the raw data and saved directly onto network-attached storage drives over a 10 GbE connection. We discuss calibration techniques that support electron detection and counting (e.g. estimate electron backscattering rates, false positive rates, and data compressibility), and novel data analysis methods enabled by ReCoDe (e.g. recalibration of data post acquisition, and accurate estimation of coincidence loss).

physics.ins-det↗

arXiv2020

High-throughput ensemble characterization of individual core-shell nanoparticles with quantitative 3D density from XFEL single-particle imaging

The structures, as building-blocks for designing functional nanomaterials, have fueled the development of versatile nanoprobes to understand local structures of noncrystalline specimens. Progresses in analyzing structures of individual specimens with atomic scale accuracy have been notable recently. In most cases, however, only a limited number of specimens are inspected lacking statistics to represent the systems with structural inhomogeneity. Here, by employing single-particle imaging with X-ray free electron lasers and new algorithm for multiple-model 3D imaging, we succeeded in investigating several thousand specimens in a couple of hours, and identified intrinsic heterogeneities with 3D structures. Quantitative analysis has unveiled 3D morphology, facet indices and elastic strains. The 3D elastic energy distribution is further corroborated by molecular dynamics simulations to gain mechanical insight at atomic level. This work establishes a new route to high-throughput characterization of individual specimens in large ensembles, hence overcoming statistical deficiency while providing quantitative information at the nanoscale.

cond-mat.mtrl-sci↗

arXiv2020

An encryption-decryption framework for validating single-particle imaging

We propose an encryption-decryption framework for validating diffraction intensity volumes reconstructed using single-particle imaging (SPI) with x-ray free-electron lasers (XFELs) when the ground truth volume is absent. This framework exploits each reconstructed volumes' ability to decipher latent variables (e.g. orientations) of unseen sentinel diffraction patterns. Using this framework, we quantify novel measures of orientation disconcurrence, inconsistency, and disagreement between the decryptions by two independently reconstructed volumes. We also study how these measures can be used to define data sufficiency and its relation to spatial resolution, and the practical consequences of focusing XFEL pulses to smaller foci. This framework overcomes critical ambiguities in using Fourier Shell Correlation (FSC) as a validation measure for SPI. Finally, we show how this encryption-decryption framework naturally leads to an information-theoretic reformulation of the resolving power of XFEL-SPI, which we hope will lead to principled frameworks for experiment and instrument design.

physics.data-an↗

arXiv2020

3D diffractive imaging of nanoparticle ensembles using an X-ray laser

We report the 3D structure determination of gold nanoparticles (AuNPs) by X-ray single particle imaging (SPI). Around 10 million diffraction patterns from gold nanoparticles were measured in less than 100 hours of beam time, more than 100 times the amount of data in any single prior SPI experiment, using the new capabilities of the European X-ray free electron laser which allow measurements of 1500 frames per second. A classification and structural sorting method was developed to disentangle the heterogeneity of the particles and to obtain a resolution of better than 3 nm. With these new experimental and analytical developments, we have entered a new era for the SPI method and the path towards close-to-atomic resolution imaging of biomolecules is apparent.

physics.app-ph↗