Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “multi-view”

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 37 records · Page 2

Levenberg–Marquardt multi-classification using hinge loss function

Incorporating higher-order optimization functions, such as Levenberg-Marquardt (LM) have revealed better generalizable solutions for deep learning problems. However, these higher-order optimization functions suffer from very large processing time and training complexity especially as training datasets become large, such as in multi-view classification problems, where finding global optima is a very costly problem. To solve this issue, we develop a solution for LM-enabled classification with, to the best of knowledge first-time implementation of hinge loss, for multiview classification. Hinge loss allows the neural network to converge faster and perform better than other loss functions such as logistic or square loss rates. Here we prove our method by experimenting with various multiclass classification challenges of varying complexity and training data size. The empirical results show the training time and accuracy rates achieved, highlighting how our method outperforms in all cases, especially when training time is limited. Our paper presents important results in the relationship between optimization and loss functions and how these can impact deep learning problems.

97 MATHEMATICS AND COMPUTING↗

YOLO2U-Net: Detection-guided 3D instance segmentation for microscopy

Microscopy imaging techniques are instrumental for characterization and analysis of biological structures. As these techniques typically render 3D visualization of cells by stacking 2D projections, issues such as out-of-plane excitation and low resolution in the z-axis may pose challenges (even for human experts) to detect individual cells in 3D volumes as these non-overlapping cells may appear as overlapping. In this paper a comprehensive method for accurate 3D instance segmentation of cells in the brain tissue is introduced. The proposed method combines the 2D YOLO detection method with a multi-view fusion algorithm to construct a 3D localization of the cells. Next, the 3D bounding boxes along with the data volume are input to a 3D U-Net network that is designed to segment the primary cell in each 3D bounding box, and in turn, to carry out instance segmentation of cells in the entire volume. The promising performance of the proposed method is shown in comparison with current deep learning-based 3D instance segmentation methods.

3D instance segmentation↗

Fast and accurate view classification of echocardiograms using deep learning

Echocardiography is essential to cardiology. However, the need for human interpretation has limited echocardiography’s full potential for precision medicine. Deep learning is an emerging tool for analyzing images but has not yet been widely applied to echocardiograms, partly due to their complex multi-view format. The essential first step toward comprehensive computer-assisted echocardiographic interpretation is determining whether computers can learn to recognize these views. We trained a convolutional neural network to simultaneously classify 15 standard views (12 video, 3 still), based on labeled still images and videos from 267 transthoracic echocardiograms that captured a range of real-world clinical variation. Our model classified among 12 video views with 97.8% overall test accuracy without overfitting. Even on single low-resolution images, accuracy among 15 views was 91.7% vs. 70.2–84.0% for board-certified echocardiographers. Data visualization experiments showed that the model recognizes similarities among related views and classifies using clinically relevant image features. Our results provide a foundation for artificial intelligence-assisted echocardiographic interpretation.

Madani, Ali↗

Tunable diode laser absorption spectroscopy-based tomography system for on-line monitoring of two-dimensional distributions of temperature and H{sub 2}O mole fraction

To monitor two-dimensional (2D) distributions of temperature and H{sub 2}O mole fraction, an on-line tomography system based on tunable diode laser absorption spectroscopy (TDLAS) was developed. To the best of the authors’ knowledge, this is the first report on a multi-view TDLAS-based system for simultaneous tomographic visualization of temperature and H{sub 2}O mole fraction in real time. The system consists of two distributed feedback (DFB) laser diodes, a tomographic sensor, electronic circuits, and a computer. The central frequencies of the two DFB laser diodes are at 7444.36 cm{sup −1} (1343.3 nm) and 7185.6 cm{sup −1} (1391.67 nm), respectively. The tomographic sensor is used to generate fan-beam illumination from five views and to produce 60 ray measurements. The electronic circuits not only provide stable temperature and precise current controlling signals for the laser diodes but also can accurately sample the transmitted laser intensities and extract integrated absorbances in real time. Finally, the integrated absorbances are transferred to the computer, in which the 2D distributions of temperature and H{sub 2}O mole fraction are reconstructed by using a modified Landweber algorithm. In the experiments, the TDLAS-based tomography system was validated by using asymmetric premixed flames with fixed and time-varying equivalent ratios, respectively. The results demonstrate that the system is able to reconstruct the profiles of the 2D distributions of temperature and H{sub 2}O mole fraction of the flame and effectively capture the dynamics of the combustion process, which exhibits good potential for flame monitoring and on-line combustion diagnosis.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Structure of Self-Generated Magnetic Fields in Laser-Solid Interaction from Proton Tomography

Self-generated magnetic fields in laser-solid interactions are experimentally characterized to reveal the 3D location and local field strength, rather than path-integrated quantities, using multi-view proton radiography and tomographic inversion. We infer magnetic fields that extend several millimeters off the target into the hot, rarefied corona, sufficient to strongly magnetize the plasma (Ω e τ e ≫ 1). The data are compared to MHD simulations incorporating recent improvements in modeling magnetic field generation and transport; the volume-averaged coronal field strength and magnetic flux agree to within 25% using a model with magnetic re-localization of transport, although the near-target morphology is not reproduced. This work demonstrates tomographic proton radiography as a valuable tool for investigating magnetic fields in laser-produced plasmas.

High-energy-density plasmas↗

High-speed 3D shape measurement using the optimized composite fringe patterns and stereo-assisted structured light system

In this study, we propose a high-speed 3D shape measurement technique based on the optimized composite fringe patterns and stereo-assisted structured light system. Stereo phase unwrapping, as a new-fashioned method for absolute phase retrieval based on the multi-view geometric constraints, can eliminate the phase ambiguities and obtain a continuous phase map without projecting any additional patterns. However, in order to ensure the stability of phase unwrapping, the period of fringe is generally around 20, which limits the accuracy of 3D measurement. To solve this problem, we develop an optimized method for designing the composite pattern, in which the speckle pattern is embedded into the conventional 4-step phase-shifting fringe patterns without compromising the fringe modulation, and thus the phase measurement accuracy. We also present a simple and effective evaluation criterion for the correlation quality of the designed speckle pattern in order to improve the matching accuracy significantly. When the embedded speckle pattern is demodulated, the periodic ambiguities in the wrapped phase can be eliminated by combining the adaptive window image correlation with geometry constraint. Finally, some mismatched regions are further corrected based on the proposed regional diffusion compensation technique (RDC). These proposed techniques constitute a complete computational framework that allows to effectively recover an accurate, unambiguous, and distortion-free 3D point cloud with only 4 projected patterns. Experimental results verify that our method can achieve high-speed, high-accuracy, robust 3D shape measurement with dense (64-period) fringe patterns at 5000 frames per second.

42 ENGINEERING↗

Bistatic SAR: Imagery & Image Products.

While typical SAR imaging employs a co-located (monostatic) RADAR transmitter and receiver, bistatic SAR imaging separates the transmitter and receiver locations. The transmitter and receiver geometry determines if the scattered signal is back scatter, forward scatter, or side scatter. The monostatic SAR image is backscatter. Therefore, depending on the transmitter/receiver collection geometry, the captured imagery may be quite different that that sensed at the monostatic SAR. This document presents imagery and image products formed from captured signals during the validation stage of the bistatic SAR research. Image quality and image characteristics are discussed first. Then image products such as two-color multi-view (2CMV) and coherent change detection (CCD) are presented.

Yocky, David A.↗

Development of 3-D Gamma-camera

BNL and Gamma Medical Technologies (GMT) will jointly develop a compact three-dimensional (3-D) gamma camera capable of localizing cancerous tumors within the prostate gland. The new gamma-ray camera will consist of pixilated cadmium-zinc-telluride (CZT) detector arrays, which will be flip-chip bonded to a suitable substrate. Application specific integrated circuits (ASIC) will also be developed to read out the electronic signals from the pixels. The detector/ASIC coupled to a multi-view collimator would be integrated into a small probe capable of imaging shaped radioactive phantoms to simulate the presence of cancerous tumors within the prostate gland.

99 GENERAL AND MISCELLANEOUS↗

Applications Of Machine Learning to Gas Plume Analysis In Longwave Infrared Hyperspectral Images

Longwave infrared hyperspectral images can be used for gas plume analysis, as many gases exhibit distinct absorption features in this portion of the electromagnetic spectrum. In practice, accurately identifying weak gas signatures is difficult because the observed radiance is dominated by background radiance, which varies with material, temperature, and viewing conditions. Many gas plume analysis pipelines operate on single images, limiting the ability to leverage spatial and multi-view information that could enhance the analysis. The goal of this dissertation is to explore how machine learning and deep learning methods can complement classical approaches to improve gas plume identification in longwave infrared hyperspectral imagery, and to investigate the use of neural radiance fields for hyperspectral scene reconstruction.

3D Scene Reconstruction↗

Multi-modal dynamic radiography using short-pulse laser-generated probe beams

Radiography is an important tool for the interrogation of dynamic experiments in the fields of dynamic properties of materials, and in condensed matter, high explosive, and high-energy-density physics. Multi-modal radiography advances the hypothesis that combining the information delivered by multiple radiographic modalities can lead to more constrained (improved) “reconstruction” of the scene than can be obtained from a single probe. We identify four modalities: multi-probe, time sequence, multi-view, and multi-messenger. Multi-probe radiography is a promising candidate for a next-generation dynamic radiographic facility. High-energy X-rays are the most frequently used probe for dynamic radiography, although recent developments show the utility of proton (pRad), electron (eRad), and neutron probe beams. Because each probing species interacts with material in the radiographic scene through quantitatively different mechanisms, each returns independent information about the scene, which can add extra constraints to the reconstruction process. How to conduct detailed, quantitative “co-analysis” of multiple data streams remains an area of active research. Multi-beam, short-pulse, laser-generated probes offer sufficient dose, an appropriate spectrum, and appropriate spatio-temporal resolution to produce high-quality dynamic radiographs. This paper reports on technology development to advance the state of the art of multi-modal/multi-probe radiography and the pursuit of both deterministic and inferential (AI/ML assisted) co-analysis methodologies to produce more constrained reconstructions from multi-modal data.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Vehicular Re-Identification from Uncontrolled Multiple Views

Vehicle re-identification (re-ID) across disparate sensing modalities remains a fundamental challenge for transportation research. In this work, we introduce a deep multi-view vehicle re-ID framework that leverages Siamese networks to compare pairs of vehicle images and produce matching scores, enabling robust association across drastically different viewpoints such as those from UAVs, surveillance cameras, and ground sensors. The model exploits convolutional neural networks to learn features that remain discriminative under changes in angle, distance, and illumination, supporting more generalizable re-ID performance. As part of this effort, we also developed an automated pipeline to synchronize roadside and UAV video streams, producing a multi-perspective dataset that complements preexisting real collections and a synthetic dataset generated in this study. Together, these contributions advance the capability to re-identify vehicles across wide viewing baselines; establish a foundation for scalable, reproducible research in vehicle re-ID; and open pathways for future applications, such as inferring routine behaviors, movement patterns, and daily habits of the individual associated with the vehicle.

convolutional neural networks↗

Deep Learning and Photogrammetric Reconstruction for Automated Crack Detection and Dimensional Measurement in Mining Operations

Surface crack detection and dimensional measurement at active mining sites present significant safety and operational challenges. Manual inspection methods are labor-intensive, spatially incomplete, and expose personnel to hazardous environments, while existing automated approaches have been developed primarily for concrete civil infrastructure and have not been validated on the complex, variable surfaces characteristic of mining environments. This dissertation presents an automated pipeline that integrates deep learning semantic segmentation with Structure-from-Motion photogrammetry to detect surface cracks and measure their aperture, length, and vertical displacement from standard RGB imagery acquired during routine Uncrewed Aerial Vehicle (UAV) survey operations, without requiring additional sensor hardware or manual measurement. The pipeline combines a U-Net architecture with an EfficientNet-B0 encoder, pretrained on the SDNET2018 concrete crack dataset and fine-tuned on a mining-specific dataset spanning laboratory concrete specimens, coal refuse impoundment embankments, and post-blast limestone quarry benches. Photogrammetric reconstruction is performed using COLMAP Structure-from-Motion and Multi-View Stereo, with crack segmentation masks projected into the reconstructed point cloud to enable three-dimensional vertical displacement measurement through local plane fitting and bimodal surface detection. The pipeline was validated across 36 controlled laboratory specimens at three imaging distances and four vertical displacement levels, achieving aperture measurement RMSE of 0.047 cm and R² of 0.954, and vertical displacement RMSE of 0.140 cm and R² of 0.966, against independent caliper measurements. Field application at a coal refuse impoundment in southwestern Pennsylvania detected 71 crack components across the embankment crest, with a dominant longitudinal crack exhibiting aperture values reaching 28 cm and a 95th percentile vertical displacement of 35.53 cm, consistent in magnitude and spatial distribution with simultaneously acquired LiDAR-derived estimates. Application across four post-blast limestone quarry bench datasets in California successfully characterized blast-induced fracture networks at ground sampling distances ranging from 0.59 to 1.23 cm/pixel, with detected crack geometries physically consistent with observable surface conditions at each site. The results demonstrate that deep learning-based crack detection and photogrammetric measurement can be integrated into routine UAV inspection workflows at mining sites, providing repeatable, scalable, and quantitative crack characterization across surface types, crack scales, and displacement magnitudes not previously addressed in the literature. The pipeline requires no dedicated surveying equipment beyond the UAV platforms already deployed at mine sites for survey and monitoring purposes, supporting practical adoption within existing operational workflows.

Crack detection, Dimensional Measurement↗

Using snowflake surface-area-to-volume ratio to model and interpret snowfall triple-frequency radar signatures

Abstract. The snowflake microstructure determines the microwave scattering properties of individual snowflakes and has a strong impact on snowfall radar signatures. In this study, individual snowflakes are represented by collections of randomly distributed ice spheres where the size and number of the constituent ice spheres are specified by the snowflake mass and surface-area-to-volume ratio (SAV) and the bounding volume of each ice sphere collection is given by the snowflake maximum dimension. Radar backscatter cross sections for the ice sphere collections are calculated at X-, Ku-, Ka-, and W-band frequencies and then used to model triple-frequency radar signatures for exponential snowflake size distributions (SSDs). Additionally, snowflake complexity values obtained from high-resolution multi-view snowflake images are used as an indicator of snowflake SAV to derive snowfall triple-frequency radar signatures. The modeled snowfall triple-frequency radar signatures cover a wide range of triple-frequency signatures that were previously determined from radar reflectivity measurements and illustrate characteristic differences related to snow type, quantified through snowflake SAV, and snowflake size. The results show high sensitivity to snowflake SAV and SSD maximum size but are generally less affected by uncertainties in the parameterization of snowflake mass, indicating the importance of snowflake SAV for the interpretation of snowfall triple-frequency radar signatures.

54 ENVIRONMENTAL SCIENCES↗

pccp.c1

Point Cloud of Cloud Points (PCCP) VAP provides a stereo reconstructed Point Cloud of Cloud Points (PCCP) that correspond to the x,y,z coordinates of cloud features detected from stereo image pairs genrated from a multi-view stereo system deployed to collect exclusive set of data on shallow cumulus clouds overhead.

54 ENVIRONMENTAL SCIENCES↗

pccpa.c1

Point Cloud of Cloud Points (PCCP) VAP provides a stereo reconstructed Point Cloud of Cloud Points (PCCP) that correspond to the x,y,z coordinates of cloud features detected from stereo image pairs genrated from a multi-view stereo system deployed to collect exclusive set of data on shallow cumulus clouds overhead.

54 ENVIRONMENTAL SCIENCES↗

pccpb.c1

Point Cloud of Cloud Points (PCCP) VAP provides a stereo reconstructed Point Cloud of Cloud Points (PCCP) that correspond to the x,y,z coordinates of cloud features detected from stereo image pairs generated from a multi-view stereo system deployed to collect exclusive set of data on shallow cumulus clouds overhead.

54 ENVIRONMENTAL SCIENCES↗

Novel Gamma Attenuation Methods Utilized for Waste Characterization of UO{sub 2}F{sub 2}-Bearing Materials at the Paducah GDP - 17327

The Paducah Gaseous Diffusion Plant (PGDP) is currently under deactivation with demolition in the decades to follow. UO{sub 2}F{sub 2}-bearing waste is generated at PGDP by UF{sub 6} hydrolysing to materials such as pipes, valves, PPE, etc. when UF{sub 6} is exposed to humid air. In addition, Al{sub 2}O{sub 3} and NaF trap materials contain significant amounts of Uranium. The UO{sub 2}F{sub 2} deposits within the generated waste can be of sub-gram levels to several hundreds of localized grams and are much denser than the surrounding interfering matrix materials and may be of very complex geometric and chemical forms such as with the Al{sub 2}O{sub 3} and NaF trap materials. The main objective is to characterize the fissile (U-235) content of the UO{sub 2}F{sub 2}-bearing waste so as to meet the Data Quality Objectives (DQO) for proper disposal. If a waste container cannot meet the DQO then it must be subjected to costly remediation, repackaging and re-characterization processes. We have developed, at PGDP, several novel model-based methods and techniques that vastly improve the standard gamma attenuation correction methods and do not require Working Reference Materials (WRM) for calibration. These new methods involve rotating and non-rotating 18.9 liter (5 U.S. gallon) waste containers with and without a transmission modality. A new model-based variant of differential attenuation has been developed that provides a high-quality self-absorption correction that does not require knowledge of the Uranium enrichment or isotopic ratios. For more challenging waste containers, a method using multi-view (non-rotating) 18.9 liter waste containers has been considered using a recently developed Advanced ISOCS{sup C} Uncertainty Estimator (AIUE) package. The new methods suffice to significantly improve accuracy and precision and to lower the Total Measurement Uncertainty (TMU) leading to satisfactory DQO's that were not previously achievable using the standard gamma attenuation correction methods. In this paper we present real 18.9 liter waste container and Working Reference Material (WRM) data subjected to the new methods and compared to standard methods depicting the significant improvement in the DQO that is realized. (authors)

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

A Real2Sim Digital Twin Pipeline for Photorealistic Robot Simulation: Evaluating VLA Policy Deployment on a Bimanual Mobile Robot

Digital twins that are automatically constructed from robot sensor data offer a promising pathway for scalable Real2Sim and Sim2Real transfer. However, it remains an open question whether photorealistic reconstruction alone is sufficient to support reliable deployment of vision-language-action (VLA) policies. We present a generative-AI-assisted Real2Sim pipeline that generates simulation-ready digital twins from real-world RGB observations with minimal manual intervention. The pipeline uses prompted segmentation to isolate scene components and a generative 3D model to directly produce simulation assets, eliminating the need for traditional multi-view reconstruction or manual 3D modeling.\r\nTo evaluate simulation fidelity, we deploy and compare policies from two VLA models in both the real robot and the reconstructed\r\nsimulation under identical tasks and initial conditions. We compare joint-level action trajectories and analyze how divergence evolves over time in closed-loop execution. Although the reconstructed environments are visually accurate, we observe increasing trajectory divergence during closedloop operation. These results indicate that photorealistic reconstruction alone is insufficient to preserve closed-loop control behavior\r\nin VLA policies, particularly in contact-rich manipulation settings where small perceptual errors compound over time.

97 MATHEMATICS AND COMPUTING↗