SEARCH · Engineering Papers
Results for “image generation”
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.
Using Deep Image Priors to Generate Counterfactual Explanations
Explore the source record for details and available documents.
Machine learning informed control systems for extrusion printing processes
Systems and methods for controlling a material extrusion device to extrude a filament of an ink are provided. An extrusion printing control system collects from one or more sensors measurements representing an internal state of material extrusion processing during extrusion of the filament. In addition, the system collects an image of the filament as the filament is extruded. The system applies a classifier to the collected image to generate an image-derived state characterizing the filament. Based on the internal state and the image-derived state, the system estimates a derived state using a model. The system determines control parameters using the model to achieve a desired quality of the filament by minimizing a cost function based on the internal state, the image-derived state, the derived state, and constraints of the material extrusion device. Finally, the system provides the control parameters to a controller of the material extrusion device.
Grating-Based Imaging-Scattering with Portable Neutron Generator
Company: Adelphi Technology, Inc. Title: Grating-Based Imaging-Scattering with Portable Neutron Generator PI: Dr. Jay Theodore Cremer, Jr. Topic: 26a Statement of the problem or situation that is being addressed. Successful plant growth depends upon an efficient and robust root system. The plant root is part of a larger system of water and microbial flows in the soil system. While much effort has been exerted to develop an imaging system for water, microbes, and roots, the problem is challenging, and no widely accepted imaging method currently exists. The optical solutions use a highly modified soil system. X-ray imaging methods are insensitive to the soft tissues in the presence of sand. Thermal neutron imaging has been often tested, but found inadequate, due to limited access and low image resolution. This project will develop a new strategy for neutron imaging of plant/soil systems. The project will allow long duration experiments in greenhouse environments and increase the image information content to the micron scale. General statement of how this problem is being addressed. Portable, rugged thermal and fast neutron sources are being developed where portable means a two-soldier team can carry the source and power unit to survey rough terrain for explosives. In the past decade, microfabrication of X-ray and thermal/cold neutron optics has opened a new imaging strategy. The standard transmission image is now supplemented with simultaneous acquisition of a phase contrast image and an image revealing scattering features. In materials science, the interferometric neutron scattering image has been used to detect early crack formation in stressed additive manufacturing test samples. The detection requires sensitivity to scattering features at the 1-micron scale. By addition of our proposed grating-optic to Adelphi Technology’s radiographic/tomographic imaging system, which is based on portable thermal neutron source, the resulting thermal neutron scatter image of the plant/soil system, will reveal details at 1-micron. Commercial Applications and Other Benefits Neutron interferometry imaging has greater penetration through large metal components compared to industrial X-ray imaging. The low-cost, large area optics developed for greenhouse applications, combined with the robust, portable neutron generator, can then be marketed as a system for inspection of additive manufactured components. In the aerospace industry, all freshly printed components are validated with X-ray CT. Scheduled maintenance again requires X-ray CT as the ability to predict aerospace component lifetime does not yet exist. Large aerospace components are only partially observed with X-ray imaging. Key Words Neutron radiography/tomography, grating interferometry, thermal neutron generator imaging, plant root and soil imaging, rhizosphere imaging, deployable neutron imaging systems Summary for Members of Congress A rugged, portable source of thermal neutrons is adapted for neutron interferometry imaging with the addition of low-cost, 3D printed optics. The first application of our proposed deployable, compact thermal neutron generator imaging system, using a grating optic, is plant root/soil science in greenhouse settings and agricultural laboratories.
Using Generative AI for Lunar Image Denoising: Instantaneous Clarity of Ambient eNvironment Capability (ICAN-C)
Using Generative AI for Lunar Image Denoising
Using Generative AI for Lunar Image Denoising: Instantaneous Clarity of Ambient eNvironment Capability (ICAN-C)
Using Generative AI for Lunar Image Denoising
Generation-V dual-Purkinje-image eyetracker
Major advances characterize the Generation-V dual-Purkinje-image eyetracker compared with the Generation-III version previously described. These advances include a large reduction in size, major improvements in frequency response and noise level, automatic alignment to a subject, and automatic adjustment for different separation between the visual and optic axes, which can vary considerably from subject to subject. In a number of applications described in the paper, the eyetracker is coupled with other highly specialized optical devices. These applications include accurately stabilizing an image on a subject's retina; accurately simulating a visually dead retinal region (i.e., a scotoma) of arbitrary shape, size, and position; and, for clinical purposes, stabilizing the position of a laser coagulator beam on a patient's retina so that the point of contact is unaffected by the patient's own eye movements.
Adaptive optical third-harmonic generation microscopy for in vivo imaging of tissues
Third-harmonic generation microscopy is a powerful label-free nonlinear imaging technique, providing essential information about structural characteristics of cells and tissues without requiring external labelling agents. In this work, we integrated a recently developed compact adaptive optics module into a third-harmonic generation microscope, to measure and correct for optical aberrations in complex tissues. Taking advantage of the high sensitivity of the third-harmonic generation process to material interfaces and thin membranes, along with the 1,300-nm excitation wavelength used here, our adaptive optical third-harmonic generation microscope enabled high-resolution in vivo imaging within highly scattering biological model systems. Examples include imaging of myelinated axons and vascular structures within the mouse spinal cord and deep cortical layers of the mouse brain, along with imaging of key anatomical features in the roots of the model plant Brachypodium distachyon. In all instances, aberration correction led to enhancements in image quality.
Randomization of particle motions and the observed morphology of cometary heads
A great diversity is known to exist in the coma morphology of comets. In particular, some comets show much structural detail in their heads (such as jets, halos, fans, plumes, streamers), while others have completely structureless comas. Obvious questions arise as to why is this so and what does the presence or absence of features tell us about the emission processes on cometary nuclei. In an effort to investigate these problems, a computer code that generates synthetic images of dust comets was modified by introducing random perturbations into motions of ejected particles. It is noted that the introduction of perturbations has made the computer generated images simulate the appearance of comets quite faithfully and that by increasing the perturbations beyond a certain limit it has been possible to erase the coma morphology diagnostic of the details of the ejection process. It is proposed that the degree of collimation of an ejecta flow from discrete active sources on the nucleus surface and possible emissions of dust coma. Molecules of comet gases that radiate in the spectral region employed (and whose velocity distribution is much more chaotic than that of dust particles) and limited atmospheric seeing likewise contribute to blurring structural detail in ground-based imaging observation of comets. The absence of discrete features in the coma does no means imply the absence of localized sources of activity on the nucleus.
Geometric ortho-rectification and generation of sigma(0) image products from multiple incidence synthetic aperture radar images
The results of the first phase of a cooperative effort in geometric orthorectification and generation of sigma(0) images of multiple incidence SAR images are presented. The geometric accuracy of the final image products is approximately 18 m or 1.5 pixels. A method for registering radar imagery collected from an airborne platform to an existing digital elevation model despite the effects of unmodeled variations in the flight path of the platform is demonstrated. The results indicate the requirements for a more detailed digital elevation model.
Average Cross-Sectional Area of DebriSat Fragments Using Volumetrically Constructed 3D Representations
Debris fragments from the hypervelocity impact testing of DebriSat are being collected and characterized for use in updating existing satellite breakup models. One of the key parameters utilized in these models is the ballistic coefficient of the fragment which is directly related to its area‐to‐mass ratio. However, since the attitude of fragments varies during their orbital lifetime, it is customary to use the average cross‐sectional area in the calculation of the area‐to‐mass ratio. The average cross‐sectional area is defined as the average of the projected surface areas perpendicular to the direction of motion and has been shown to be equal to one‐fourth of the total surface area of a convex object. Unfortunately, numerous fragments obtained from the DebriSat experiment show significant concavity (i.e., shadowing) and thus we have explored alternate methods for computing the average cross‐sectional area of the fragments. An imaging system based on the volumetric reconstruction of a 3D object from multiple 2D photographs of the object was developed for use in determining the size characteristic (i.e., characteristics length) of the DebriSat fragments. For each fragment, the imaging system generates N number of images from varied azimuth and elevation angles and processes them using a space‐carving algorithm to construct a 3D point cloud of the fragment. This paper describes two approaches for calculating the average cross‐sectional area of debris fragments based on the 3D imager. Approach A utilizes the constructed 3D object to generate equally distributed cross‐sectional area projections and then averages them to determine the average cross‐sectional area. Approach B utilizes a weighted average of the area of the 2D photographs to directly compute the average cross‐sectional area. A comparison of the accuracy and computational needs of each approach is described as well as preliminary results of an analysis to determine the "optimal" number of images needed for the 3D imager to accurately measure the average cross sectional area of objects with known dimensions.
Report on the AAPM grand challenge on deep generative modeling for learning medical image statistics
Abstract Background The findings of the 2023 AAPM Grand Challenge on Deep Generative Modeling for Learning Medical Image Statistics are reported in this Special Report. Purpose The goal of this challenge was to promote the development of deep generative models for medical imaging and to emphasize the need for their domain‐relevant assessments via the analysis of relevant image statistics. Methods As part of this Grand Challenge, a common training dataset and an evaluation procedure was developed for benchmarking deep generative models for medical image synthesis. To create the training dataset, an established 3D virtual breast phantom was adapted. The resulting dataset comprised about 108 000 images of size 512 512. For the evaluation of submissions to the Challenge, an ensemble of 10 000 DGM‐generated images from each submission was employed. The evaluation procedure consisted of two stages. In the first stage, a preliminary check for memorization and image quality (via the Fréchet Inception Distance [FID]) was performed. Submissions that passed the first stage were then evaluated for the reproducibility of image statistics corresponding to several feature families including texture, morphology, image moments, fractal statistics, and skeleton statistics. A summary measure in this feature space was employed to rank the submissions. Additional analyses of submissions was performed to assess DGM performance specific to individual feature families, the four classes in the training data, and also to identify various artifacts. Results Fifty‐eight submissions from 12 unique users were received for this Challenge. Out of these 12 submissions, 9 submissions passed the first stage of evaluation and were eligible for ranking. The top‐ranked submission employed a conditional latent diffusion model, whereas the joint runners‐up employed a generative adversarial network, followed by another network for image superresolution. In general, we observed that the overall ranking of the top 9 submissions according to our evaluation method (i) did not match the FID‐based ranking, and (ii) differed with respect to individual feature families. Another important finding from our additional analyses was that different DGMs demonstrated similar kinds of artifacts. Conclusions This Grand Challenge highlighted the need for domain‐specific evaluation to further DGM design as well as deployment. It also demonstrated that the specification of a DGM may differ depending on its intended use.
Some new techniques for processing remotely obtained images by self-generated spectral masks.
An extension of a new technique that makes possible parallel, simultaneous processing of remotely obtained images is presented. The technique holds out promise for automatic onboard classification of data. The central feature involves the generation of binary masks, directly from the image, based on object reflectance data, that group objects into equivalence classes. These masks, called equivalence class masks, can be used in various logical combinations to isolate classes of objects with a priori known reflectance or radiance signatures. Experimental verification of the technique is furnished for simple scenes. A computational scheme, based on a sequence of integrated irradiance measurements on the image, that makes it possible to identify objects within an equivalence class is suggested.
Simulation of ocean SAR images via phase history generation
A method for simulating an ocean synthetic-aperture radar (SAR) image is illustrated for a simple internal wave current pattern. The method calculates both the amplitude image and the radar signal history. The simulation model used consists of six stages. In the first stage, the full wave spectrum in two spatial coordinates is calculated from the wind speed and direction using the action spectral density equation. In the second stage, the pixel size is selected and used as the basis for dividing the spectrum into large and small scale motions. Realizations for the large-scale ocean surface height and velocity are then calculated. In the third stage, the sensor wavelength and geometry are used to calculate the small scale statistics (radar cross section, coherence time, root-mean square (RMS), radial velocity and RMS slope). In the fourth stage, an autoregressive method is used to generate a realization of the surface reflectivity history that is consistent with the radial velocity and the radial velocity variance. In the fifth stage, the signal history is generated by summing the reflectivities at the proper times with the appropriate antenna weighting function for the SAR. As a consequence of this process, speckle is automatically included in the signal. In the final stage, the SAR image can be created by using any of the traditional ways to process the signal history, including variable focusing and multilook processing.
Grain2mesh: A Python and cubit mesh generator from unprocessed mesoscale images
Predicting bulk behavior from microscale features constitutes a key objective in multiscale modeling research, often involving numerical models composed of finite elements that capture the diversity of constituent phases, shapes, and orientations within the material. The Grain2mesh toolbox allows the user to input unprocessed mesoscopic images for automatic segmentation, pre-processing, quality control, and numerical mesh generation. The numerical mesh generation incorporates Cubit routines to generate robust multi-phase mesh structure for use in computational mechanics solvers. The python classes developed contain detailed documentation and examples to support standard usage and case-specific alternative options.
Impact of LANDSAT MSS sensor differences on change detection analysis
Differences in overall sensor response, due to both factors within the sensor system, and external factors such as atmospheric effects, must be dealt with before change detection techniques using image differencing can be successfully implemented. The approach used to be coregister data to a base, followed by between image scatterogram generation and evaluation, and image differencing. Data from LANDSAT 2 and 3 were registered to LANDSAT 4 data. The area was limited to a small protion of the scene near the City of Santa Cruz because this was the only area cloud and haze free on all the images. Visual examination of each image, as well as the between image scatterograms, provide information on radiometric characteristics of each sensor. The image differencing, after normalizing for radiometric differences, provides a visual depiction of geometric distortions.
Ideal Resampling Of Discrete Sequences
Spectral information preserved to extent possible. Technique developed to shrink or expand discrete input sequence of numbers into output sequence in manner preserving input spectrum up to Nyquist limit of smaller of two sequences. In case of expansion, technique also prevents introduction of spurious frequencies not present in input. While applicable to many kinds of data sampled at regular interval, particularly useful in processing and enhancement of images, where discrete sequences spatially ordered sets of picture-element brightness values. Image coarsened by resampling to reduce number of picture elements while preserving as much as possible of original image information. Used to generate intermediate images at small intervals between frames, thereby suppressing appearance of jerky motion caused by sudden jumps between frames at low frame rate.