Beyond JWST: A Technology Path to the Next Great UVOIR Space Telescope
No abstract available
James Web Space Telescope↗
SEARCH · Search NASA
Results for “Image segmentation”
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.
Patch-Based Convolutional Neural Networks for Multiple Microstructural Features Detection in FIB-SEM Micrographs of Irradiated Nuclear Fuel
Focused ion beam scanning electron microscopy (FIB-SEM) tomography has increasingly been utilized for acquiring three-dimensional (3D) microstructure features at the sub-micron scale in irradiated nuclear materials. This technique involves sequential ion beam slicing followed by electron beam imaging and compositional mapping using energy dispersive spectroscopy (EDS). Despite its growing use, several challenges persist. These include the time-intensive nature of data collection of EDS data, difficulties in distinguishing between various microstructures, and issues with image alignment. These challenges currently limit the broader application of FIB-SEM tomography in the field. To overcome these limitations, we propose using convolutional neural networks (CNNs) to automate microstructure identification in SEM images. Our study introduces a new framework for identifying microstructures in irradiated U-10Zr (wt. %) metallic fuel with limited annotated data. The framework includes the creation of a reliable annotated dataset with paired SEM and ground truth data from EDS maps, the applications of CNNs for microstructure identification, and the validation of model performance. Specifically, we employed the Segment Anything Model (SAM) to align SEM images with corresponding EDS maps and focused ion beam (FIB) tomography SEM data. We evaluate several models, including Patch-based U-Net, Attention U-Net, and Residual U-Net, finding that patch-based U-Net exhibits superior segmentation performance and consistency. This approach reduces reliance on EDS detectors and aids in accelerating nuclear material analysis process, highlighting the potential of advanced deep learning techniques to improve microstructural understanding in nuclear material. This is the first framework to integrate SAM and Patch-based CNN models for semantic segmentation of irradiated nuclear materials, with potential applicability to other tomography datasets.
36 - MATERIALS SCIENCE↗
Real-time sensing of optical alignment
The Large Deployable Reflector and other future segmented optical systems may require autonomous, real-time alignment of their optical surfaces. Researchers have developed gratings located directly on a mirror surface to provide interferometric sensing of the location and figure of the mirror. The grating diffracts a small portion of the incident beam to a diffractive focus where the designed diagnostics can be performed. Mirrors with diffraction gratings were fabricated in two separate ways. The formation of a holographic grating over the entire surface of a mirror, thereby forming a Zone Plate Mirror (ZPM) is described. Researchers have also used computer-generated hologram (CGH) patches for alignment and figure sensing of mirrors. When appropriately illuminated, a grid of patches spread over a mirror segment will yield a grid of point images at a wavefront sensor, with the relative location of the points providing information on the figure and location of the mirror. A particular advantage of using the CGH approach is that the holographic patches can be computed, fabricated, and replicated on a mirror segment in a mass production 1-g clean room environment.
Stier, Mark T.↗
Segmented mirror coarse phasing with white light interferometry: modeling and experiment on NGST's wavefront control testbed
A method of coarse phasing segmented mirrors using white light interferometry (WLI) has been developed for Next Generation Space Telescope (NGST) wavefront sensing and control. Using the broadband point spread function (PSF) of the segmented mirrors taken during a segment piston scan, the WLI can accurately detect small residual piston errors. WLI does not rely on any extra optics and uses only the final imaging camera. With its high sensitivity to small segment piston error WLI can be used as a complementary phasing algorithm to the dispersed fringe sensor (DFS) for NGST. The paper will present the results from modeling and experiment on the NGST's Wavefront Control Testbed (WCT).
Space↗
Classifying Urban Land Covers Using Local Indices of Spatial Complexity
The skewed statistical distributions of land cover types in complex, heterogeneous urban areas limits the effectiveness of traditional spectrally based maximum-likelihood classifiers. This work examines the utility of fractal dimension and Moran's I index of spatial autocorrelation in segmenting high-resolution panchromatic and lower-resolution multispectral imagery. Tools available in the Image Characterization and Modeling System (ICAMS) were used to analyze multi-temporal and multi-platform imagery of Atlanta, Georgia. In this example, land cover change trajectories from forest or grassland to built up land covers lead to decreased spatial autocorrelation. In lower resolution imagery such as Landsat MSS, the complex details of forested land covers and urbanized areas are smoothed, and texture-based change detection is less effective. Although segmentation of panchromatic images is possible using fractal dimension or Moran's I, widely differing land covers often yield similar values of these indices. Better results are obtained when a surface of local fractal dimension or spatial autocorrelation is combined as an additional layer in a supervised maximum-likelihood multispectral classification.
Arumugam, Mahesh↗
Automated segmentation of soft X-ray tomography: Native cellular structure with submicron resolution at high-throughput for whole-cell quantitative imaging in yeast
Soft X-ray tomography (SXT) is an invaluable tool for quantitatively analyzing cellular structures at suboptical isotropic resolution. However, it has traditionally depended on manual segmentation, limiting its scalability for large datasets. Here, we leverage a deep learning-based autosegmentation pipeline to segment and label cellular structures in hundreds of cells across three Saccharomyces cerevisiae strains. This task-based pipeline uses manual iterative refinement to improve segmentation accuracy for key structures, including the cell body, nucleus, vacuole, and lipid droplets, enabling high-throughput and precise phenotypic analysis. Using this approach, we quantitatively compared the three-dimensional (3D) whole-cell morphometric characteristics of wild-type, VPH1-GFP, and vac14 strains, uncovering detailed strain-specific cell and organelle size and shape variations. We show the utility of SXT data for precise 3D curvature analysis of entire organelles and cells and detection of fine morphological features using surface meshes. Our approach facilitates comparative analyses with high spatial precision and statistical throughput, uncovering subtle morphological features at the single-cell and population level. This workflow significantly enhances our ability to characterize cell anatomy and supports scalable studies on the mesoscale, with applications in investigating cellular architecture, organelle biology, and genetic research across diverse biological contexts.
Chen, Jianhua [Lawrence Berkeley National Laborato↗
Experiments in encoding multilevel images as quadtrees
Image storage requirements for several encoding methods are investigated and the use of quadtrees with multigray level or multicolor images are explored. The results of encoding a variety of images having up to 256 gray levels using three schemes (full raster, runlength and quadtree) are presented. Although there is considerable literature on the use of quadtrees to store and manipulate binary images, their application to multilevel images is relatively undeveloped. The potential advantage of quadtree encoding is that an entire area with a uniform gray level may be encoded as a unit. A pointerless quadtree encoding scheme is described. Data are presented on the size of the quadtree required to encode selected images and on the relative storage requirements of the three encoding schemes. A segmentation scheme based on the statistical variation of gray levels within a quadtree quadrant is described. This parametric scheme may be used to control the storage required by an encoded image and to preprocess a scene for feature identification. Several sets of black and white and pseudocolor images obtained by varying the segmentation parameter are shown.
Lansing, Donald L.↗
Integration of Optical Coherence Tomography Scan Patterns to Augment Clinical Data Suite
Vision changes identified in long duration spaceflight astronauts has led Space Medicine at NASA to adopt a more comprehensive clinical monitoring protocol. Optical Coherence Tomography (OCT) was recently implemented at NASA, including on board the International Space Station in 2013. NASA is collaborating with Heidelberg Engineering to increase the fidelity of the current OCT data set by integrating the traditional circumpapillary OCT image with radial and horizontal block images at the optic nerve head. The retinal nerve fiber layer was segmented by two experienced individuals. Intra-rater (N=4 subjects and 70 images) and inter-rater (N=4 subjects and 221 images) agreement was performed. The results of this analysis and the potential benefits will be presented.
Mason, S.↗
Focal Gray Matter Plasticity as a Function of Long Duration Head Down Tilted Bed Rest: Preliminary Results
Long duration spaceflight (i.e., 22 days or longer) has been associated with changes in sensorimotor systems, resulting in difficulties that astronauts experience with posture control, locomotion, and manual control. The microgravity environment is an important causal factor for spaceflight induced sensorimotor changes. Whether these sensorimotor changes are solely related to peripheral changes from reduced vestibular stimulation, body unloading, body fluid shifts or that they may be related to structural and functional brain changes is yet unknown. However, a recent study reported associations between microgravity and flattening of the posterior eye globe and protrusion of the optic nerve [1] possibly as the result of increased intracranial pressure due to microgravity induced bodily fluid shifts [3]. Moreover, elevated intracranial pressure has been related to white matter microstructural damage [2]. Thus, it is possible that spaceflight may affect brain structure and thereby cognitive functioning. Long duration head down tilt bed rest has been suggested as an exclusionary analog to study microgravity effects on the sensorimotor system [4]. Bed rest mimics microgravity in body unloading and bodily fluid shifts. In consideration of the health and performance of crewmembers both in- and post-flight, we are conducting a prospective longitudinal 70-day bed rest study as an analog to investigate the effects of microgravity on brain structure [5]. Here we present results of the first six subjects. Six subjects were assessed at 12 and 7 days before-, at 7, 30, and ~70 days in-, and at 8 and 12 days post 70 days of bed rest at the NASA bed rest facility in UTMB, Galveston, TX, USA. At each time point structural MRI scans (i.e., high resolution T1-weighted imaging and Diffusion Tensor Imaging (DTI)) were obtained using a 3T Siemens scanner. Focal changes over time in gray matter density were assessed using the voxel based morphometry 8 (VBM8) toolbox under SPM. Longitudinal processing in VBM8 includes linear registration of each scan to the mean of the subject and subsequently transforming all scans in to MNI space by applying the warp from the mean subject to MNI to the individual gray matter segmentations. Modulation was applied so that all images represented the volume of the original structure in native space. Voxel wise analysis was carried out on the gray matter images after smoothing, using a flexible factorial design with family wise error correction. Focal changes in white matter microstructural integrity were assessed using tract based spatial statistics (TBSS) as part of FMRIB software library (FSL). TBSS registers all DTI scans to standard space. It subsequently creates a study specific white matter skeleton of the major white matter tracts. For each subject, for each DTI metric (i.e. fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD)), the maximum value in a line perpendicular to the skeleton tract is projected to the skeleton. Non-parametric permutation based t-tests and ANOVA's were used for voxel-wise comparison of the skeletons. For both VBM and TBSS, comparison of pre bed rest measurements did not show significant differences. VBM analysis revealed decreased gray matter density in bilateral areas including the frontal medial cortex, the insular cortex and the caudate (see Figure) from 'pre to in bed rest'. Over the same time period, there was an increase in gray matter density in the cerebellum, occipital-, and parietal cortex, including the precuneus (see Figure). The majority of these changes did not recover from 'during to post bed rest'. TBSS analysis did not reveal significant changes in white matter microstructural integrity after correction for multiple comparisons. Uncorrected analyses (p<.015) revealed an increase in RD in the cerebellum and brainstem from pre bed rest to the first week in bed rest that did not recover post bed rest. Extended bed rest, which is an analog for microgravity, can result in gray matter changes and potentially in microstructural white matter changes in areas that are important for neuro motor behavior and cognition. These changes did not recover at two weeks post bed rest. Whether the effects of bed rest wear off at longer times post bed rest, and if they are associated with behavior are important questions that warrant further research.
Koppelmans, V.↗
Hubble Images from 1996
Primarily composed of animation, movies, and stills, this video is divided into 12 segments or slugs as the video refers to them. They are: Global Map of Pluto, Images of Pluto, Surface Map of Pluto, Helix Nebula- NGC 7293, Gaseous Knots, Animation of the Formation of the Helix Nebula, Crab Nebula, Jupiter Aurora Movie, Birth of a Quasar, Merging Galaxies, and Spiral Galaxies.
Source record↗
Deep learning‑based metal artefact reduction in X-ray computed tomography of TRISO fuel compacts
TRISO (TRi-structural ISOtropic) – compact-type micro-particle fuels – are next generation nuclear fuel compacts designed with safety in mind. Structural integrity and characterisation before and after irradiation are important to determine the performance of the fuel compacts under real reactor conditions. X-ray computed tomography can be an important tool for non-destructive evaluation of these fuel compacts. The fuel particles are highly attenuating for X-rays which creates metal artefact, rendering the images unusable. Our artefact correction can mitigate these artefacts significantly. The proposed method works by first segmenting highly-attenuating structures (metal) and forward projecting to localise the source of artefacts in the projection domain, then, using a traditional or U-net-based deep learning architecture, contextually interpolating those regions to remove the artefacts. Finally the reconstruction from the modified projection is fused with the segmented metal. The proposed method shows a significant improvement in the image quality while significantly reducing the reconstruction time compared to the standard technique.
Rahman, Obaid [ORNL] (ORCID:0000000277810840)↗
Vision based obstacle detection and grouping for helicopter guidance
Electro-optical sensors can be used to compute range to objects in the flight path of a helicopter. The computation is based on the optical flow/motion at different points in the image. The motion algorithms provide a sparse set of ranges to discrete features in the image sequence as a function of azimuth and elevation. For obstacle avoidance guidance and display purposes, these discrete set of ranges, varying from a few hundreds to several thousands, need to be grouped into sets which correspond to objects in the real world. This paper presents a new method for object segmentation based on clustering the sparse range information provided by motion algorithms together with the spatial relation provided by the static image. The range values are initially grouped into clusters based on depth. Subsequently, the clusters are modified by using the K-means algorithm in the inertial horizontal plane and the minimum spanning tree algorithms in the image plane. The object grouping allows interpolation within a group and enables the creation of dense range maps. Researchers in robotics have used densely scanned sequence of laser range images to build three-dimensional representation of the outside world. Thus, modeling techniques developed for dense range images can be extended to sparse range images. The paper presents object segmentation results for a sequence of flight images.
Sridhar, Banavar↗
A quantitative assay for intercellular aggregation
In an earlier communication (Munn et al., J Immunol. Methods 166: 11-25, 1993), we presented the initial development of a quantitative assay for monitoring the rates of cellular aggregation based on digital image processing and video microscopy. This study describes some important enhancements and modifications to the procedure. A new index is introduced to characterize the three-dimensional morphology of the aggregates. This index is based on temporal changes in the projected area of the cells and cell aggregates during the course of the experiment. By drawing an analogy with the kinetic theory of gases, we have also introduced a procedure to normalize for variations in cell seeding density among different experiments. In addition, the image analysis technique has been improved by introducing a background subtraction algorithm to remove illumination defects and an adaptive segmentation procedure. These improvements allowed us to completely automate the image analysis procedure, thus minimizing user intervention and improving the reproducibility of the measurements. The enhanced visual assay is evaluated using some recent results from our studies on homotypic lymphocyte aggregation.
Non-NASA Center↗
Apparatus and method for a light direction sensor
The present invention provides a light direction sensor for determining the direction of a light source. The system includes an image sensor; a spacer attached to the image sensor, and a pattern mask attached to said spacer. The pattern mask has a slit pattern that as light passes through the slit pattern it casts a diffraction pattern onto the image sensor. The method operates by receiving a beam of light onto a patterned mask, wherein the patterned mask as a plurality of a slit segments. Then, diffusing the beam of light onto an image sensor and determining the direction of the light source.
Leviton, Douglas B.↗
Reuniting crystallography with real space: Ab initio structure elucidation with 4D-STEM
Structure elucidation via single-crystal methods has historically lacked experimental access to real-space information, instead relying exclusively on diffraction-space measurements of Bragg reflections. Here we exploit the dual-space imaging power of 4D scanning transmission electron microscopy to meaningfully integrate real-space information into the crystallographic workflow. We show that virtual apertures assembled by segmentation of high-angle annular dark-field images enable i) pixel-by-pixel separation of coherent Bragg signal from clusters of closely spaced nanocrystals and ii) selective extraction of integrated intensities from thinner subregions of individual specimens, facilitating retroactive tuning of multiple scattering artifacts. This strategy empowers us to simply pick and choose whichever nanoscale regions of interest generate the highest-quality diffraction patterns, allowing us to solve several independent structures of the metal-organic framework UiO-66 from specimens whose agglomerated morphology proved intractable for conventional microcrystal electron diffraction. Our method is compatible with both rotational and serial approaches to data processing, ultimately divulging the first scanning nanobeam electron diffraction structures determined by direct methods at subangstrom resolution.
Saha, Ambarneil↗
Comparison of edges detected at different polarisations in MAESTRO data
Edge detection would appear to be a crucial tool for analyzing multi-polarized, multi-frequency, and multi-temporal Synthetic Aperture Radar (SAR) images. Edge structure provides a simple means for comparing different polarizations and frequencies, and for detecting changes over time. Due to the fact that edges and segments (homogeneous regions) are dual concepts, edge detection has an important role to play in identifying segments within which mean backscatter measurements for use in image classification can be made. As part of a general investigation into edge detection in SAR imagery, an initial investigation was carried out into the detectability and nature of edges in multi-polarized and multi-frequency SAR images. The contrast ratio (CR) operator was used to detect edges. This operator was previously shown to perform well at detecting edges in single-polarized and single-frequency SAR images.
Caves, Ronald G.↗
A parametric study of various synthetic aperture telescope configurations for coherent imaging applications
The comparative advantages of synthetic aperture telescopes (SATs) of segmented primary mirror and common secondary mirror type, on the one hand, and on the other those employing an array of independent telescopes, are discussed. The diffraction-limited optical performance of both redundant and nonredundant subaperture configurations are compared in terms of point spread function characteristics and encircled energy plots. Coherent imaging with afocal telescope SATs involves a pupil-mapping operation followed by a Fourier transform one. A quantitative analysis of the off-axis optical performance degradation due to pupil-mapping errors is presented, together with the field-dependent effects of residual design aberrations of independent telescopes.
Harvey, James E.↗
Microstructure Segmentation With Deep Learning Encoders Pre-Trained on a Large Microscopy Dataset
This study examined the improvement of microscopy segmentation intersection over union accuracy by transfer learning from a large dataset of microscopy images called MicroNet. Many neural network encoder architectures were trained on over 100,000 labeled microscopy images from 54 material classes. These pre-trained encoders were then embedded into multiple segmentation architectures including UNet and DeepLabV3+ to evaluate segmentation performance on created benchmark microscopy datasets. Compared to ImageNet pre-training, models pre-trained on MicroNet generalized better to out-of-distribution micrographs taken under different imaging and sample conditions and were more accurate with less training data. When training with only a single Ni-superalloy image, pre-training on MicroNet produced a 72.2% reduction in relative intersection over union error. These results suggest that transfer learning from large in-domain datasets generate models with learned feature representations that are more useful for downstream tasks and will likely improve any microscopy image analysis technique that can leverage pre-trained encoders.
machine learning↗