Search NASA⌕ Search

SEARCH · Search NASA

Results for “3D 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.

At least 199 records · Page 11

Three-dimensional (3D) microstructural characterization and quantification of reflow porosity in Sn-rich alloy/copper joints by X-ray tomography

In this paper high resolution X-ray tomography was used to characterize reflow porosity in Sn-3.9Ag-0.7Cu/Cu solder joints. The combination of two segmentation techniques was applied for the three-dimensional (3D) visualization of pores in the joints and the quantification on the characteristics of reflow porosity, including pore size, volume fraction and morphology. The size, morphology and distribution of porosity were visualized in 3D for three different solder joints. Since the results are relatively similar for all three, only the results of one joint are presented. Solder reflow porosity was mostly spherical, segregated along the solder/Cu interface, and had an average pore size of 30 {mu}m in diameter. A few large pores (larger than 100 {mu}m in diameter) were present, some of which had lower sphericity, i.e., they were more irregular. The presence of these large pores may significantly influence the mechanical behavior of solder joints. - Highlights: {yields} Non-destructive 3D characterization and quantification of porosity in Pb-free solders by X-ray tomography {yields} Two new image analysis and reconstruction tools are presented that can be used by the community at large {yields} Pore size, volume fraction, and sphericity, is critical to understanding microstructure and modeling of these systems.

36 MATERIALS SCIENCE↗

Beam Path Toxicities to Non-Target Structures During Intensity-Modulated Radiation Therapy for Head and Neck Cancer

Background: Intensity-modulated radiation therapy (IMRT) beams traverse nontarget normal structures not irradiated during three-dimensional conformal RT (3D-CRT) for head and neck cancer (HNC). This study estimates the doses and toxicities to nontarget structures during IMRT. Materials and Methods: Oropharyngeal cancer IMRT and 3D-CRT cases were reviewed. Dose-volume histograms (DVH) were used to evaluate radiation dose to the lip, cochlea, brainstem, occipital scalp, and segments of the mandible. Toxicity rates were compared for 3D-CRT, IMRT alone, or IMRT with concurrent cisplatin. Descriptive statistics and exploratory recursive partitioning analysis were used to estimate dose 'breakpoints' associated with observed toxicities. Results: A total of 160 patients were evaluated for toxicity; 60 had detailed DVH evaluation and 15 had 3D-CRT plan comparison. Comparing IMRT with 3D-CRT, there was significant (p {<=} 0.002) nonparametric differential dose to all clinically significant structures of interest. Thirty percent of IMRT patients had headaches and 40% had occipital scalp alopecia. A total of 76% and 38% of patients treated with IMRT alone had nausea and vomiting, compared with 99% and 68%, respectively, of those with concurrent cisplatin. IMRT had a markedly distinct toxicity profile than 3D-CRT. In recursive partitioning analysis, National Cancer Institute's Common Toxicity Criteria adverse effects 3.0 nausea and vomiting, scalp alopecia and anterior mucositis were associated with reconstructed mean brainstem dose >36 Gy, occipital scalp dose >30 Gy, and anterior mandible dose >34 Gy, respectively. Conclusions: Dose reduction to specified structures during IMRT implies an increased beam path dose to alternate nontarget structures that may result in clinical toxicities that were uncommon with previous, less conformal approaches. These findings have implications for IMRT treatment planning and research, toxicity assessment, and multidisciplinary patient management.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Technical Note: Unsupervised C-arm pose tracking with radiographic fiducial

Purpose: C-arm fluoroscopy reconstruction, such as that used in prostate brachytherapy, requires that the relative poses of the individual C-arm fluoroscopy images must be known prior to reconstruction. Radiographic fiducials can provide excellent C-arm pose tracking, but they need to be segmented in the image. The authors report an automated and unsupervised method that does not require prior segmentation of the fiducial. Methods: The authors compute the individual C-arm poses relative to a stationary radiographic fiducial of known geometry. The authors register a filtered 2D fluoroscopy image of the fiducial to its 3D model by using image intensity alone without prior segmentation. To enhance the C-arm images, the authors investigated a three-step cascade filter and a line enhancement filter. The authors tested the method on a composite fiducial containing beads, straight lines, and ellipses. Ground-truth C-arm pose was provided by a clinically proven method. Results: Using 111 clinical C-arm images and {+-}10 deg. and {+-}10 mm random perturbation around the ground-truth pose, a total of 2775 cases were evaluated. The average rotation and translation errors were 0.62 deg. (STD=0.31 deg.) and 0.72 mm (STD=0.55 mm) for the three-step filter and 0.67 deg. (STD=0.40 deg.) and 0.87 mm (STD=0.27 mm) using the line enhancement filter. Conclusions: The C-arm pose tracking method was sufficiently accurate and robust on human patient data for subsequent 3D implant reconstruction.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

SU-E-J-238: Monitoring Lymph Node Volumes During Radiotherapy Using Semi-Automatic Segmentation of MRI Images

Purpose: Identification and image-based monitoring of lymph nodes growing due to disease, could be an attractive alternative to prophylactic head and neck irradiation. We evaluated the accuracy of the user-interactive Grow Cut algorithm for volumetric segmentation of radiotherapy relevant lymph nodes from MRI taken weekly during radiotherapy. Method: The algorithm employs user drawn strokes in the image to volumetrically segment multiple structures of interest. We used a 3D T2-wturbo spin echo images with an isotropic resolution of 1 mm3 and FOV of 492×492×300 mm3 of head and neck cancer patients who underwent weekly MR imaging during the course of radiotherapy. Various lymph node (LN) levels (N2, N3, N4'5) were individually contoured on the weekly MR images by an expert physician and used as ground truth in our study. The segmentation results were compared with the physician drawn lymph nodes based on DICE similarity score. Results: Three head and neck patients with 6 weekly MR images were evaluated. Two patients had level 2 LN drawn and one patient had level N2, N3 and N4'5 drawn on each MR image. The algorithm took an average of a minute to segment the entire volume (512×512×300 mm3). The algorithm achieved an overall DICE similarity score of 0.78. The time taken for initializing and obtaining the volumetric mask was about 5 mins for cases with only N2 LN and about 15 mins for the case with N2,N3 and N4'5 level nodes. The longer initialization time for the latter case was due to the need for accurate user inputs to separate overlapping portions of the different LN. The standard deviation in segmentation accuracy at different time points was utmost 0.05. Conclusions: Our initial evaluation of the grow cut segmentation shows reasonably accurate and consistent volumetric segmentations of LN with minimal user effort and time.

60 APPLIED LIFE SCIENCES↗

3D printed droplet generation devices for serial femtosecond crystallography enabled by surface coating

The role of surface wetting properties and their impact on the performance of 3D printed microfluidic droplet generation devices for serial femtosecond crystallography (SFX) are reported herein. SFX is a novel crystallography method enabling structure determination of proteins at room temperature with atomic resolution using X-ray free-electron lasers (XFELs). In SFX, protein crystals in their mother liquor are delivered and intersected with a pulsed X-ray beam using a liquid jet injector. Owing to the pulsed nature of the X-ray beam, liquid jets tend to waste the vast majority of injected crystals, which this work aims to overcome with the delivery of aqueous protein crystal suspension droplets segmented by an oil phase. For this purpose, 3D printed droplet generators that can be easily customized for a variety of XFEL measurements have been developed. The surface properties, in particular the wetting properties of the resist materials compatible with the employed two-photon printing technology, have so far not been characterized extensively, but are crucial for stable droplet generation. This work investigates experimentally the effectiveness and the long-term stability of three different surface treatments on photoresist films and glass as models for our 3D printed droplet generator and the fused silica capillaries employed in the other fluidic components of an SFX experiment. Finally, the droplet generation performance of an assembly consisting of the 3D printed device and fused silica capillaries is examined. Stable and reproducible droplet generation was achieved with a fluorinated surface coating which also allowed for robust downstream droplet delivery. Experimental XFEL diffraction data of crystals formed from the large membrane protein complex photosystem I demonstrate the full compatibility of the new injection method with very fragile membrane protein crystals and show that successful droplet generation of crystal-laden aqueous droplets intersected by an oil phase correlates with increased crystal hit rates.

47 OTHER INSTRUMENTATION↗

Quant-CT: Segmenting and Quantifying Computed Tomography

Quant-CT is currently a plugin to ImageJ, designed as a Java-class that provides control mechanism for the user to choose volumes of interest within porous material, followed by the selection of image subsamples for automated tuning of parameters for filters and classifiers, and finally measurement of material geometry, porosity, and visualization. Denoising is mandatory before any image interpretation, and we implemented a new 3D java code that performs bilateral filtering of data. Segmentation of the dense material is essential before any quantifications about geological sample structure, and we invented new schemes to deal with over segmentation when using statistical region merging algorithm to pull out grains that compose imaged material. It make uses of ImageJ API and other standard and thirty-party APIs. Quant-CT conception started in 2011 under Scidac-e sponsor, and details of the first prototype were documented in publications below. While it is used right now for microtomography images, it can potentially be used by anybody with 3D image data obtained by experiment or produced by simulation.

Daniela Ushizima, Wes Bethel↗

ABWR start-up test analysis using BWR core simulator with three-dimensional direct response matrix method

The ABWR start-up test analysis has been done with the BWR core simulator using the three--dimensional direct response matrix (3D-DRM) method. The Monte Carlo code VMONT made the sub-response matrices for the 3D-DRM method. Each boundary surface was subdivided by 4 x 4 for transverse segments, by 4 for angular segments and by 4 for axial zones in a node. For the calculation speedup, the 3D-DRM code used the divided sub-response matrices data set. The code used the MPI and OpenMP for the parallelized method. The median value is set as the average critical eigenvalues. The changes from the maximum value to the minimum value are 0.34 %{Delta}k with the spectral history method and 0.40 %{Delta}k without it, and the respective standard deviations were 0.12 % and 0.14 %. Using the spectral history method decreased the variation by 0.06 %{Delta}k. The root mean square differences of the axial power distribution were about 6 % between the analysis results and the plant data. Using the currents which converged in the previous exposure step reduced the number of iterations when the CR pattern changed only slightly. The averaged calculation time for each exposure step was about 5 hours on 12 PC Linux cluster servers with Core 2 Quad 3 GHz. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Data-driven Mapping of the Mouse Connectome: The utility of transfer learning to improve the performance of deep learning models performing axon segmentation on light-sheet microscopy images

Light sheet microscopy has made possible the high temporal and spatial 3D imaging of both fixed and live biological tissue, with samples as large as the entire mouse brain. However, segmentation and quantification of that data remains a time-consuming manual process. Machine learning methods promise the possibility of automating this process. This study seeks to advance the performance of prior models through the application of refinements such as transfer learning.

59 BASIC BIOLOGICAL SCIENCES↗

Spatio-Temporal Video Segmentation with Shape Growth or Shrinkage Constraint

We propose a new method for joint segmentation of monotonously growing or shrinking shapes in a time sequence of noisy images. The task of segmenting the image time series is expressed as an optimization problem using the spatio-temporal graph of pixels, in which we are able to impose the constraint of shape growth or of shrinkage by introducing monodirectional infinite links connecting pixels at the same spatial locations in successive image frames. The globally optimal solution is computed with a graph cut. The performance of the proposed method is validated on three applications: segmentation of melting sea ice floes and of growing burned areas from time series of 2D satellite images, and segmentation of a growing brain tumor from sequences of 3D medical scans. In the latter application, we impose an additional intersequences inclusion constraint by adding directed infinite links between pixels of dependent image structures.

Segmentation↗

Effect of thermally stable Cu- and Mg-rich aluminides on the high temperature strength of an AlSi12CuMgNi alloy

The internal architecture of an AlSi12CuMgNi piston alloy, revealed by synchrotron tomography, consists of three dimensional interconnected hybrid networks of Cu-rich aluminides, Mg-rich aluminides and eutectic/primary Si embedded in an α-Al matrix. The strength at room temperature and at 300°C is studied as a function of solution treatment time at 490°C and compared with results previously reported for an AlSi12Ni alloy. The addition of 1 wt% Cu and 1 wt% Mg to AlSi12CuMgNi increases the room temperature strength by precipitation hardening while the strength at 300°C is similar for both alloys in as-cast condition. The strength of AlSi12CuMgNi decreases with solution treatment time and stabilizes at 4 h solution treatment. The effect of solution treatment time on the strength of the AlSi12CuMgNi alloy is less pronounced than for the AlSi12Ni alloy both at room temperature and at 300°C. - Highlights: • The 3D microstructure of AlSi12CuMgNi is revealed by synchrotron tomography. • An imaging analysis procedure to segment phases with similar contrasts is presented. • 1 wt% Cu and Mg results in the formation of 3D networks of rigid phases. • AlSi12CuMgNi is stronger than AlSi12Ni owing to the stability of the 3D networks.

36 MATERIALS SCIENCE↗

SU-C-9A-01: Parameter Optimization in Adaptive Region-Growing for Tumor Segmentation in PET

Purpose: To design a reliable method to determine the optimal parameter in the adaptive region-growing (ARG) algorithm for tumor segmentation in PET. Methods: The ARG uses an adaptive similarity criterion m - fσ ≤ I-PET ≤ m + fσ, so that a neighboring voxel is appended to the region based on its similarity to the current region. When increasing the relaxing factor f (f ≥ 0), the resulting volumes monotonically increased with a sharp increase when the region just grew into the background. The optimal f that separates the tumor from the background is defined as the first point with the local maximum curvature on an Error function fitted to the f-volume curve. The ARG was tested on a tumor segmentation Benchmark that includes ten lung cancer patients with 3D pathologic tumor volume as ground truth. For comparison, the widely used 42% and 50% SUVmax thresholding, Otsu optimal thresholding, Active Contours (AC), Geodesic Active Contours (GAC), and Graph Cuts (GC) methods were tested. The dice similarity index (DSI), volume error (VE), and maximum axis length error (MALE) were calculated to evaluate the segmentation accuracy. Results: The ARG provided the highest accuracy among all tested methods. Specifically, the ARG has an average DSI, VE, and MALE of 0.71, 0.29, and 0.16, respectively, better than the absolute 42% thresholding (DSI=0.67, VE= 0.57, and MALE=0.23), the relative 42% thresholding (DSI=0.62, VE= 0.41, and MALE=0.23), the absolute 50% thresholding (DSI=0.62, VE=0.48, and MALE=0.21), the relative 50% thresholding (DSI=0.48, VE=0.54, and MALE=0.26), OTSU (DSI=0.44, VE=0.63, and MALE=0.30), AC (DSI=0.46, VE= 0.85, and MALE=0.47), GAC (DSI=0.40, VE= 0.85, and MALE=0.46) and GC (DSI=0.66, VE= 0.54, and MALE=0.21) methods. Conclusions: The results suggest that the proposed method reliably identified the optimal relaxing factor in ARG for tumor segmentation in PET. This work was supported in part by National Cancer Institute Grant R01 CA172638; The dataset is provided by AAPM TG211.

60 APPLIED LIFE SCIENCES↗

Segmental Refinement: A Multigrid Technique for Data Locality

We investigate a technique - segmental refinement (SR) - proposed by Brandt in the 1970s as a low memory multigrid method. The technique is attractive for modern computer architectures because it provides high data locality, minimizes network communication, is amenable to loop fusion, and is naturally highly parallel and asynchronous. The network communication minimization property was recognized by Brandt and Diskin in 1994; we continue this work by developing a segmental refinement method for a finite volume discretization of the 3D Laplacian on massively parallel computers. An understanding of the asymptotic complexities, required to maintain textbook multigrid efficiency, are explored experimentally with a simple SR method. A two-level memory model is developed to compare the asymptotic communication complexity of a proposed SR method with traditional parallel multigrid. Performance and scalability are evaluated with a Cray XC30 with up to 64K cores. We achieve modest improvement in scalability from traditional parallel multigrid with a simple SR implementation.

3D Laplacian↗

WE-F-16A-01: Commissioning and Clinical Use of PC-ISO for Customized, 3D Printed, Gynecological Brachytherapy Applicators

Purpose: (1) Evaluate the safety and radiation attenuation properties of PCISO, a bio-compatible, sterilizable 3D printing material by Stratasys, (2) establish a method for commissioning customized multi- and single-use 3D printed applicators, (3) report on use of customized vaginal cylinders used to treat a series of serous endometrial cancer patient. Methods: A custom film dosimetry apparatus was designed to hold a Gafchromic radio film segment between two blocks of PC-ISO and 3D-printed using a Fortus 400mc (StrataSys). A dose plan was computed using 13 dwell positions at 2.5 mm spacing and normalized to 1500 cGy at 1 cm. Film exposure was compared to control tests in only air and only water. The average Hounsfield Unit (HU) was computed and used to verify water equivalency. For the clinical use cases, the physician specifies the dimensions and geometry of a custom applicator from which a CAD model is designed and printed. Results: The doses measured from the PC-ISO Gafchromic film test were within 1% of the dose measured in only water between 1cm and 6cm from the channel. Doses increased 7–4% measured in only air. HU range was 11–43. The applicators were sterilized using the Sterrad system multiple times without damage. As of submission 3 unique cylinders have been designed, printed, and used in the clinic. A standardizable workflow for commissioning custom 3D printed applicators was codified and will be reported. Conclusions: Quality assurance (QA) evaluation of the PC-ISO 3D-printing material showed that PC-ISO is a suitable material for a gynecological brachytherapy vaginal cylinder in a clinical setting. With the material commissioning completed, if the physician determines that a better treatment would Result, a customized design is fabricated with limited additional QA necessary. Although this study was specific to PC-ISO, the same setup can be used to evaluate other 3D-printing materials.

60 APPLIED LIFE SCIENCES↗

Automated segmentation of dental CBCT image with prior-guided sequential random forests

Purpose: Cone-beam computed tomography (CBCT) is an increasingly utilized imaging modality for the diagnosis and treatment planning of the patients with craniomaxillofacial (CMF) deformities. Accurate segmentation of CBCT image is an essential step to generate 3D models for the diagnosis and treatment planning of the patients with CMF deformities. However, due to the image artifacts caused by beam hardening, imaging noise, inhomogeneity, truncation, and maximal intercuspation, it is difficult to segment the CBCT. Methods: In this paper, the authors present a new automatic segmentation method to address these problems. Specifically, the authors first employ a majority voting method to estimate the initial segmentation probability maps of both mandible and maxilla based on multiple aligned expert-segmented CBCT images. These probability maps provide an important prior guidance for CBCT segmentation. The authors then extract both the appearance features from CBCTs and the context features from the initial probability maps to train the first-layer of random forest classifier that can select discriminative features for segmentation. Based on the first-layer of trained classifier, the probability maps are updated, which will be employed to further train the next layer of random forest classifier. By iteratively training the subsequent random forest classifier using both the original CBCT features and the updated segmentation probability maps, a sequence of classifiers can be derived for accurate segmentation of CBCT images. Results: Segmentation results on CBCTs of 30 subjects were both quantitatively and qualitatively validated based on manually labeled ground truth. The average Dice ratios of mandible and maxilla by the authors’ method were 0.94 and 0.91, respectively, which are significantly better than the state-of-the-art method based on sparse representation (p-value < 0.001). Conclusions: The authors have developed and validated a novel fully automated method for CBCT segmentation.

61 RADIATION PROTECTION AND DOSIMETRY↗

Computerized analysis of coronary artery disease: Performance evaluation of segmentation and tracking of coronary arteries in CT angiograms

Purpose: The authors are developing a computer-aided detection system to assist radiologists in analysis of coronary artery disease in coronary CT angiograms (cCTA). This study evaluated the accuracy of the authors’ coronary artery segmentation and tracking method which are the essential steps to define the search space for the detection of atherosclerotic plaques. Methods: The heart region in cCTA is segmented and the vascular structures are enhanced using the authors’ multiscale coronary artery response (MSCAR) method that performed 3D multiscale filtering and analysis of the eigenvalues of Hessian matrices. Starting from seed points at the origins of the left and right coronary arteries, a 3D rolling balloon region growing (RBG) method that adapts to the local vessel size segmented and tracked each of the coronary arteries and identifies the branches along the tracked vessels. The branches are queued and subsequently tracked until the queue is exhausted. With Institutional Review Board approval, 62 cCTA were collected retrospectively from the authors’ patient files. Three experienced cardiothoracic radiologists manually tracked and marked center points of the coronary arteries as reference standard following the 17-segment model that includes clinically significant coronary arteries. Two radiologists visually examined the computer-segmented vessels and marked the mistakenly tracked veins and noisy structures as false positives (FPs). For the 62 cases, the radiologists marked a total of 10191 center points on 865 visible coronary artery segments. Results: The computer-segmented vessels overlapped with 83.6% (8520/10191) of the center points. Relative to the 865 radiologist-marked segments, the sensitivity reached 91.9% (795/865) if a true positive is defined as a computer-segmented vessel that overlapped with at least 10% of the reference center points marked on the segment. When the overlap threshold is increased to 50% and 100%, the sensitivities were 86.2% and 53.4%, respectively. For the 62 test cases, a total of 55 FPs were identified by radiologist in 23 of the cases. Conclusions: The authors’ MSCAR-RBG method achieved high sensitivity for coronary artery segmentation and tracking. Studies are underway to further improve the accuracy for the arterial segments affected by motion artifacts, severe calcified and noncalcified soft plaques, and to reduce the false tracking of the veins and other noisy structures. Methods are also being developed to detect coronary artery disease along the tracked vessels.

60 APPLIED LIFE SCIENCES↗

Expanding NeMO-Net Machine Learning Capabilities for Citizen Science

NASA NeMO-Net, the neural multi-modal observation and training network for global coral reef assessment, is an open-source deep convolutional neural network and interactive active learning training software aiming to accurately assess the present and past dynamics of coral reef ecosystems through determination of percent living cover and morphology as well as mapping of spatial distribution. We present an interactive citizen science video game, released this April, for desktop and iOS devices where users interactively label morphology classifications over mm-scale 3D coral reef imagery captured using diver photomosaic imagery, the UAV enabled NASA FluidCam instrument, and satellite datasets. To date, the application has had over 40,000 downloads and over60,000 unique coral reef classifications, each filtered through a user-based rating and expert evaluation system. We also present results from NeMO-Net’s convolutional neural network (CNN) models used to semantically segment 2D satellite imagery as well as projections of 3D coral reconstructions using user input data as training datasets. Fusing datasets using machine learning from multiple remote sensing platforms presents novel methodologies for assessing the health of coral ecosystems, which are critically endangered by a changing climate. In partnering with Mission Blue, the National Oceanic and Atmospheric Administration (NOAA), and the Living Oceans Foundation (LOF), NeMO-Net leverages an international consortium of subject matter experts to provide both proper training for citizen scientists and the generation of a labeled datasets to ingest into machine learning algorithms for global coral reef identification.

NeMO-Net↗

Comparative Assessment of U-Net-Based Deep Learning Models for Segmenting Microfractures and Pore Spaces in Digital Rocks

Segmentation of high-resolution X-ray microcomputed tomography (µCT) images is crucial in digital rock physics (DRP), affecting the characterization and analysis of microscale phenomena in the porous media. The complexity of geological structures and nonideal scanning conditions pose significant challenges to conventional image segmentation approaches. Motivated by the recent increasing popularity of deep learning (DL) techniques in image processing, this work undertakes a comparative study of DL models, specifically U-Net and its variants, for segmenting multiple targets with distinguished features in digital rocks, including discrete fracture networks (DFNs), pore spaces, and solid rock. Particularly, DFNs have a smaller volumetric fraction over others, bringing in a substantial challenge of imbalanced segmentation. The primary focus is to evaluate the architecture and feature enhancement strategies of various DL models, including U-Net, attention U-Net, residual U-Net, U-Net++, and residual U-Net++. The models were designed as 2.5D, utilizing a central 2D image and its two adjacent upper and lower 2D images as input to provide a pseudo-3D context. In addition, because the ground truth of segmentation was unknown for real-world digital rocks, we created a benchmark data set following the inverse operations of segmentation. The data synthesis started from the label images (i.e., solid rock, pore spaces, and DFNs), followed by simulating partial volume blurring, adding random background noise, and introducing ring artifacts to mimic real raw X-ray µCT images. The data set, which included various rock types (i.e., sandstone and artificial data), scanning resolution, and magnitudes of noise and artifacts, was divided into training and testing data sets with a 90% and 10% ratio, respectively. Moreover, in addition to the conventional pixel-wise evaluation metrics, the physics-based metric of the lattice-Boltzmann method (LBM) simulated permeability provided more comprehensive assessments. The results demonstrated that the residual connections, nested architectures, and redesigned skip connections contribute to the model performance and give the residual U-Net++ the highest accuracy. The improvements were mainly on the boundaries and small targets, especially the DFNs, which dominate the interconnectivity and therefore affect the permeability greatly. This study also rigorously evaluated the efficiency and generalization of each model, demonstrating that the sophisticated architectures achieved excellent practicability and maintained robust performance on completely unseen data, ensuring their suitability for diverse and challenging DRP applications.

58 GEOSCIENCES↗

Expanding NeMO-Net Machine Learning Capabilities for Citizen Science

NASA NeMO-Net, the neural multi-modal observation and training network for global coral reef assessment, is an open-source deep convolutional neural network and interactive active learning training software aiming to accurately assess the present and past dynamics of coral reef ecosystems through determination of percent living cover and morphology as well as mapping of spatial distribution. We present an interactive citizen science video game, released this April, for desktop and iOS devices where users interactively label morphology classifications over mm-scale 3D coral reef imagery captured using diver photomosaic imagery, the UAV enabled NASA FluidCam instrument, and satellite datasets. To date, the application has had over 40,000 downloads and over 60,000 unique coral reef classifications, each filtered through a user-based rating and expert evaluation system. We also present results from NeMO-Net’s convolutional neural network (CNN) models used to semantically segment 2D satellite imagery as well as projections of 3D coral reconstructions using user input data as training datasets. Fusing datasets using machine learning from multiple remote sensing platforms presents novel methodologies for assessing the health of coral ecosystems, which are critically endangered by a changing climate. In partnering with Mission Blue, the National Oceanic and Atmospheric Administration (NOAA), and the Living Oceans Foundation (LOF), NeMO-Net leverages an international consortium of subject matter experts to provide both proper training for citizen scientists and the generation of a labeled datasets to ingest into machine learning algorithms for global coral reef identification.

NeMO-Net↗