Distance Field Research Articles

Over 20‐Year Global Magnetohydrodynamic Simulation of Earth's Magnetosphere

AbstractWe present our approach to modeling over 20 years of the solar wind‐magnetosphere‐ionosphere system using version 5 of the Grand Unified Magnetosphere‐Ionosphere Coupling Simulation (GUMICS‐5). As input we use 16‐s resolution magnetic field and 1‐min plasma measurements by Advanced Composition Explorer satellite from 1998 to 2020. The modeled interval is divided into 28 hr simulations including 4 hr overlap. We use maximum magnetospheric resolution of 0.5 Earth radii (RE) up to about 15 RE from Earth and decreasing resolution further away. In the ionosphere we use a maximum resolution of approximately 100 km poleward of ±58° magnetic latitude and decreasing resolution toward equator. With respect to previous version GUMICS‐4, we have parallelized the magnetosphere of GUMICS‐5 using the Message Passing Interface and have made several improvements which have for example, decreased its numerical diffusion. In total we have performed over 8,000 simulations which have produced over 10,000,000 ionospheric files and 2,000,000 magnetospheric files requiring over 100 TB of disk space. We compare these results to several empirical models and geomagnetic indices derived from ground magnetic field measurements. GUMICS‐5 reproduces observed solar cycle trends in magnetopause stand‐off distance and magnetospheric lobe field strength but consistency in plasma sheet pressure and ionospheric cross‐polar cap potential is lower. Comparisons with geomagnetic indices show better results for Kp index than for auroral electrojet index. Our extensive results can serve, for example, as a foundation for combined physics‐based and black‐box approach to real‐time prediction of near‐Earth space, or as input to other physics‐based models of the inner magnetosphere, upper and middle atmosphere, etc.

Nasal-Skull Base Fast Path Extraction and Automatic Navigation

In the nasal-skull base virtual endoscope, the existing navigation path extraction algorithms have low operation efficiency, and automatic navigation algorithm is easy to cause violent rotation of image. Therefore, a center path extraction algorithm based on navigation area pre-planning and an automatic navigation algorithm based on adaptive speed and rotation are proposed. The former improves the distance transformation method. The navigation area is pre-planned based on the fast expanding random tree and calculate the boundary distance field map, and construct the maximum cost tree to extract the center path. The latter improves the DoTween method. The camera moving speed is adjusted according to the length of the shortest line segment in the path, and the predicted steering distance is adjusted according to the cavity diameter at the vertex in the path. Experiments are conducted with nasal models from DICOM. The results show that the path extraction speed of the center path extraction algorithm based on navigation area pre-planning method is about 49 times faster than the distance transformation method, and the automatic navigation algorithm of adaptive speed and rotation is more natural at the vertices comparing with methods such as DoTween.

Satisfaction of female students at the college of education with distance field training in light of some demographic variables

The purpose of this research is to identify female students’ satisfaction level with distance field training at the College of Education in light of some of the demographic variables. A questionnaire on students' satisfaction with distance field training was applied to a sample of 102 female students (mean age = 20.029, SD = 0.813). Data were analysed by mean, standard deviations, frequencies and percentages; and t-test was used for differences between independent samples. The research findings revealed that the degree of satisfaction of the female students with distance field training was high. Besides, there were also no differences in their satisfaction level due to the academic specialisation (special education–kindergarten) and the academic average (high, medium and low), while there were significant differences due to the academic level variables (seventh–eighth), in favour of the eighth level. The research results were discussed in light of the previous related literature and further research suggestions were provided to researchers and stakeholders. Keywords: Demographic variables, distance field training, College of Education, satisfaction.

A level set-based procedure for the cohesive modeling of yarn–yarn contacts in woven composite RVEs

Decohesion processes involved in debonding between yarns and matrix or sliding between contacting yarns constitute important failure mechanisms in woven composites. The geometry of the yarn architecture however makes it difficult to avoid interpenetrations between neighboring yarns in their geometrical description. This motivated recent works using a level set-based methodology to generate smooth geometries without interpenetration for generated Representative Volume Elements (RVEs) and experimental image-based samples. Such tools require introducing gaps between contacting yarns, which may be debatable when taken too large or may be computationally costly for thin gap, locally dictating the mesh size. To avoid this, the yarns contacting surfaces can be discretized using interface elements. Yet, identifying such contacting surfaces is complex. An implicit geometry description based on distance fields is used here to extract a geometry that supports cohesive zones, identifying the interface of yarn–yarn contacts and eliminating the gaps using manipulations on distance fields of the individual yarns. This allows generating conforming tetrahedral meshes with inserted cohesive elements in the contact areas. The methodology is shown to reduce the size of finite element models of woven composite RVEs by over two orders of magnitude, thereby paving the way towards affordable damage simulations on such RVEs.

Omnipose: a high-precision morphology-independent solution for bacterial cell segmentation

Advances in microscopy hold great promise for allowing quantitative and precise measurement of morphological and molecular phenomena at the single-cell level in bacteria; however, the potential of this approach is ultimately limited by the availability of methods to faithfully segment cells independent of their morphological or optical characteristics. Here, we present Omnipose, a deep neural network image-segmentation algorithm. Unique network outputs such as the gradient of the distance field allow Omnipose to accurately segment cells on which current algorithms, including its predecessor, Cellpose, produce errors. We show that Omnipose achieves unprecedented segmentation performance on mixed bacterial cultures, antibiotic-treated cells and cells of elongated or branched morphology. Furthermore, the benefits of Omnipose extend to non-bacterial subjects, varied imaging modalities and three-dimensional objects. Finally, we demonstrate the utility of Omnipose in the characterization of extreme morphological phenotypes that arise during interbacterial antagonism. Our results distinguish Omnipose as a powerful tool for characterizing diverse and arbitrarily shaped cell types from imaging data.

Online Distance Field Priors for Gaussian Process Implicit Surfaces

Gaussian process (GP) implicit surface models provide environment and object representations which elegantly address noise and uncertainty while remaining sufficiently flexible to capture complex geometry. However, GP models quickly become intractable as the size of the observation set grows—a trait which is difficult to reconcile with the rate at which modern range sensors produce data. Furthermore, naïve applications of GPs to implicit surface models allocate model resources uniformly, thus using precious resources to capture simple geometry. In contrast to prior work addressing these challenges though model sparsification, spatial partitioning, or ad-hoc filtering, we propose introducing model bias online through the GP’s mean function. We achieve more accurate distance fields using smaller models by creating a distance field prior from features which are easy to extract and have analytic distance fields. In particular, we demonstrate this approach using linear features. We show the proposed distance field halves model size in a 2D mapping task using data from a SICK S300 sensor. When applied to a single 3D scene from the TUM RGB-D SLAM dataset, we achieve a fivefold reduction in model size. Our proposed prior results in more accurate GP implicit surfaces, while allowing existing models to function in larger environments or with larger spatial partitions due to reduced model size.

Vector field-based curved layer slicing and path planning for multi-axis printing

• A new vector field-based curved layer slicing and printing path planning method is proposed. • A quantitative optimization model is established to find the optimal filament orientation and printing orientation vector fields. • Multi-factors including support-free requiement, collision-free condition, and mechanical performance are considered. • A continuous printing path planning algorithm is developed to reduce the number of nozzle retractions. The emergence of multi-axis printing systems provides a new solution to print complex parts. However, the current process planning methods fail to balance the support-free requirement, the collision-free condition, as well as the mechanical performance at the same time when printing parts with complicated features. This paper proposes a new curved layer slicing and continuous printing path planning method that considers these factors. The proposed method makes use of two mutually orthogonal unit vector fields embedded on the tetrahedral mesh of the part: the filament orientation vector field and the printing orientation vector field, which indicate the filament orientation and the nozzle orientation at each position during the printing process respectively. A quantitative optimization model is established to compute the two optimal unit vector fields. Then a monotonically increasing scalar distance field is generated by integrating along the optimal printing orientation vector field, and the isosurfaces of the distance field can be used to naturally slice the part into curved layers. Finally, at each isosurface, a continuous printing path is planned along the optimal filament orientation field by interpolating isolines of a surface embedded scalar distance field. Both the simulation and physical experiments were conducted to confirms the effectiveness of the proposed method and its advantage over the existing methods in balance different factors including support-free condition, collision-free requirement, and mechanical performance.

SDFEst: Categorical Pose and Shape Estimation of Objects From RGB-D Using Signed Distance Fields

Rich geometric understanding of the world is an important component of many robotic applications such as planning and manipulation. In this paper, we present a modular pipeline for pose and shape estimation of objects from RGB-D images given their category. The core of our method is a generative shape model, which we integrate with a novel initialization network and a differentiable renderer to enable 6D pose and shape estimation from a single or multiple views. We investigate the use of discretized signed distance fields as an efficient shape representation for fast analysis-by-synthesis optimization. Our modular framework enables multi-view optimization and extensibility. We demonstrate the benefits of our approach over state-of-the-art methods in several experiments on both synthetic and real data. We open-source our approach at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/roym899/sdfest</uri> .

A generic framework for geotechnical subsurface modeling with machine learning

This study introduces a generic framework for geotechnical subsurface modeling, which accounts for spatial autocorrelation with local mapping machine learning (ML) methods. Instead of using XY coordinate fields directly as model input, a series of autocorrelated geotechnical distance fields (GDFs) is designed to enable the ML models to infer the spatial relationship between the sampled locations and unknown locations. The whole framework using GDF with ML methods is named GDF-ML. This framework is purely data-driven which avoids the tedious work in the scale of fluctuations (SOFs) estimating and data detrending in the conventional spatial interpolation methods. Six local mapping ML methods (extra trees (ETs), gradient boosting (GB), extreme gradient boosting (XGBoost), random forest (RF), general regression neural network (GRNN) and k -nearest neighbors (KNN)) are compared in the GDF-ML framework. The results show that the GDFs are better than the conventional XY coordinate fields based ML methods in both accuracy and spatial continuity. GDF-ML is flexible which can be applied to high-dimensional, multi-variable and incomplete datasets. Among these six methods, GDF with ET method (GDF-ET) clearly shows the best accuracy and best spatial continuity. The proposed GDF-ET method can provide a fast and accurate interpretation of the soil property profile. Sensitivity analysis shows that this method is applicable to very small training dataset size. The associated statistical uncertainty can also be quantified so that the reliability of the subsurface modeling results can be estimated objectively and explicitly. The uncertainty results clearly show that the prediction becomes more accurate when more sampled data are available. • A generic framework is proposed for geotechnical subsurface modeling with machine learning. • The framework is data-driven and does not require pre-stratification of subsurface properties. • Geotechnical distance fields (GDFs) can be used as input for machine learning models. • Interpolation performances of six local mapping machine learning methods are compared. • The interpolation uncertainties associated with the extra tree model can be quantified.

Using R-Functions to Control the Shape of Soft Robots

In this letter, we introduce a new approach for soft robot shape formation and morphing using approximate distance fields. The method uses concepts from constructive solid geometry, R-functions, to construct an approximate distance function to the boundary of a domain in <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\Re ^{d}$</tex-math></inline-formula> . The gradients of the R-functions can then be used to generate control algorithms for shape formation tasks for soft robots. By construction, R-functions are smooth and convex everywhere, possess precise differential properties, and easily extend from <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\Re ^{2}$</tex-math></inline-formula> to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\Re ^{3}$</tex-math></inline-formula> if needed. Furthermore, R-function theory provides a straightforward method to creating composite distance functions for any desired shape by combining subsets of distance functions. The process is highly efficient since the shape description is an analytical expression, and in this sense, it is better than competing control algorithms such as those based on potential fields. Although the method could also apply to swarm robots, in this letter it is applied to soft robots to demonstrate shape formation and morphing in 2-D (simulation and experimentation) and 3-D (simulation).

Detection of Localization Failures Using Markov Random Fields With Fully Connected Latent Variables for Safe LiDAR-Based Automated Driving

Most of the recent automated driving systems assume the accurate functioning of localization. Unanticipated errors cause localization failures and result in failures in automated driving. An exact localization failure detection is necessary to ensure safety in automated driving; however, detection of the localization failures is challenging because sensor measurement is assumed to be independent of each other in the localization process. Owing to the assumption, the entire relation of the sensor measurement is ignored. Consequently, it is difficult to recognize the misalignment between the sensor measurement and the map when partial sensor measurement overlaps with the map. This paper proposes a method for the detection of localization failures using Markov random fields with fully connected latent variables. The full connection enables to take the entire relation into account and contributes to the exact misalignment recognition. Additionally, this paper presents localization failure probability calculation and efficient distance field representation methods. We evaluate the proposed method using two types of datasets. The first dataset is the SemanticKITTI dataset, whereby four methods are compared with the proposed method. The comparison results reveal that the proposed method achieves the most accurate failure detection. The second dataset is created based on log data acquired from the demonstrations that we conducted in Japanese public roads. The dataset includes several localization failure scenes. We apply the failure detection methods to the dataset and confirm that the proposed method achieves exact and immediate failure detection.

PlaneSDF-Based Change Detection for Long-Term Dense Mapping

The ability to process environment maps across multiple sessions is critical for robots operating over extended periods of time. Specifically, it is desirable for autonomous agents to detect changes amongst maps of different sessions so as to gain a conflict-free understanding of the current environment. In this paper, we look into the problem of change detection based on a novel map representation, dubbed Plane Signed Distance Fields (PlaneSDF), where dense maps are represented as a collection of planes and their associated geometric components in SDF volumes. Given point clouds of the source and target scenes, we propose a three-step PlaneSDF-based change detection approach: (1) PlaneSDF volumes are instantiated within each scene and registered across scenes using plane poses; 2D height maps and object maps are extracted per volume via height projection and connected component analysis. (2) Height maps are compared and intersected with the object map to produce a 2D change location mask for changed object candidates in the source scene. (3) 3D geometric validation is performed using SDF-derived features per object candidate for change mask refinement. We evaluate our approach on both synthetic and real-world datasets and demonstrate its effectiveness via the task of changed object detection. Supplementary video: https://youtu.be/oh-MQPWTwZI

Acoustic Resonance Frequencies of Underwater Toroidal Bubbles.

Underwater bubbles display an acoustic resonance frequency close to spherical ones. In order to obtain a resonance significantly deviating from the spherical case, we stabilize bubbles in toroidal frames, resulting in bubbles which can be slender while still compact. For thin tori the resonance frequency increases greatly. Between a pair of bubble rings, we can achieve a flat acoustic pressure field for a critical distance between rings, a condition reminiscent of Helmholtz coils in magnetostatics. This opens the possibility to shape the acoustic field using long tunnels of rings.

On the Contribution of Wall Distance Fields to the Adjoint of a RANS Model

The adjoint method has been extensively used in many areas of CFD such as gradient-based shape optimisation. When utilising the RANS equations for simulating turbulent flows, the adjoint method requires a scrupulous differentiation of the RANS equations, including the wall distance contribution. This can be a challenging task and a potential source of inaccuracy for functional sensitivities if not correctly executed. This paper presents a formulation for including the contribution of an equation-based wall distance model to the discrete adjoint of a RANS model. The proposed formulation is tested in a gradient-based optimisation scenario and the effects of the wall distance adjoint fields on the functional sensitivities are investigated. Neglecting the contribution of the wall distance adjoint yields an error in the functional sensitivities with respect to volume mesh nodes. Including the wall distance adjoint restores the accuracy of the functional sensitivities yielding better convergence of the design optimisation.

Prediction and Generation of 3D Functional Scene Based on Relation Graph

Functionality analysis of 3D shapes, as a means to understand and manipulate 3D environments, plays an important role in artificial intelligence. Human can predict the functionality of the object based on knowledge and experience without any surroundings. To this end, a method which uses relation graphs guide the generation of 3D functional scenes of a given object is proposed. The method consists of two steps: relation graph generation and scene generation. In the relation graph generation phase, a deep graph convolution generative network is used to predict a relation graph for the given object, which encodes the central and interacting objects as nodes and spatial relationships between objects as edges. In the scene generation phase, the generated graph and the given central object is taken as input to output the shapes and locations of interacting objects. Proposed method is tested on the interaction scene dataset for scene generation and compared with PG-DNN and PQ-NET. Chamfer distance and light field distance are used as evaluation indicators. Experimental results show that proposed method outperforms the state-of-the-art methods in functional scene generation and high-resolution functional scenes can be generated by using implicit representation of shapes.

Topology-controllable Implicit Surface Reconstruction Based on Persistent Homology

In geometric design, reconstructing an implicit surface with high-quality geometry and expected topology from point clouds has been a challenging problem. However, traditional topological invariants, such as Betti numbers, are not easily used in controlling topology in implicit surface reconstruction. Persistent homology provides a quantitative measurement, i.e. , persistence diagram (PD), and allows tracking of pairs of key points that generate and destroy topological holes. In this study, we propose a topology-controllable implicit surface reconstruction method from point clouds based on persistent homology. Specifically, given a point cloud with normals, the signed distance field is first constructed, and then a B-spline function represented distance function is generated by fitting the signed distance values through progressive iterative approximation. By designing a topological target function using the persistent pairs in PDs according to topological priors, the control coefficients of the B-spline function are optimized to extract an implicit surface, i.e. , an iso-surface of the B-spline function, with expected topology. Experiments show that the proposed method can reconstruct surfaces with higher topological quality than other reconstruction methods. Moreover, the proposed method can be used to edit the topology of an implicit surface. • The topology-controllable framework for implicit B-spline surface reconstruction improves geometric quality and ensures the desired topology of the extracted surface. • Principles of designing the topological target function for implicit surface reconstruction are proposed. • Reconstruction on sparse point clouds and the application of topology editing are explored.

Self-calibration method of optical fiber shape sensor placement angle deviation based on GA

A self-calibration method based on genetic algorithm (GA) is used to achieve the placement angle correction of optical fiber shape sensor. This method can automatically calibrate the placement angle of the sensor directly, with no need of complex experimental calibration process. The method is verified by simulation and experiment under different bending directions and curvature, the maximum error of curvature and bending directions is decreased by 9.6%, 7.0°, and 11.3%, 7.2° by the self-calibration method in the simulation and experiment, respectively. This method provides a new correction method for optical fiber shape sensing, which is more convenient and effective than the traditional calibration method, especially in the field of long distance and large structure, which needs a large number of measuring points and needs to be laid and calibrated in the scene.

Neuronavigation maximizes accuracy and precision in TMS positioning: Evidence from 11,230 distance, angle, and electric field modeling measurements

BackgroundResearchers and clinicians have traditionally relied on elastic caps with markings to reposition the transcranial magnetic stimulation (TMS) coil between trains and sessions. Newer neuronavigation technology co-registers the patient's head and structural magnetic resonance imaging (MRI) scan, providing the researcher with real-time feedback about how to adjust the coil to be on-target. However, there has been no head to head comparison of accuracy and precision across treatment sessions. Objective/Hypothesis: In this two-part study, we compared elastic cap and neuronavigation targeting methodologies on distance, angle, and electric field (E-field) magnitude values. MethodsIn 42 participants receiving up to 50 total accelerated rTMS sessions in 5 days, we compared cap and neuronavigation targeting approaches in 3408 distance and 6816 angle measurements. In Experiment 1, TMS administrators saved an on-target neuronavigation location at Beam F3, which served as the landmark for all other measurements. Next, the operators placed the TMS coil based on cap markings or neuronavigation software to measure the distance and angle differences from the on-target sample. In Experiment 2, we saved each XYZ coordinate of the TMS coil from cap and neuronavigation targeting in 12 participants to compare the E-field magnitude differences at the cortical prefrontal target in 1106 cap and neuronavigation models. ResultsCap targeting was significantly off-target for distance, placing the coil an average of 10.66 mm off-target (Standard error of the mean; SEM = 0.19 mm) compared to 0.3 mm (SEM = 0.03 mm) for neuronavigation (p < 0.0001). Cap targeting also significantly deviated for angles off-target, averaging 7.79 roll/pitch degrees (SEM = 1.07°) off-target and 5.99 yaw degrees (SEM = 0.12°) off-target; in comparison, neuronavigation targeting positioned the coil 0.34 roll/pitch degrees (SEM = 0.01°) and 0.22 yaw (SEM = 0.004°) off-target (both p < 0.0001). Further analyses revealed that there were significant inter-operator differences on distance and angle positioning for F3 (all p < 0.05), but not neuronavigation. Lastly, cap targeting resulted in significantly lower E-fields at the intended prefrontal cortical target, with equivalent E-fields as 110.7% motor threshold (MT; range = 58.3–127.4%) stimulation vs. 119.9% MT (range = 115–123.3%) from neuronavigated targeting with 120% MT stimulation applied (p < 0.001). ConclusionsCap-based targeting is an inherent source of target variability compared to neuronavigation. Additionally, cap-based coil placement is more prone to differences across operators. Off-target coil placement secondary to cap-based measurements results in significantly lower amounts of stimulation reaching the cortical target, with some individuals receiving only 48.6% of the intended on-target E-field. Neuronavigation technology enables more precise and accurate TMS positioning, resulting in the intended stimulation intensities at the targeted cortical level.

Knowledge-driven based three-dimensional prospectivity modeling of Fe–Cu skarn deposits; a case study of the Fanchang volcanic basin, anhui province, Eastern China

Mineral exploration at depth is becoming increasingly important as a result of the decades of discovery and mining of surficial or near-surface mineralization, leading to targeting of deeper-seated mineralization and mineral systems. The numerous challenges posed by this transition to deeper exploration can be partly overcome by the application of 3D prospectivity modeling. This study provides an outline of this type of prospectivity modeling and focuses on the under-explored Fanchang volcanic basin, an area prospective for skarn-type Fe–Cu mineralization within the Middle–Lower Yangtze River Metallogenic belt (MLYRB) of eastern China. The approach used utilizes a 3D geological model of the Fanchang volcanic basin that enables the identification and extraction of favorable elements for mineralization using 3D spatial and mathematical morphology analysis methods. A 3D distance field method was also used to obtain distance field information for the favorable ore-forming factors delineated during this study. The limited exploration to date in the study area means that there is a lack of specific information (e.g., location, size, grade) for the mineralization in this region. As such, knowledge-driven Index Overlay with Multi-Class (IOMC), Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS), and Fuzzy Comprehensive Evaluation (FCE) approaches were used to generate 3D prospectivity models for the Fanchang volcanic basin. The resulting weight coefficients relating to favorable ore-forming factors were quantified using an Analytic Hierarchy Process (AHP) approach, yielding random consistency ratio (CR) values of 0.0125. These are all less than a value of 0.1, which would indicate inconsistencies between the matrix values and values for the entirety of the hierarchy, indicating that our calculated weighting coefficients are precise and accurate. Comparison of the 3D prospectivity modeling results of the three methods using a prediction area (P–A) plot indicates that 88% of the known mineralization and mineralized occurrences within the basin were located within an area of only 12% of the study area considered prospective during modeling using the IOMC method. In comparison, TOPSIS and FCE modeling positively identified 82% (18% of the study area) and 91% (9% of the study area) of the mineralization in the study area, respectively. These results indicate that these approaches are viable tools for exploration targeting in relatively unexplored areas without specific orebody size, location, and grade data, and the FCE method appears more efficient and reliable than the other two methods in predicting locations to explore for Fe–Cu skarn mineralization. The approaches outlined during this study are widely applicable for exploration for other mineralizing systems in under-explored regions elsewhere and the specific high prospectivity targets identified during this study and the links between exploration indicators and high prospectivity regions also provide key areas and insights for both future modeling and mineral exploration within the Fanchang volcanic basin and the wider MLYRB.

Signed distance field framework for unified DEM modeling of granular media with arbitrary particle shapes

Signed distance field framework for unified DEM modeling of granular media with arbitrary particle shapes