Three-dimensional Data Research Articles

Parallel B-Scan YOLO Plus A-Scan Voting Method for Pavement Integrity Assessment Based on Three-Dimensional Ground-Penetrating Radar Data

Pavement integrity assessment based on ground-penetrating radar (GPR) remains challenging because of the uncertainty of the relationship between pavement damage and GPR reflection patterns. Further, existing image-processing techniques for analyzing GPR data still need significant improvements. In this study, a novel framework termed “parallel B-scan You Only Look Once (YOLO) plus A-scan voting” (P-BYPAV) was proposed to improve the performance of YOLO classification in the presence of ambiguous B-scan image patterns. The framework was developed by utilizing GPR-aided coring and falling-weight deflectometer test data. GPR B-scan reflection patterns can be classified into four categories: (1) the valley-peak-valley pattern (V-P-V) representing intact and healthy interfaces where moisture damage is minimal; (2) the peak-valley-peak pattern (P-V-P) representing the presence of air voids at layer interfaces, with a strong correlation to severe debonding; (3) the amplified valley-peak-valley pattern (Amp. V-P-V) representing the presence of moisture at layer interfaces, corresponding to various level of pavement damages; and (4) the symmetric peak-valley pattern, which indicates ambiguity in distinguishing moisture content and air voids. A dataset of 300 labeled images was created using the knowledge of these patterns. The dataset was then used to train the proposed P-BYPAV framework, demonstrating that P-BYPAV successfully integrated the advantages of YOLO and A-scan data recognition. The results show that P-BYPAV resolved more than 97% of the ambiguous regions outlined by YOLO alone.

Exploration of crystal chemical space using text-guided generative artificial intelligence

The vastness of chemical space presents a long-standing challenge for the exploration of new compounds with pre-determined properties. In materials science, crystal structure prediction has become a mature tool for mapping from composition to structure based on global optimisation techniques. Generative artificial intelligence now offers the means to efficiently navigate larger regions of crystal chemical space informed by structure-property datasets of materials. Here, we introduce a model, named Chemeleon, designed to generate chemical compositions and crystal structures by learning from both textual descriptions and three-dimensional structural data. The model employs denoising diffusion techniques for compound generation using textual inputs aligned with structural data via cross-modal contrastive learning. The potential of this approach is demonstrated for multi-component compound generation, including the Zn-Ti-O ternary space, and the prediction of stable phases in the Li-P-S-Cl quaternary space of relevance to solid-state batteries.

VIRTual autOPSY-applying CT and MRI for modern forensic death investigations.

Virtual autopsy, an advanced forensic technique, utilizes cutting-edge imaging technologies such as computed tomography (CT) and magnetic resonance imaging (MRI) to investigate the cause and manner of death without the need for physical dissection. By creating detailed, three-dimensional data of the entire body or specific areas of interest, these post-mortem imaging modalities provide a comprehensive, non-invasive approach to examining decedents. This article explores the historical development of virtual autopsy, its current applications in forensic medicine, and its promising future. It highlights the crucial roles of CT and MRI in forensic death investigations, while also addressing the challenges and limitations associated with these imaging techniques in post-mortem examinations.

Study on the Relationship Between 3D Landscape Patterns and Residents’ Comfort in Urban Multi-Unit High-Rise Residential Areas: A Case Study of High-Density Inland City

As urbanization accelerates, the increasing density of urban buildings and the prevalence of multi-unit high-rise residential areas have emerged as significant factors affecting residents’ comfort. Effective green space planning within residential areas can mitigate residents’ thermal discomfort. This study utilizes methods including the construction of two-dimensional and three-dimensional landscape indices and meteorological data simulation to examine the relationship between residents’ comfort levels at various heights in residential buildings and the 3D landscape patterns of residential areas, based on semantic three-dimensional grid data from a residential complex in Wuhan. The results indicate that (1) The characteristics of 3D landscape patterns vary across different regions within multi-unit high-rise residential areas. The landscape patches in the central and southern regions are more balanced compared to other areas, while there is minimal height variation in residential buildings in the northeastern region. (2) There are notable differences in comfort levels at varying heights across different areas of the residential district. In summer, residents expressing satisfaction with environmental comfort are primarily located in high-rise buildings in the central-southern region, whereas in winter, satisfaction is concentrated among residents in lower and mid-rise buildings in both the northern center and southern areas. (3) The degree of landscape fragmentation, the dominance of certain patches, and the distribution of buildings and vegetation at different heights significantly influence residents’ comfort. Achieving a balanced distribution of green spaces, reducing building density, and ensuring a uniform arrangement of trees of varied heights can effectively enhance the living environment for residents on lower floors, providing practical strategies for the planning of green spaces and built environments that improve overall resident quality of life. This research provides a theoretical foundation and reference for evaluating thermal comfort in high-rise residential areas and optimizing green space configurations.

AIVT: Inference of turbulent thermal convection from measured 3D velocity data by physics-informed Kolmogorov-Arnold networks.

We propose the artificial intelligence velocimetry-thermometry (AIVT) method to reconstruct a continuous and differentiable representation of the temperature and velocity in turbulent convection from measured three-dimensional (3D) velocity data. AIVT is based on physics-informed Kolmogorov-Arnold networks and trained by optimizing a loss function that minimizes residuals of the velocity data, boundary conditions, and governing equations. We apply AIVT to a set of simultaneously measured 3D temperature and velocity data of Rayleigh-Bénard convection, obtained by combining particle image thermometry and Lagrangian particle tracking. This enables us to directly compare machine learning results to true volumetric, simultaneous temperature and velocity measurements. We demonstrate that AIVT can reconstruct and infer continuous, instantaneous velocity and temperature fields and their gradients from sparse experimental data at a high resolution, providing an additional approach for understanding thermal turbulence.

Flight route EDR estimation using MHE to fuse three-dimensional wind information in the QAR data analysis.

To address the problem that vertical wind data can provide information only on turbulence fluctuations in the vertical direction, meaning that the consideration of insufficiently compre-hensive turbulence information leads to low accuracy in eddy dissipation rate(EDR) estimation based on vertical wind (VWE), the paper proposes a flight route EDR estimation using MHE to fuse three-dimensional wind information in the Quick Access Recorder(QAR) data analysis. Quality control is performed on QAR data to eliminate irrelevant features for computing vertical wind and the two horizontal wind components. The corresponding formulas are then applied to obtain three-dimensional wind data information. Subsequently, a multi-head attention mechanism is employed to quantify the relationship between three-dimensional wind characteristics and the feature matrix (vertical acceleration). This process generates feature fusion weights, which are mapped onto the three-dimensional wind matrix and used to convert the three-dimensional wind data into one-dimensional wind data, thereby achieving the fusion of three-dimensional wind features. Finally, through maximum likelihood estimation(MLE) combine the new wind data in the frequency domain, the estimation of the flight route EDR is realized. Experimental results confirmed that the proposed estimation algorithm exhibits optimal performance compared to EDR estimations based on PCE and VWE, and the estimated values have smaller errors relative to the true EDR values. The proposed algorithm provides highly accurate EDR estimations for flight paths and demonstrates good practical applicability for assessing turbulence intensity along flight routes, thereby enhancing the safety of aircraft during flight.

A Novel 3D Approach with a CNN and Swin Transformer for Decoding EEG-Based Motor Imagery Classification.

Motor imagery (MI) is a crucial research field within the brain-computer interface (BCI) domain. It enables patients with muscle or neural damage to control external devices and achieve movement functions by simply imagining bodily motions. Despite the significant clinical and application value of MI-BCI technology, accurately decoding high-dimensional and low signal-to-noise ratio (SNR) electroencephalography (EEG) signals remains challenging. Moreover, traditional deep learning approaches exhibit limitations in processing EEG signals, particularly in capturing the intrinsic correlations between electrode channels and long-distance temporal dependencies. To address these challenges, this research introduces a novel end-to-end decoding network that integrates convolutional neural networks (CNNs) and a Swin Transformer, aiming at enhancing the classification accuracy of the MI paradigm in EEG signals. This approach transforms EEG signals into a three-dimensional data structure, utilizing one-dimensional convolutions along the temporal dimension and two-dimensional convolutions across the EEG electrode distribution for initial spatio-temporal feature extraction, followed by deep feature exploration using a 3D Swin Transformer module. Experimental results show that on the BCI Competition IV-2a dataset, the proposed method achieves 83.99% classification accuracy, which is significantly better than the existing deep learning methods. This finding underscores the efficacy of combining a CNN and Swin Transformer in a 3D data space for processing high-dimensional, low-SNR EEG signals, offering a new perspective for the future development of MI-BCI. Future research could further explore the applicability of this method across various BCI tasks and its potential clinical implementations.

Quantitative susceptibility mapping study of deep gray matter iron content in glioma patients.

Iron is associated with the pathophysiology of gliomas. The measurement of brain iron levels plays a vital role in defining the pathophysiologic changes caused by gliomas. This study aimed to analyze the effect of gliomas with different histopathology and molecular characteristics on deep gray matter (DGM) iron content and the relationship between DGM iron content and cognition in glioma patients (PT). In this retrospective study, we included 81 PT, categorized according to different histopathology and molecular characteristics, and 30 age- and gender-matched healthy controls (HC). Brain quantitative susceptibility mapping (QSM) maps were computed from three-dimensional (3D) multi-echo gradient-echo data using Laplacian-based phase unwrapping, a variational sharpening (V-SHARP) background field correction and the streaking artifacts reduction (STAR)-QSM method. ITK-SNAP was used to measure the susceptibility values reflecting the iron content in the regions of interest. Differences in DGM magnetic susceptibility between groups were compared. Pearson's correlation analysis assessed the relationship between DGM magnetic susceptibility and Montreal Cognitive Assessment (MoCA). Compared to HC, PT showed higher DGM magnetic susceptibility. QSM analysis exhibited higher DGM magnetic susceptibility in high-grade gliomas (HGG) than low-grade gliomas (LGG). Isocitrate dehydrogenase (IDH) wild-type gliomas showed higher magnetic susceptibility than IDH mutant-type in the putamen [IDH mutant-type: interquartile range (IQR), 0.052, 0.019 ppm vs. IDH wild-type: IQR, 0.062, 0.015 ppm, P=0.015]. A higher Ki-67 index also correlated with higher magnetic susceptibility in the putamen (high Ki-67 index group: IQR, 0.063, 0.019 ppm vs. low Ki-67 index group: IQR, 0.052, 0.016 ppm, P=0.005). The PT group had significantly lower MoCA scores than the HC group (patients group vs. HC group: 22.30±4.92 vs. 24.70±2.20, P=0.021). There was a negative correlation between MoCA scores and DGM magnetic susceptibility. The DGM iron content, as measured by QSM, was higher in PT than in HC and correlated with tumor grade. Cognitive decline is associated with increased DGM iron content in PT. The DGM iron content could be a potential biomarker in glioma differentiation and prognosis.

An alternative approach to code, store, and regenerate 3D data in dental medicine using open-source software: A scripting-based technique.

An alternative approach to code, store, and regenerate 3D data in dental medicine using open-source software: A scripting-based technique.

Advancements in Wind Turbine Wake Modelling using 3D scanning LiDAR measurements

Abstract This study investigates the capability of numerical models to estimate wind turbine wake characteristics under varying atmospheric stability conditions using three-dimensional wind data obtained from a scanning LiDAR. The LiDAR was set to measure the wake of a 3.8 MW test turbine up to 5 rotor diameters downstream from the hub, for a period of four months, capturing both vertical and horizontal wake profiles. These measurements were compared against results from the parabolized Navier-Stokes (PNS) solver FarmFlow (FF) and Unsteady Reynolds-Averaged Navier-Stokes (URANS) simulations. The results showed a good correlation between the measured wake profiles and the simulation results for both models. The URANS model effectively captured horizontal wake profiles, including detailed features such as double-Gaussian profiles and turbine misalignment effects, performing best under stable atmospheric conditions. However, URANS had limitations in accurately matching vertical profiles, especially under neutral atmospheric conditions. The FarmFlow model provided reliable estimates without tuning, demonstrating good agreement with measurements for all conditions, though it underestimated the wake strength in neutral atmospheric conditions. This study provides insights into atmospheric and wind turbine wake measurements as well as into wake models capabilities, bringing forward directions for wind turbine wake simulations improvements.

Impacts of Vertical Resolution and Diurnally Varying Background Error Covariance on FY-4A/AGRI Data Assimilation over Land

Abstract This study aims at improving the assimilation of Fengyun-4A (FY-4A)/Advanced Geostationary Radiation Imager (AGRI) clear-sky surface-sensitive brightness temperature over land using the Weather Research and Forecasting and the Gridpoint Statistical Interpolation three-dimensional variational data assimilation for summer convective cases over eastern China. We investigated the influences of the planetary boundary layer on the assimilation of FY-4A/AGRI surface-sensitive channels, focusing on model low-level vertical resolution and diurnal variation of background error covariance. The vertical levels are increased from 48 to 91 by adding more near-surface levels. First, stronger low-level vertical background error correlations of humidity and temperature are found during daytime than during nighttime. A total of five numerical experiments are then conducted to examine the impacts of FY-4A/AGRI assimilation with two vertical resolutions. The experiments WV48 and WV91 assimilate only the water vapor channels, and ALL48, ALL91, and ALL91M assimilate the water vapor and surface-sensitive channels. The diurnally varying background error covariance, which is estimated at 3-h intervals, is incorporated into ALL91M. The 6-h cycling data assimilation at 1-h interval prior to convection initiation is conducted for each of the seven cases. Overall, a comparison between WV48/WV91 and ALL48/ALL91M revealed a consistent positive impact of AGRI brightness temperature assimilation over land on convective precipitation forecasts. The added values of assimilating AGRI surface-sensitive channels are more pronounced by using the finer vertical resolution due to low-level humidity enhancements that are related to the rise of the planetary boundary layer. With the diurnally varying background error covariance, ALL91M outperformed ALL91 because of improved low-level analyzed structures that are associated with proper low-level vertical error correlations of humidity and thermal variables.

Improving PM2.5 forecasts with three-dimensional variation data assimilation of visibility observations in China

Improving PM2.5 forecasts with three-dimensional variation data assimilation of visibility observations in China

Research on Brinell hardness measurement method based on three-dimensional point cloud data

Research on Brinell hardness measurement method based on three-dimensional point cloud data

Three-dimensional heterogeneous panel data models with multi-level interactive fixed effects

Three-dimensional heterogeneous panel data models with multi-level interactive fixed effects

Nonlinear finite element modeling and seismic analysis of Thai historical masonry wall structures

Abstract Ayutthaya is a significant heritage tourism destination in Thailand, with numerous historical sites featuring ancient brick walls. Assessing their stability and structural integrity requires accurate material data and appropriate modeling techniques. This paper describes nonlinear finite element modeling of Thai historical masonry wall structures. The presented models are used to perform nonlinear seismic analysis. A section of the historical wall in Wat Phra Si Sanphet, Ayutthaya, is selected as a research case. Three-dimensional laser scanning data show that the wall is arranged in an S-shape along the east–west direction. The wall materials are red bricks and lime mortar produced locally in the 15th century CE. Affected by natural weathering and geological movement, the surface mortar of the historical wall has visible peeling and damage, and the brick units are partially broken or displaced. Three finite element modeling methods using the concrete damage plasticity theory, namely, a detailed micro-model, a simplified micro-model, and a macro-model, are proposed and compared with previously published laboratory tests of masonry prisms. The results reveal that the simplified micro-model is the most suitable. Consequently, a seismic analysis of the historical wall in Wat Phra Si Sanphet, Ayutthaya, is performed using the simplified micro-model. Examples of possible wall damage profiles are provided, and guidance for structural reinforcement is provided.

Prediction of vegetation pattern evolution in arid ecosystems using 3D-Var data assimilation.

In arid regions, ecosystems are fragile, and vegetation exhibits high sensitivity to changes in climatic conditions. Vegetation patterns-non-uniform macroscopic structures formed by vegetation through temporal and spatial self-organization-serve as critical indicators of an ecosystem's adaptive capacity, post-disturbance resilience, and early warning signals of ecosystem degradation. Investigating the formation mechanisms of vegetation patterns using reaction-diffusion (RD) models represents a vital approach to deciphering vegetation evolution dynamics, with significant implications for protecting arid ecosystems. However, heterogeneous steady-state solutions of RD systems, such as Turing patterns, often reside in multistable regions. This implies that minute variations in initial conditions may lead to markedly divergent outcomes. When initial vegetation distribution data are imprecise, predictions of vegetation evolution trends and steady-state distributions in a given spatial position using RD models become highly sensitive to initial errors-a case where "minor discrepancies in input yield vastly divergent results." This study applies the three-dimensional variational data assimilation method to a RD model coupling vegetation, soil moisture, and surface water dynamics in arid regions. The results demonstrate that incorporating a modest amount of observational data can substantially enhance the model's predictive accuracy for vegetation evolution trajectories.

Study on Wind-eroded Concave Surface Morphology Based on Laser Radar and Color Space Encoding

Abstract This article takes a test area in the eastern edge of Greenbulongtan, Barunbireli Town, Alxa Left Banner, Alxa League, Inner Mongolia Autonomous Region, where a polylactic acid fiber (PLA) sand barrier is installed, as the research object. By using the three-dimensional scanning technology and software such as Sketch Up with the portable device iPhone12 Pro Max equipped with a LiDAR, the boundary length, erosion depth, erosion coefficient, and erosion intensity of the concave surface of the sand barrier in the test area were measured through the method of model size measurement. In addition, the model size measurement method and the novel RGB numerical simulation algorithm proposed in this paper were mutually corroborated in terms of the accumulation rate of the concave surface. The experiment shows that the erosion coefficient of the PLA wind erosion concave surface with a size of 1.5 m×1.5 m in the test area is about 1.34, the accumulation intensity is about 0.8 ml/cm2, and the erosion intensity is about 1.2 ml/cm2. The overall erosion amount in the test area is greater than the accumulation amount, and the stability of the 1.5 m×1.5 m PLA sand barrier in the experimental stage is poor. Exploring the relationship between the specific RGB values and eleva-tion is the highlight of this article. The novel RGB numerical simulation algorithm proposed in this paper calculates the erosion rate, which fits with the erosion rate obtained by measuring the three-dimensional model size data, with a fitting degree of more than 98%. This proves that using LiDAR three-dimensional modeling to study wind erosion concave surfaces is feasible, and the RGB numerical simulation algorithm is simple and effective in quickly and efficiently processing complex concave terrain.

Zygomatic Osteotomy surgery design software based on skull CT scans - Self-supervised algo reduces workload.

Zygomatic Osteotomy surgery design software based on skull CT scans - Self-supervised algo reduces workload.

A personalized time-resolved 3D mesh generative model for unveiling normal heart dynamics

Understanding the structure and motion of the heart is crucial for diagnosing and managing cardiovascular diseases, the leading cause of global death. There is wide variation in cardiac shape and motion patterns, influenced by demographic, anthropometric and disease factors. Unravelling normal patterns of shape and motion, and understanding how each individual deviates from the norm, would facilitate accurate diagnosis and personalized treatment strategies. Here, to this end, we developed a conditional generative model, MeshHeart, to learn the distribution of shape and motion patterns for the left and right ventricles of the heart. To model the high-dimensional spatio-temporal mesh data, MeshHeart uses a geometric encoder to represent cardiac meshes in a latent space and a temporal transformer to model the motion dynamics of latent representations. Based on MeshHeart, we investigate the latent space of 3D + t cardiac mesh sequences and propose a distance metric, latent delta, which quantifies the deviation of a real heart from its personalized normative pattern. Here, 3D + t refers to three-dimensional data evolving over time. In experiments using a large cardiac magnetic resonance image dataset of 38,309 participants from the UK Biobank, MeshHeart demonstrates high performance in cardiac mesh sequence reconstruction and generation. Latent space features are discriminative for cardiac disease classification, whereas latent delta exhibits strong correlations with clinical phenotypes in phenome-wide association studies.

Inter-microscope comparability of dental microwear texture data obtained from different optical profilometers: Part II Deriving instrument-specific correction equations for meta-analyses using published data.

Dental microwear texture analysis (DMTA) has emerged as a valuable method for investigating the feeding ecology of vertebrates. Over the past decade, three-dimensional topographic data from microscopic regions of tooth surfaces have been collected, and surface texture parameters have been published for both extant and fossil species. However, different types of measurement instruments and surface processing used by respective laboratories conducting DMTA have limited the potential data comparison. In this study, we propose correction formulae for the DMTA data produced by different instruments to facilitate intercomparison. We used six confocal instruments from five laboratories to scan standard tooth samples with strictly defined scan areas. We found significant differences in DMTA parameter values among the different machines, despite scanning the exact same spots. The degree of discrepancy varied considerably, with instruments from the same manufacturer and similar models showing less variation. Some parameters exhibited high correlations between instruments, enabling the development of regression equations for correction formulae. Using these correction formulae, we adjusted published DMTA data and conducted a meta-analysis of extant herbivores to examine the effects of internal and external abrasives. Our findings indicate overall positive effects of internal and external abrasives on DMTA, with varying responses to the abrasives between ruminant and non-ruminant herbivores. The meta-analysis supports the hypothesis that ruminants effectively "wash ingesta" in their rumens, mitigating the impact of external abrasives and reducing overall dental wear.