Search Results (1,614)

Background/Objectives: The clinical management of adolescent idiopathic scoliosis (AIS) is hindered by the inability to accurately predict curve progression. Although skeletal maturity and the initial Cobb angle are established predictors of progression, their combined predictive accuracy remains limited. This study aimed to develop a robust and interpretable artificial intelligence (AI) system using deep learning (DL) models to predict the progression of scoliosis using only standing frontal radiographs. Methods: We conducted a multicenter study involving 542 patients with AIS. After excluding 52 borderline progression cases (6–9° progression in the Cobb angle), 294 and 196 patients were assigned to progression (≥10° increase) and non-progression (≤5° increase) groups, respectively, considering a 2-year follow-up. Frontal whole spinal radiographs were preprocessed using histogram equalization and divided into two regions of interest (ROIs) (ROI 1, skull base–femoral head; ROI 2, C7–iliac crest). Six pretrained DL models, including convolutional neural networks (CNNs) and transformer-based models, were trained on the radiograph images. Gradient-weighted class activation mapping (Grad-CAM) was further performed for model interpretation. Results: Ensemble models outperformed individual ones, with the average ensemble model achieving area under the curve (AUC) values of 0.769 for ROI 1 and 0.755 for ROI 2. Grad-CAM revealed that the CNNs tended to focus on the local curve apex, whereas the transformer-based models demonstrated global attention across the spine, ribs, and pelvis. Models trained on ROI 2 performed comparably with respect to those using ROI 1, supporting the feasibility of image standardization without a loss of accuracy. Conclusions: This study establishes the clinical potential of transformer-based DL models for predicting the progression of scoliosis using only plain radiographs. Our multicenter approach, high AUC values, and interpretable architectures support the integration of AI into clinical decision-making for the early treatment of AIS. Full article

Ultra-short-term wind power exhibits significant minute-level fluctuation characteristics, leading to substantial instantaneous prediction errors. To mitigate the impact of instantaneous wind power prediction errors, the following steps are taken: First, the correlation between instantaneous prediction errors and meteorological factors is determined, and strongly associated variables are selected as model inputs. Next, the particle swarm optimization algorithm is employed to optimize the initial weights and threshold parameters of the spatiotemporal backpropagation neural network prediction model to enhance its performance. Subsequently, based on the nonlinear relationship between wind speed/direction data and instantaneous prediction errors, a wind speed matrix gradient correction method and a deep learning correction method with physical constraints on prediction errors are constructed to address errors caused by declining model generalization under strong disturbances. To validate the effectiveness of the proposed prediction algorithm integrating parameter optimization and the error correction method, it is compared with typical convolutional neural networks, long short-term memory networks, and backpropagation neural algorithms. The results demonstrate that compared to other wind power prediction strategies, this method reduces the mean absolute percentage error, root mean square error, and mean absolute error by 48.49%, 45.51%, and 50.8%, respectively. These results confirm that combining error correction strategies with prediction model parameter optimization effectively enhances the ability to reduce instantaneous wind power prediction errors, providing a practical technical solution for optimizing ultra-short-term wind power prediction accuracy and offering valuable insights for ensuring the stability of wind power grid integration. Full article

This study develops a novel analytical solution for three-dimensional axisymmetric consolidation of unsaturated soils incorporating radial–vertical composite seepage mechanisms and anisotropic permeability characteristics. A groundbreaking dual orthogonal expansion framework is established, utilizing innovative Bessel–trigonometric function coupling to solve the inherently complex spatiotemporal coupled partial differential equations in cylindrical coordinate systems. The mathematical approach synergistically combines modal expansion theory with Laplace transform methodology, achieving simultaneous spatial expansion of gas–liquid two-phase pressure fields through orthogonal function series, thereby transforming the three-dimensional problem into solvable ordinary differential equations. Rigorous validation demonstrates exceptional accuracy with coefficient of determination R² exceeding 0.999 and relative errors below 2% compared to numerical simulations, confirming theoretical correctness and practical applicability. The analytical solutions reveal four critical findings with quantitative engineering implications: (1) dual-directional drainage achieves 28% higher pressure dissipation efficiency than unidirectional drainage, providing design optimization criteria for vertical drainage systems; (2) normalized matric suction variation exhibits characteristic three-stage evolution featuring rapid decline, plateau stabilization, and slow recovery phases, while water phase follows bidirectional inverted S-curve patterns, enabling accurate consolidation behavior prediction under varying saturation conditions; (3) gas-water permeability ratio k_a/k_w spanning 0.1 to 1000 produces two orders of magnitude time compression effect from 10⁻² s to 10⁻⁴ s, offering parametric design methods for construction sequence control; (4) initial pressure gradient parameters λ_a and λ_w demonstrate opposite regulatory mechanisms, where increasing λ_a retards consolidation while λ_w promotes the process, providing differentiated treatment strategies for various geological conditions. The unified framework accommodates both uniform and gradient initial pore pressure distributions, delivering theoretical support for refined embankment engineering design and construction control. Full article

Clarifying the differential effects of temperature gradient and temperature change rate on the evolution of rock fractures and damage mechanism under thermal shock is of great significance for the development and utilization of deep geothermal resources. In this study, granite samples at different temperatures (20 °C, 150 °C, 300 °C, 450 °C, 600 °C, and 750 °C) were subjected to rapid cooling treatment with liquid nitrogen. After the thermal treatment, a series of tests were conducted on the granite, including wave velocity test, uniaxial compression experiment, computed tomography scanning, and scanning electron microscopy test, to explore the influence of thermal shock on the physical and mechanical parameters as well as the meso-structural damage of granite. The results show that with the increase in heat treatment temperature, the P-wave velocity, compressive strength, and elastic modulus of granite gradually decrease, while the peak strain gradually increases. Additionally, the failure mode of granite gradually transitions from brittle failure to ductile failure. Through CT scanning experiments, the spatial distribution characteristics of the pore–fracture structure of granite under the influence of different temperature gradients and temperature change rates were obtained, which can directly reflect the damage degree of the rock structure. When the heat treatment temperature is 450 °C or lower, the number of thermally induced cracks in the scanned sections of granite is relatively small, and the connectivity of the cracks is poor. When the temperature exceeds 450 °C, the micro-cracks inside the granite develop and expand rapidly, and there is a gradual tendency to form a fracture network, resulting in a more significant effect of fracture initiation and permeability enhancement of the rock. The research results are of great significance for the development and utilization of hot dry rock and the evaluation of thermal reservoir connectivity and can provide useful references for rock engineering involving high-temperature thermal fracturing. Full article

This study addresses the challenge of cognitive waveform design for multiple-input–multiple-output (MIMO) radar systems operating in cluttered environments. It focuses on the key practical requirements for transmitting time-domain waveforms and proposes a novel approach. This method first determines the optimal frequency-domain waveform and then designs a time-domain waveform that closely approximates the frequency-domain solution. The primary objective is to enable MIMO radar systems to transmit orthogonal waveforms while accommodating various constraints. A frequency-domain waveform optimization model was initially developed using the principle of maximizing dual mutual information (DMI), and the energy spectral density (ESD) of the optimal waveform was derived using the water-filling method. Next, a time-domain waveform approximation model is constructed based on the minimum mean square error (MMSE) criterion, which incorporates constant modulus and peak-to-average power ratio (PAPR) constraints. To minimize the performance degradation of the waveform, an improved adaptive gradient descent genetic algorithm (GD-AGA) was proposed to synthesize multichannel orthogonal time-domain waveforms for MIMO radars. The simulation results demonstrate the effectiveness of the proposed model for enhancing the performance of MIMO radar. Compared with traditional genetic algorithms (GA) and two enhanced GA alternatives, the proposed algorithm achieves a lower ESD loss and better orthogonal performance. Full article

Soil aggregate stability plays a central role in mediating slope erosion, a key ecological process in North China. This study aimed to investigate how aggregate structures (reflected by rainfall intensity and vegetation-type differences) influence the erosion process. Using wasteland as the control, we conducted artificial simulated rainfall experiments on soils covered by Quercus variabilis, Platycladus orientalis, and shrubs, with three rainfall intensity gradients. Key findings showed that Platycladus orientalis exhibited the strongest infiltration capacity and longest runoff initiation delay due to its high proportion of stable macroaggregates (>0.25 mm), while barren land readily formed surface crusts, leading to the fastest runoff. Increased rainfall intensity significantly exacerbated runoff and erosion. When the macroaggregate content exceeded 60%, sediment yield rates dropped sharply, with a significant negative exponential relationship between the mean weight diameter (MWD) and sediment yield; barren land (dominated by microaggregates) faced the highest erosion risk and fell into an erosion–fragmentation vicious cycle. Redundancy analysis revealed that microbial communities (e.g., Ascomycota) and fine roots were dominant erosion-controlling factors under heavy rainfall. Ultimately, the synergistic system of the macroaggregate architecture and root-microbial cementation enabled Platycladus orientalis and other tree stands to reduce soil erodibility via maintaining aggregate stability, whereas shrubs and barren land amplified rainfall intensity effects. barren landbarren landmm·h⁻¹ mm·h⁻¹ mm·h⁻¹ barren land. Full article

Accurate prediction of the flow field is crucial to evaluating the aerodynamic performance of an aircraft. While traditional computational fluid dynamics (CFD) methods solve the governing equations to capture both global flow structures and localized gradients, they are computationally intensive. Deep learning-based surrogate models offer a promising alternative, yet often struggle to simultaneously model long-range dependencies and near-wall flow gradients with sufficient fidelity. To address this challenge, this paper introduces the Message-passing And Global-attention block (MAG-BLOCK), a graph neural network module that combines local message passing with global self-attention mechanisms to jointly learn fine-scale features and large-scale flow patterns. Building on MAG-BLOCK, we propose FLOW-GLIDE, a cross-architecture deep learning framework that learns a mapping from initial conditions to steady-state flow fields in a latent space. Evaluated on the AirfRANS dataset, FLOW-GLIDE outperforms existing models on key performance metrics. Specifically, it reduces the error in the volumetric flow field by 62% and surface pressure prediction by 82% compared to the state-of-the-art. Full article

Neutrophils are an essential protective component of the innate immune system. However, in severe bacterial infections, neutrophils are known to mis-localise from the primary site of infection to other organs, where excessive release of cytokines, chemokines, and neutrophil extracellular traps (NETs) can induce organ damage and death. In this study, we use an animal model of bacterial infection originating in the peritoneum to show that hydrogen peroxide (H₂O₂, a potent neutrophil chemoattractant) is initially released in high concentrations both in the peritoneum and in multiple ‘off-target’ organs (lungs, liver and kidneys). The initial high H₂O₂ release inhibits neutrophil chemotaxis, but after 24 h concentrations of H₂O₂ reduce and can promote neutrophil migration to organs, where they release pro-inflammatory cytokines and chemokines along with NETs. The antimalarial compound artesunate potently inhibits neutrophil migration to off-target organs. It also abolishes cytokine, chemokine, and NET production, suggesting that artesunate may be a valuable novel therapy for preventing off-target organ inflammation associated with severe bacterial infections. Finally, the potency of H₂O₂ as a chemoattractant is shown by in vitro experiments in which, faced with competing gradients of H₂O₂ and other chemoattractants, neutrophils preferentially migrate towards H₂O₂. Full article

The Yangtze River Delta (YRD) region has experienced rapid development in its tourism industry, establishing itself as a leading force within China’s tourism sector. However, significant regional disparities continue to hinder its sustainable development. This study adopts a mixed-methods approach to analyze the spatiotemporal evolution of domestic tourist flows and tourism industry agglomeration patterns in the region. Using city-level data from 2016 to 2022, the analysis employs a comprehensive methodology including standard deviation, coefficient of variation, standard deviation ellipse, and locational entropy. The main findings are as follows: (1) In the pre-pandemic period (2016–2019), absolute disparities in tourist flows widened, whereas relative disparities narrowed. During the pandemic (2020–2022), absolute disparities decreased, while relative disparities initially increased before contracting. (2) Tourist flows displayed a southeast–northwest gradient, with high-value areas clustered along the southeastern coast. Standard deviation ellipse analysis reveals that tourist flows were primarily distributed along the eastern coastal corridor, parallel to the coastline. Prior to the pandemic, tourism growth showed a tendency toward spatial equilibrium; however, this trend was disrupted during the pandemic, resulting in a more decentralized spatial pattern. (3) Throughout the pandemic, tourism industry concentration increased significantly in most cities. Cities with renowned scenic attractions and diversified economic structures demonstrated stronger resilience, while those heavily reliant on tourism were more vulnerable to the pandemic’s effects. Full article

Plants vary in number of chromosomes (ploidy levels), which can influence morphological traits, including the size and density of stomata cells. Although biologists can detect these differences under a microscope, the process is often time-consuming and tedious. This study aims to automate the classification of blackcurrant (Ribes nigrum L.) ploidy levels—diploid, triploid, and tetraploid—by leveraging deep learning techniques. Convolutional Neural Networks and Vision Transformers are employed to perform microscopic image classification across two distinct blackcurrant datasets. Initial experiments demonstrate that these models can effectively classify ploidy levels when trained and tested on subsets derived from the same dataset. However, the primary challenge lies in proposing a model capable of yielding satisfactory classification results across different datasets ensuring robustness and generalization, which is a critical step toward developing a universal ploidy classification system. In this research, a variety of experiments is performed including application of augmentation technique. Model efficacy is evaluated with standard metrics and its interpretability is ensured through Gradient-weighted Class Activation Mapping visualizations. Finally, future research directions are outlined with application of other advanced state-of-the-art machine learning methods to further refine ploidy level prediction in botanical studies. Full article

Lung cancer remains one of the most lethal cancers globally. Its early detection is vital to improving survival rates. In this work, we propose a hybrid computer-aided diagnosis (CAD) pipeline for lung cancer classification using Computed Tomography (CT) scan images. The proposed CAD pipeline integrates ten image preprocessing techniques and ten pretrained deep learning models for feature extraction including convolutional neural networks and transformer-based architectures, and four classical machine learning classifiers. Unlike traditional end-to-end deep learning systems, our approach decouples feature extraction from classification, enhancing interpretability and reducing the risk of overfitting. A total of 400 model configurations were evaluated to identify the optimal combination. The proposed approach was evaluated on the publicly available Lung Image Database Consortium and Image Database Resource Initiative dataset, which comprises 1018 thoracic CT scans annotated by four thoracic radiologists. For the classification task, the dataset included a total of 6568 images labeled as malignant and 4849 images labeled as benign. Experimental results show that the best performing pipeline, combining Contrast Limited Adaptive Histogram Equalization, Swin Transformer feature extraction, and eXtreme Gradient Boosting, achieved an accuracy of 95.8%. Full article

The rapid advancement of kinetic pursuit technologies has significantly increased the difficulty of evasion for high-speed UAVs (HSUAVs), particularly in scenarios where two collaboratively operating pursuers approach from the same direction with optimized initial space intervals. This paper begins by deriving an optimal initial space interval to enhance cooperative pursuit effectiveness and introduces an evasion difficulty classification framework, thereby providing a structured approach for evaluating and optimizing evasion strategies. Based on this, an intelligent maneuver evasion strategy using semantic classification progressive learning with twin delayed deep deterministic policy gradient (SCPL-TD3) is proposed to address the challenging scenarios identified through the analysis. Training efficiency is enhanced by the proposed SCPL-TD3 algorithm through the employment of progressive learning to dynamically adjust training complexity and the integration of semantic classification to guide the learning process via meaningful state-action pattern recognition. Built upon the twin delayed deep deterministic policy gradient framework, the algorithm further enhances both stability and efficiency in complex environments. A specially designed reward function is incorporated to balance evasion performance with mission constraints, ensuring the fulfillment of HSUAV’s operational objectives. Simulation results demonstrate that the proposed approach significantly improves training stability and evasion effectiveness, achieving a 97.04% success rate and a 7.10–14.85% improvement in decision-making speed. Full article

Toxins are naturally poisonous small compounds, peptides and proteins that are produced in all three kingdoms of life. Venoms are animal toxins and can contain even hundreds of different compounds. Numerous approaches have been used to detect toxins, including prediction methods. We developed a novel machine learning-based predictor for detecting protein toxins from their sequences. The gradient boosting method was trained on carefully selected training data. Initially, we tested 2614 features, which were reduced to 88 after a comprehensive feature selection procedure. Out of the four tested algorithms, XGBoost was chosen to train the final predictor. Comparison to available predictors indicated that ProToxin showed significant improvement compared to state-of-the-art predictors. On a blind test dataset, the accuracy was 0.906, the Matthews correlation coefficient was 0.796, and the overall performance measure was 0.796. ProToxin is a fast and efficient method and is freely available. It can be used for small and large numbers of sequences. Full article

The Qiongdongnan Basin constitutes a sedimentary basin characterized by elevated temperatures, significant overpressures, and abundant hydrocarbons. Investigations within this basin have identified hydrothermal fluid movements linked to overpressure conditions, comprising two vertically separated overpressured intervals. The shallow overpressure compartment is principally caused by a combination of undercompaction and clay diagenesis. In contrast, the deeper high-pressure compartment results from hydrocarbon gas generation. Numerical pressure modeling indicates late-stage (post-5 Ma) development of significant overpressure within the deep compartment. It is proposed that accelerated subsidence in the Pliocene-Quaternary initiated substantial gas generation, thereby promoting the formation of the deep overpressured system. Multiple organic maturation parameters, combined with fluid inclusion microthermometry, reveal a thermal anomaly adjacent to the upper boundary of the deep overpressured zone. This anomaly indicates vertical transport of hydrothermal fluids ascending from the underlying high-pressure zone. Laser Raman spectroscopy confirms the presence of both hydrocarbons and carbon dioxide within these migrating fluids. Integration of fluid inclusion thermometry with burial history modeling constrains the timing of hydrocarbon-carrying fluid charge to the interval from 4.2 Ma onward, synchronous with modeled peak gas generation and a phase of pronounced overpressure buildup. We propose that upon exceeding the fracture gradient threshold, fluid pressure triggered upward migration of deeply sourced, hydrocarbon-enriched fluids through hydrofracturing pathways. This process led to localized dissolution and fracturing near the top of the deep overpressured system, while simultaneously facilitating significant hydrocarbon accumulation and forming preferential accumulation zones. These findings provide critical insights into petroleum exploration in overpressured sedimentary basins. Full article

Cracking in asphalt layers of rigid–flexible composite pavements under coupled ambient temperature fields and traffic loading represents a critical failure mode. Traditional models based on uniform temperature assumptions inadequately capture the crack propagation mechanisms. This study developed a thermo-mechanical coupling model that incorporates realistic temperature-modulus gradients to analyze the compressive-shear behavior and simulate crack propagation using the extended finite element method (XFEM) coupled with a modified Paris’ law. Key findings reveal that the asphalt layer exhibits a predominant compressive-shear stress state; increasing the base modulus from 10,000 MPa to 30,000 MPa reduces the maximum shear stress by 22.8% at the tire centerline and 8.6% at the edge; thermal stress predominantly drives crack initiation, whereas vehicle loading governs the propagation path; field validation via cored samples confirms inclined top-down cracking under thermo-mechanical coupling; and the fracture energy release rate (G_f) reaches a minimum of 155 J·m⁻² at 14:00, corresponding to a maximum fatigue life of 32,625 cycles, and peaks at 350 J·m⁻² at 01:00, resulting in a reduced life of 29,933 cycles—reflecting a 9.0% temperature-induced fatigue life variation. The proposed model, which integrates non-uniform temperature gradients, offers enhanced accuracy in capturing complex boundary conditions and stress states, providing a more reliable tool for durability design and assessment of composite pavements. Full article