Search Results (10,394)

Automated landslide change detection using remote sensing imagery is critical for rapid disaster response. However, landslide change detection using bi-temporal optical imagery is frequently degraded by cross-region domain shifts and by the elongated, anisotropic morphology of landslide boundaries, leading to substantial pseudo-change alarms. To suppress pseudo-changes and improve cross-region robustness, we propose a DEM-assisted topography-conditioned and orientation-adaptive Siamese network (DEMO-Net) that injects topographic inductive bias through terrain-conditioned feature modulation and orientation-adaptive convolutions. Specifically, DEM-derived multi-channel priors are encoded to predict spatially varying FiLM parameters that recalibrate shallow optical features, suppressing spurious changes while preserving discriminative cues. In addition, we introduce an adaptive-oriented attention convolution that leverages a DEM-derived aspect to guide sparse multi-orientation aggregation via shared-kernel transformation, enabling direction-aware receptive-field alignment for elongated and direction-varying landslide structures without costly global attention. Experiments on the GVLM benchmark under a 5-fold site-wise cross-region protocol show that DEMO-Net achieves 85.17% F1 and 74.26% mIoU, outperforming the strongest CNN baseline FC-EF by 5.05% and 7.20%, respectively. These results demonstrate the effectiveness of jointly leveraging terrain-conditioned calibration and physically consistent orientation-aligned feature extraction for robust cross-region landslide change detection. Full article

Pediatric gait abnormalities are closely intertwined with musculoskeletal dysfunctions and heightened injury risk, underscoring the urgency of early and accessible screening tools. Here, we develop and validate a video-based semi-supervised Abnormal Gait Recognition Module (AGRM) to address unmet needs in pediatric gait assessment, with a focus on diagnostic performance and clinical interpretability. The AGRM is built on a 3D ResNet backbone, synergistically integrated with a Mean Teacher Module (MTM) to mitigate the limitations of limited labeled clinical data, and a Spatial Hierarchical Pooling Module (SHPM) for robust multiscale spatiotemporal feature extraction—two core innovations tailored to gait dynamics. We trained and validated the model on a hybrid dataset combining self-collected pediatric gait videos and the public CASIA-B dataset, evaluating its performance in binary (normal vs. abnormal) and three-class (normal, genu varum, genu valgum) classification tasks using accuracy, macro-precision, macro-recall, and macro-F1 score. Ablation studies quantified the incremental contributions of MTM and SHPM, while Grad-CAM visualization was employed to enhance model interpretability. In the three-class classification task, the AGRM achieved a 70.5% accuracy, 72.1% macro-precision, 71.5% macro-recall, and a macro-F1 score of 0.718; in the binary task, it yielded a 80.3% precision and 79.2% recall. SHPM significantly augmented spatiotemporal feature aggregation, capturing fine-grained gait dynamics, whereas MTM improved model generalization under constrained labeled data scenarios—findings corroborated by ablation experiments. Grad-CAM visualization confirmed the model’s targeted attention to lower extremity regions, particularly the knee joints, aligning with the pathological loci of gait abnormalities. Collectively, our AGRM demonstrates robust performance and generalization in identifying pediatric gait abnormalities, while effectively capturing key pathological gait characteristics. This video-based intelligent approach offers a promising tool for early gait screening in both clinical and community settings, addressing barriers to accessible pediatric musculoskeletal assessment. Full article

Lensless on-chip microscopy enables high-throughput, wide-FOV imaging; however, the Bayer color filter array (CFA) in standard color sensors spatially multiplexes spectral channels, introducing sub-sampling and spectral crosstalk that degrade phase retrieval. We propose a Wirtinger Poly-Gradient Solver (WPGS) for quantitative phase reconstruction with Bayer-filtered color sensors under sequential Red–Green–Blue Light-Emitting Diode (RGB-LED) illumination. The method combines Transport of Intensity Equation (TIE)-based initialization with polychromatic Wirtinger optimization to suppress CFA-induced artifacts and enable pixel super-resolution (PSR). Experiments resolve a $2.76 \mu \mathrm{m}$ linewidth using a $1.85 \mu \mathrm{m}$ pixel-pitch sensor, exceeding the nominal Nyquist limit imposed by pixel sampling. We further demonstrate label-free imaging of HeLa cells and unstained tissue sections, supporting high-throughput digital pathology and offering potential for longitudinal biological observation. Full article

Local Climate Zone (LCZ) classification provides a standardized framework for characterizing urban morphology and its climatic implications. However, most existing remote sensing-based LCZ mapping methods rely on pixel-level classification and multispectral data alone, which limits their ability to capture urban scene heterogeneity and to distinguish structurally similar LCZ classes. In this paper, we propose LCZ-HMSSNet, a deep learning framework for scene-level LCZ classification that integrates PRISMA hyperspectral images with Sentinel-2 multispectral data. The proposed approach exploits both the spectral richness of hyperspectral data and the spatial context provided by multispectral observations, and incorporates a spatial–spectral feature separation mechanism to enhance the discriminability of the fused representations. Experiments conducted across six representative European cities evaluate the proposed method from multiple perspectives, including comparisons with different classification models, data contribution analysis, and structural ablation studies. The results demonstrate that the proposed method consistently outperforms MSI-only and existing LCZ classification approaches, achieving an overall accuracy (OA) of 0.988 and a Kappa of 0.985. In addition, the small-sample experiments indicate the robustness and potential of the proposed model, providing a practical reference for future LCZ mapping in data-scarce scenarios. Full article

Temporal knowledge graphs aim to enhance the dynamic and evolutionary representation of knowledge while enabling time-based reasoning. However, the reasoning based on temporal knowledge graphs in real geographic environments suffers from low accuracy due to the difficulty in effectively utilizing complex spatio-temporal information. Spatial attributes within entities typically encompass both relative and absolute spatial information types. However, during spatio-temporal reasoning, the deep coupling between the quadruple $(entities, relations, timestamp)$ and these two spatial information types is frequently overlooked, as they remain unintegrated in inference predictions. This paper proposes a novel Multi-Task Spatial Recurrent Evolution Graph Convolutional Network (MTS-RE-GCN) framework to enable temporal knowledge graph methods to better reason about spatial entities under time-varying conditions. Experiments on the spatio-temporal dataset and the benchmark dataset (i.e., ICEWS14s, ICEWS18) with spatio-temporal features demonstrate that MTS-RE-GCN significantly outperforms the baseline models (e.g., RE-GCN, TiRGN). For entity prediction tasks, MTS-RE-GCN achieves mean reciprocal rank (MRR) scores of 0.848, 0.739, 0.566, representing improvements of 9.00%, 6.03%, 3.28%, correspondingly. This provides a comprehensive and efficient solution for spatio-temporal entity prediction in temporal knowledge graphs, holding significant implications for spatio-temporal data analysis, event prediction, and related fields. Full article

Deep learning models typically rely on large-scale datasets with accurate annotations, yet real-world applications inevitably suffer from label noise, which severely degrades generalization—particularly for lightweight neural networks with limited capacity. Existing learning with noisy labels methods are mainly designed for over-parameterized models and are often unsuitable for resource-constrained deployment. To address this challenge, we propose a robust framework that integrates a Micro Hybrid Attention Module (MHAM) with knowledge distillation (KD) for lightweight architectures such as MobileNetV3. MHAM employs a decoupled channel–spatial attention design to enhance discriminative feature extraction while suppressing noise-sensitive background responses. From a graph–signal perspective, MHAM can be interpreted as a spectral smoothing operator that improves optimization stability. In addition, knowledge distillation with soft teacher supervision mitigates overfitting to corrupted hard labels and reduces prediction uncertainty. Extensive experiments demonstrate the effectiveness of the proposed method. On FER2013, a real-world noisy facial expression recognition benchmark, our approach achieves 68.5% accuracy with only 0.52M parameters, while reducing optimization variance by 24%. On CIFAR-10 with 40% symmetric label noise, it improves accuracy from 54.85% to 60.10%. On CIFAR-10N with multiple types of real-world human annotation noise, the proposed method consistently achieves 63.9–71.9% accuracy under different noise protocols. These results show that the proposed framework provides an efficient and robust solution for noisy label learning in lightweight facial expression and object classification on edge devices. Full article

Accurate identification of hydrodynamic parameters is essential for high-fidelity modeling and control of unmanned underwater vehicles (UUVs). Compared with towing tank experiments and computational fluid dynamics simulations, system identification based on free-running trial data offers a cost-effective and scalable alternative. However, in real ocean environments, unmodeled lumped disturbances—such as shear currents, stratification-induced buoyancy variations, and wave-induced drift forces—strongly couple with the vehicle’s intrinsic dynamics. Conventional least-squares estimators and physics-informed neural networks tend to absorb environmental effects into the physical parameters, leading to physically inconsistent estimates. To address this challenge, this paper proposes a physics-guided hybrid network (PG-HyNet) with input-domain structural decoupling. The architecture explicitly separates the intrinsic rigid-body dynamics from spatially varying environmental disturbances by assigning dynamics-related states to a physics-constrained branch and position-dependent variables to a residual disturbance branch. A staged training strategy is introduced to stabilize identification and suppress parameter drift during optimization. The framework is validated using high-fidelity simulations incorporating shear currents, density stratification, and wave drift effects, as well as real-world lake trial data. The results demonstrate that PG-HyNet significantly improves robustness against disturbance-induced parameter compensation, enabling physically consistent hydrodynamic parameter recovery while accurately capturing spatially varying environmental disturbance effects. Full article

We investigate the discrete energy behavior and long-time stability of a second-order Crank–Nicolson mixed finite element discretization for the shallow water equations with nonlinear bottom friction. The method combines a compatible ${\mathrm{BDM}}_1$–${\mathrm{DG}}_0$ spatial approximation with a skew-symmetric formulation of the advective terms and a midpoint treatment of dissipative source terms. At the fully discrete level, we derive a precise mechanical energy identity showing that the scheme is energy-consistent;the discrete energy satisfies a balance law consisting of a nonnegative frictional dissipation term and a higher-order midpoint defect of the order $\mathcal{O}\left(\Delta t^3\right)$. Although the method is not unconditionally energy-dissipative, we prove that strict Lyapunov decay holds under a mild CFL-type restriction on the time step. Furthermore, we establish uniform long-time boundedness of the discrete energy and asymptotic recovery of the continuous dissipation law as $\Delta t\to 0$. We also analyze the interaction between nonlinear solver tolerances and energy diagnostics, showing that the observed positive energy increments are controlled, non-accumulating, and intrinsic to the midpoint quadrature structure rather than solver artifacts. The scheme is proven to be precisely well balanced for lake-at-rest equilibria, including nonlinear bottom friction. Comprehensive numerical experiments confirm second-order temporal accuracy, robustness under friction, asymptotic monotonicity under time step refinement, and strict equilibrium preservation. The results provide a rigorous energy-diagnostic framework clarifying when Crank–Nicolson schemes are physically reliable despite the absence of unconditional discrete dissipation. Full article

Religious buildings such as synagogues, churches, and mosques, which are central to religious, cultural, and social life, have served important purposes throughout history as sacred spaces where art, architecture and performance converge. Although these sacred spaces offer unique spatial contexts that deepen individuals’ spiritual experiences through their physical, symbolic, and atmospheric qualities, empirical studies examining this relationship remain limited. This study aims to investigate the impact of biophilic design features within the Yivli Minaret Mosque, one of the oldest Islamic monuments in Antalya, constructed during the 13th-century Anatolian Seljuk Period, on the spiritual experiences of congregation members, and to identify the key psychological mechanisms shaping this relationship. The methodology of the study is based on a mixed-methods approach that combines expert assessments conducted using the Biophilic Interior Design Matrix (BID-M), which integrates proven scientific data with artistic perspective within a historical and symbolic religious structure, with survey data obtained from 359 mosque congregation members. The findings indicate that the mosque exhibits medium-to-high levels of biophilic design characteristics and that the relationship with nature is established indirectly through historical, cultural, and ecological contexts and symbolic representations rather than directly through natural elements. In this respect, the biophilic characteristics of sacred spaces are not merely an artistic and aesthetic approach, but an element that supports individuals’ relationship with nature and their restorative and spiritual experience. Overall, the study reveals that spiritual experience cannot be considered independently of its spatial context and that sacred spaces related to nature support spiritual experience. Full article

Learning image representations with deep self-supervised models is an important task in computer vision, which aims to establish beneficial and general representations from unlabeled images. However, existing efforts train models mainly on high-level features, neglecting lower-level features and their global spatial information, thus limiting the discriminative power of the learned representations. In this work, we propose a representational learning model based on restricted attention feature fusion network (RAFFNet) to improve the quality and generalization of the learned image representations. Specifically, to fully exploit the features in the deep network, we use a self-supervised model on multi-level features to learn more general representations. Meanwhile, a new feature fusion strategy with a dual attention mechanism of channel and space is used for multi-level features, enabling the model training to obtain more important and comprehensive feature information. Furthermore, in order to better extract global spatial information, we devise a simple but effective attentional weighted mask, which restricts the weight of spatial attention and prevents the model from focusing only on local features with high attention weights. Experiments on four public classification datasets, CIFAR-10, CIFAR-100, Tiny ImageNet and ImageNet-1%, and two object detection datasets, PASCAL VOC and COCO, demonstrate that the proposed RAFFNet has better representation performance and generalization ability than most state-of-the-art image representation learning algorithms. Full article

Climate change poses significant risks to Mediterranean aquaculture, with sea surface temperature (SST) identified as a critical stressor affecting cultivated species. This study aims to assess climate-related risks for coastal aquaculture in the Valencian Community (Spain) by analyzing SST spatiotemporal variability and predicting future trends. A multi-method approach was employed, combining ARIMA models for 10-year predictions at eight coastal locations, Bayesian hierarchical models (BHM) fitted via INLA for spatiotemporal analysis of maximum SST and temperature range (2000–2024), and Generalized Additive Models (GAM) to evaluate relationships with climate indices (NAO, AMO, ENSO). Results revealed a consistent warming trend since the 1990s, with ARIMA predictions indicating maximum SST values of 27.2 ± 0.1 °C in September over the next decade. The spatiotemporal model showed effective spatial correlation ranges of 246 km for maximum SST and 207 km for SST range. Anomalous warming years (2003, 2006, 2018, 2023–2024) coincided with documented marine heatwave events. The GAM explained 98.2% of deviance, with AMO showing significant influence ($p<0.001$), while ENSO was not statistically significant. Southern locations (Altea, Campello) currently experience the highest temperatures, but projections indicate Valencia and Sagunto will become the warmest areas. These findings provide essential information for marine spatial planning and recommend a precautionary approach when considering aquaculture relocation towards northern coastal areas. Full article

The main aim of this study was to evaluate the applicability of a low-cost, double-wall gas metal arc welding (GMAW)-based wire arc additive manufacturing (WAAM) process for aluminum alloy AlMg5, with an emphasis on microstructural heterogeneity, layer-dependent defect formation, and their implications for mechanical performance and geometric characteristics. A Taguchi L9 (3³) design of experiments was employed to investigate the influence of welding current (40–60 A), shielding gas flow (10–20 L/min), and arc correction (0–40%) on wall geometry, material utilization, and overall process quality through multi-response optimization. The optimal parameter set (60 A, 15 L/min, 0% arc correction) resulted in a 54.9% improvement in the Grey Relational Grade compared to the lowest-performing configuration. Metallographic analysis revealed heterogeneous grain evolution governed by the multilayer thermal history, with porosity levels ranging from 3.20% to 3.49% and lack-of-fusion defects preferentially concentrated in interlayer and mid-height regions. The fabricated high-wall structure exhibited hardness values between 72 and 85 HV and an average ultimate tensile strength of 175 MPa. The observed mechanical scatter was consistent with localized microstructural heterogeneity and spatial defect distribution. The results demonstrate that geometric evaluation alone is insufficient as a quality metric for WAAM components and must be complemented by metallographic integrity assessment. Overall, the study highlights the importance of direct parameter optimization in double-wall WAAM structures to mitigate defect formation and enhance mechanical reliability under industrially accessible deposition conditions. Full article

When cultivating a selected field crop (alfalfa), we aimed to examine its positive effects on the variability of soil infiltration capacity. A total of 21 monitoring points were proposed for investigating soil hydraulic conductivity on the targeted plot with a total area of 47.64 ha, divided between the irrigated and non-irrigated areas. The plot is located outside the village of Oponice (Slovak Republic) and is managed by VPP Kolíňany. The study of hydraulic conductivity has been ongoing on the selected plot for several years. The presented results come from a two-year experiment, during which work operations related to the cultivation of alfalfa were carried out on the plot. The unsaturated hydraulic conductivity of the soil was assessed several times a year using a Mini Disk Infiltrometer, while soil moisture at monitoring points and the dependence of the measurement date (work operations, weather conditions) were also monitored. The average soil moisture content in the pilot measurements reached 18.77% vol. (CV = 1.44%), in the secondary measurements 17.21% vol. (CV 20.49%), in tertiary measurements 15.27% vol. (CV = 10.38%), and in the last measurements 15.26% vol. (CV = 10%), which ultimately represents a positive result of soil moisture balance. To test the significance of the differences between measurements taken across the entire surveyed plot, a one-factor ANOVA analysis was used to compare the measurement dates. The results showed a statistically significant difference when examining the effect of the time period of soil infiltration capacity monitoring between all measurements (p = 0.004). The mutual combinations of individual measurement dates were mostly significant (p = 0.03 for IDM1, IDM2; p = 0.003 for IDM2, IDM3), except for one case without a significant difference (IDM3, IDM4; p = 0.52). The second hypothesis was confirmed only at some monitoring points, and it can be stated that the irrigated area had a more significant effect on the soil infiltration capacity. The results obtained by the Shapiro–Wilk test and Welch’s test in irrigated and non-irrigated areas at individual dates showed statistically insignificant differences in three cases (IDM1, p = 0.123; IDM3, p = 0.382; IDM4, p = 0.445) and statistically significant in one case (IDM2, p = 0.0175). Based on the hypotheses and the results obtained, it can be said that the work tasks performed have a decisive influence on the infiltration capacity of the soil. The phenomenon of “water resistance” did not manifest itself in our research on soil infiltration capacity. The results were also evaluated using ArcGIS software 10.0 to display the spatial variability of soil hydraulic conductivity. The last application used to evaluate the results was Orange software 3.40.0, using clustering maps and hierarchical clustering. The results also pointed to variability depending on the dates of monitoring. Full article

LiDAR has been widely applied in autonomous driving and mobile robotics. Recently, many studies focus on real-time point cloud segmentation, aiming to achieve higher accuracy while maintaining real-time inference speed. Current real-time methods mostly rely on 2D projection, which inevitably leads to spatial information loss. To address the limitations of 2D projection methods, we propose a Geometry-Enhanced Network called GENet that exploits spatial priors. The network employs an Atrous Separable Range Attention (ASRA) module to explicitly utilize spatial priors from range images, enabling geometry-aware feature aggregation with large receptive field at linear complexity. A Geometry-Context Modulation (GCM) mechanism is then used to calibrate semantic features, incorporating geometric priors while preserving the discriminative ability of original features across different categories. Experiments show that our method achieves efficient information fusion while maintaining real-time performance. Compared to existing methods, GENet requires fewer parameters and less computation, achieving a favorable balance between accuracy and efficiency. Full article

Accurate quantification of spatial distances between fluorescent signals in multi-channel 3D microscopy is essential for understanding genomic organization and gene regulation. However, chromatic aberration introduces systematic spatial offsets between channels that significantly bias distance measurements, particularly at short genomic distances. We present FISH-Dist, an automated computational pipeline for quantitative distance measurements in 3D fluorescence in situ hybridization (FISH) experiments acquired on standard confocal microscopes. Our method combines deep learning-based spot segmentation, 3D Gaussian fitting for sub-pixel localization, and two complementary chromatic aberration correction approaches: affine (ACC) and linear (LCC). We validated the pipeline by measuring the lengths of DNA origami nanorulers and systematically evaluated FISH probe design parameters, including probe spacing, density, and target sequence length. FISH-Dist achieves sub-pixel accuracy in signal detection and substantially reduces inter-channel distance measurement errors. This enables a reproducible quantification of spatial relationships in 3D FISH datasets. Unlike existing tools optimized for long-range chromosomal interactions or requiring super-resolution microscopy, FISH-Dist specifically addresses the technical challenges of standard confocal imaging at short genomic distances, where chromatic aberration has a proportionally greater impact on measurement accuracy. Full article