Search Results (2,177)

In the literature, fatigue-loaded reinforced concrete (RC) beams have been the subject of several experimental investigations; however, few numerical studies have specifically examined this behavior. The primary goal of this study is to create and validate a comprehensive nonlinear finite element (FE) modeling framework that combines an existing concrete damage model with specialized modelling techniques (e.g., material modelling, structural modelling, mesh configuration) to forecast the behaviour of reinforced concrete beams under monotonic fatigue loads and track the failure progress. This was accomplished by implementing suitable constitutive and structural models pertaining to concrete and reinforcing steel using VecTor2 finite element software. The Lü concrete damage model, which accounts for the accumulated damage in the concrete at each loading cycle, was taken from the literature to enhance the numerical findings. A number of published experimental tests conducted under monotonic fatigue loading were used to assess the accuracy of the suggested numerical model. The obtained numerical results demonstrated that the FE model may be used to simulate the monotonic fatigue behaviour of various RC beam types. The monotonic fatigue results were significantly improved by applying the Lü concrete damage model. Additionally, the FE model was implemented into practice to offer valuable information on failure mechanisms, fracture patterns, and strain profiles at different loading cycles. Full article

This study introduces SquareSwish, a smooth, self-gated activation $f\left(x\right)=x\sigma {\left(x\right)}^2$, and benchmarks it against ten established activations (ReLU, LeakyReLU, ELU, SELU, GELU, Snake, LearnSnake, Swish, Mish, Hard-Swish) across six CNN architectures (EfficientNet-B1/B4, EfficientNet-V2-M/S, ResNet-50, and Xception) under a uniform transfer-learning protocol. Two geographically grounded datasets are used in this study. FuelRiskMap-TR comprises 7686 satellite images of urban fuel stations in Türkiye, which is semantically enriched with the OpenStreetMap context and YOLOv8-Small rooftop segmentation (mAP@0.50 = 0.724) to support AI-enabled, ICT-integrated risk screening. In a similar fashion, FuelRiskMap-UK is collected, comprising 2374 images. Risk scores are normalized and thresholded to form balanced High/Low-Risk labels for supervised training. Across identical training settings, SquareSwish achieves a top-1 validation accuracy of 0.909 on EfficientNet-B1 for FuelRiskMap-TR and reaches 0.920 when combined with SELU in a simple softmax-probability ensemble, outperforming the other activations under the same protocol. By squaring the sigmoid gate, SquareSwish more strongly attenuates mildly negative activations while preserving smooth, non-vanishing gradients, tightening decision boundaries in noisy, semantically enriched Earth-observation settings. Beyond classification, the resulting city-scale risk layers provide actionable geospatial outputs that can support inspection prioritization and integration with municipal GIS, offering a reproducible and low-cost safety-planning approach built on openly available imagery and volunteered geographic information. Full article

The growing threat of unauthorized drones to ground-based critical infrastructure necessitates efficient ground-to-air surveillance systems. This paper proposes a lightweight framework for UAV detection and altitude measurement from a fixed ground perspective. We introduce Ghost-YOLOv11n, an optimized detector that integrates GhostConv modules into YOLOv11n, reducing computational complexity by 12.7% while achieving 98.8% mAP_0.5 on a comprehensive dataset of 8795 images. Deployed on a LuBanCat4 edge device with Rockchip RK3588S NPU acceleration, the model achieves 20 FPS. For stable altitude estimation, we employ an Extended Kalman Filter to refine measurements from a monocular ranging method based on similar-triangle geometry. Experimental results under ground monitoring scenarios show height measurement errors remain within 10% up to 30 m. This work provides a cost-effective, edge-deployable solution specifically for ground-based anti-drone applications. Full article

Assurance of the integrity of every weld joint is highly desirable, and defect detection methods that can be applied to welds at high temperatures immediately after welding are required. The laser ultrasonic (LU) method, which generates ultrasonic waves in the target via pulsed laser irradiation, is a well-known technique for non-contact defect detection during welding. Ultrasonic waves excited in ablation mode exhibit large amplitudes and predominantly surface-normal propagation, which has driven extensive research into their application for weld inspection. However, owing to the size and weight of conventional equipment, such systems have largely been limited to bench-top experimental setups. To address this, we developed an LU robotic system incorporating a compact, lightweight laser source and an improved signal-processing system. We conducted experiments to measure signals and to detect backside slits in flat plates and blowholes in lap-fillet welds. Additionally, a method to improve the sensitivity of laser interferometers was investigated and demonstrated on smut-covered areas near weld beads. Full article

The Carboniferous coal seams in Northeast Kazakhstan remain insufficiently investigated, with a lack of comprehensive mineralogical and geochemical assessments necessary to understand the geological processes controlling coal quality. This study examines 15 coal samples from the Karagandy Coal Formation (KCF) at the Saradyr and Bogatyr mines using proximate and ultimate analyses, FTIR, XRD, SEM–EDS, ED-XRF, and ICP-OES, providing the first detailed comparison of mineralogical and geochemical characteristics—including depositional signals and inorganic constituent distribution—between these mines within the KCF. The coals exhibit an average ash yield of 24.1% on a dry basis, volatile matter of 21.6% on a dry and ash-free basis, and low moisture content of 1.1% (air-dry), with low sulfur levels of 0.7% in whole coal across both mines. Mineralogical composition is dominated by quartz and clay minerals, with minor pyrite, apatite, chalcopyrite, and rutile. Major oxides in the coal ash average 68.2% SiO₂ and 19.5% Al₂O₃, followed by Fe₂O₃, K₂O, and TiO₂ (3–12.1%). Among the 24 identified trace elements, Sm is the most abundant at 6.3 ppm with slight enrichment (CC = 2.8), Lu remains at normal levels (CC < 1), and most other elements are depleted (CC < 0.5). The Al₂O₃/TiO₂ ratios (3.8–10.8) indicate contributions from intermediate to mafic parent materials. The detrital mineralogy, parting compositions, and elevated ash content indicate significant accommodation space development during or shortly after peat accumulation, likely within a vegetated alluvial plain depression. These findings provide new insights into the depositional environment and coal-forming processes of the KCF and contribute to regional assessments of coal quality and resource potential. Full article

The deep learning paradigm is progressively shifting from non-smooth activation functions, exemplified by ReLU, to smoother alternatives such as GELU and SiLU. This transition is motivated by the fact that non-differentiability introduces challenges for gradient-based optimization, while an expanding body of research demonstrates that smooth activations yield superior convergence, improved generalization, and enhanced training stability. A central challenge, however, is how to systematically transform widely used non-smooth functions into smooth counterparts that preserve their proven representational strengths while improving differentiability and computational efficiency. To address this, we propose a general activation smoothing framework grounded in mollification theory. Leveraging the Epanechnikov kernel, the framework achieves statistical optimality and computational tractability, thereby combining theoretical rigor with practical utility. Within this framework, we introduce Smoothed ReLU (S-ReLU), a novel second-order continuously differentiable (C²) activation derived from ReLU that inherits its favorable properties while mitigating inherent drawbacks. Extensive experiments on CIFAR-10, CIFAR-100, and ImageNet-1K with Vision Transformers and ConvNeXt consistently demonstrate the superior performance of S-ReLU over existing ReLU variants. Beyond computer vision, large-scale fine-tuning experiments on language models further show that S-ReLU surpasses GELU, underscoring its broad applicability across both vision and language domains and its potential to enhance stability and scalability. Full article

High-risk and locally advanced prostate cancer represents 20–25% of new diagnoses of prostate cancer and is associated with high rates of recurrence, morbidity, and mortality. The neoadjuvant window provides a unique opportunity for systemic control prior to definitive therapy with radical prostatectomy or radiotherapy (RT). Early trials with first-generation androgen deprivation therapy (ADT) achieved pathological downstaging but no survival benefit. In the 2000s, the advent of chemohormonal regimes using docetaxel provided excitement but mixed results tempered expectations and is now not recommended prior to surgery. Second-generation androgen receptor pathway inhibitors (ARPIs) combined with ADT have demonstrated significant survival benefit in metastatic prostate cancer and are currently being evaluated in large phase III trials in the neoadjuvant setting. RT remains an alternative curative modality, and recent data highlights similar issues to surgery in eradicating micrometastatic disease despite excellent local control. This has driven parallel efforts to evaluate intensified systemic therapy in the pre-RT/neoadjuvant settings. In addition to the excitement surrounding ARPIs, radioligand therapy, such as [¹⁷⁷Lu]Lu-PSMA-617 has shown promise in the neoadjuvant setting and continues to be investigated. Future research aims to incorporate genomic and molecular factors to enable personalised neoadjuvant therapies by identifying damage immunologically responsive subtypes that may derive greater benefit from immune-directed therapies in the peri-operative setting. This narrative review synthesises current evidence for neoadjuvant therapies in high-risk prostate cancer and future directions. Full article

This study proposes a hybrid computational framework that integrates the Extended Finite Element Method (XFEM), Artificial Neural Network (ANN), and Monte Carlo simulation to evaluate the fatigue crack propagation and reliability of reinforced concrete (RC) bridge decks. First, XFEM was employed to simulate crack initiation and propagation under cyclic loading based on the statistical distributions of the Paris law parameters C and m. The fatigue life data generated from these simulations were used to train a multilayer feedforward ANN optimized with the Adam algorithm and the ReLU activation function. The trained network achieved a high prediction accuracy (R² = 0.99, MAPE = 0.977%) and demonstrated strong generalization capability for predicting the XFEM-derived fatigue life. Subsequently, 10,000 Monte Carlo samples of C and m were analyzed using the trained ANN to perform probabilistic fatigue life assessment. The results revealed a nonlinear degradation pattern in reliability: the structural reliability remained high at low fatigue cycles but decreased sharply once a critical threshold of approximately 1.45 × 10⁹ cycles was reached. When actual bridge traffic was considered, the deck maintained a reliability of 0.99 after 23 years and 0.95 after 67 years of service. Compared with the XFEM, the ANN-based prediction improved computational efficiency by more than 10⁴ times while maintaining satisfactory accuracy. The proposed hybrid framework effectively combines deterministic simulation, probabilistic analysis, and data-driven modeling, providing a rapid and reliable approach for predicting fatigue life and evaluating the reliability of concrete bridge structures. Full article

Human organoids are 3D cell culture models that precisely replicate the microenvironment of real organs. In organoid-based experiments, assessing whether the internal vasculature has formed normally is essential for ensuring the reliability of experimental results. However, conventional vasculature assessment relies on manual inspection by researchers, which is time-consuming and prone to variability caused by subjective judgment. This study proposes a lightweight deep learning model for automatic classification of normal and abnormal vasculature in vascular organoid images. The proposed model is based on EfficientNet by replacing the activation function SiLU with ReLU and removing the Squeeze-and-Excitation (SE) blocks to reduce computational complexity. The dataset consisted of vascular organoid images obtained from co-culture experiments. Data augmentation and noise addition were performed to alleviate class imbalance. Experimental results show that the proposed Modified 3 models (B0, B1, B2) achieved accuracy of 0.90, 0.99, and 1.00, respectively, with corresponding inference speed of 51.1, 36.0, and 32.4 FPS on the CPU, demonstrating real-time inference capability and an average speed improvement of 70% compared to the original models. This study presents an efficient automated analysis framework that enables quantitative and reproducible vasculature assessment by introducing a lightweight model that maintains high accuracy and supports real-time processing. Full article

Background: The role of diaphragmatic function in transient tachypnea of the newborn (TTN) remains poorly understood. This study aimed to compare diaphragmatic ultrasound parameters between neonates with TTN requiring non-invasive ventilation (NIV) and healthy neonates. Secondary objectives include the relationships between these parameters and gestational age (GA), birth weight (BW), and the evaluation of inter-operator reproducibility. Methods: This prospective observational pilot study involved neonates with GA ≥ 34 weeks with clinical and ultrasound diagnosis of TTN treated with NIV. An equal number of healthy neonates served as controls. Diaphragmatic and lung ultrasound were performed on day 1 (T0) and day 2 (T1) of life. Measurements included end-inspiratory and end-expiratory diaphragmatic thickness (DTi and DTe, respectively), diaphragmatic excursion (DE), and Lung Ultrasound Score (LUS). Inter-operator reproducibility was tested in 31 neonates (62 scans in total). Results: Forty neonates were enrolled (20 TTN, 20 controls). DE was significantly higher in controls compared with TTN neonates (4.6 ± 0.9 mm vs. 5.4 ± 1.3 mm, p = 0.03) and increased from T0 to T1 in the control group (4.6 ± 1.1 mm vs. 5.4 ± 1.3 mm, p = 0.04), while no significant variation was observed in TTN cases. A negative correlation, approaching significance, was found between DE and LUS at T1 (p = 0.05). DTi and DTe increased linearly with GA and BW (p < 0.001). Bland–Altman analysis showed low bias and acceptable limits of agreement between measurements. Conclusions: The underlying pulmonary disease may influence diaphragmatic function in neonates with TTN. The integration of lung and diaphragmatic ultrasound could be useful for monitoring disease progression and follow-up. Full article

Among various activation functions, the Rectified Linear Unit (ReLU) has become the most widely adopted due to its computational simplicity and effectiveness in mitigating the vanishing-gradient problem. In this work, we investigate the advantages of employing ReLU as the activation function and establish its theoretical significance. Our analysis demonstrates that ReLU-based neural networks possess the universal approximation property. In addition, we provide a theoretical explanation for the phenomenon of neuron death in ReLU-based neural networks. We further validate the effectiveness of this explanation through empirical experiments. Full article

Background: Pleural effusion (PE) is a frequent complication in patients with malignancies and is often associated with poor prognosis. Lung ultrasound (LUS) has become an indispensable bedside tool for detecting, characterizing, and guiding the management of pleural effusions. Methods: This narrative review summarizes the current evidence on the diagnostic performance of LUS for PE in cancer patients, emphasizing recent advances in functional ultrasound techniques. Results: B-mode LUS can detect small-volume effusions and estimate their volume. Sonographic features such as echogenicity, septations, and pleural abnormalities can help differentiate transudative from exudative effusions. Shear-wave elastography and contrast-enhanced ultrasound provide additional functional information on tissue stiffness and perfusion. This information may help distinguish between malignant and benign pleural lesions and facilitate targeted biopsy when cytology is nondiagnostic. Compared with computed tomography, LUS offers superior evaluation of juxtadiaphragmatic and pleural surface abnormalities. It facilitates safe, real-time thoracocentesis. Recent innovations, including improved quality, affordable handheld ultrasound systems and artificial intelligence-based analysis, are expected to further enhance diagnostic precision and accessibility. Conclusions: Although LUS is a sensitive and versatile tool for assessing PE in cancer patients, it has limited diagnostic accuracy in distinguishing between benign and malignant effusions. Advanced techniques, such as shear-wave elastography and contrast-enhanced ultrasound, may further support the differentiation of malignant and benign diseases. Ongoing technological advances are likely to enhance the diagnostic accuracy and accessibility of lung ultrasound. Full article

Understanding how land tenure and socio-economic pressures shape landscape transformation is critical for sustainable management in biodiversity-rich regions. This study examines three decades (1987–2017) of land use and land cover (LU&LC) change in the Ngerenyi area of the Taita Hills, Kenya, by integrating multispectral Landsat analysis with household survey data. Harmonized pre-processing and supervised classification of four LU&LC classes, agriculture, built-up areas, high-canopy vegetation, and low-canopy vegetation, achieved overall accuracies above 80% and Kappa values exceeding 0.75. Transition modeling using the Minimum Information Loss Transition Estimation (MILTE) approach, combined with net-versus-swap metrics, revealed persistent decline and fragmentation of high-canopy vegetation, cyclical transitions between agriculture and low-canopy vegetation, and the near-irreversible expansion of built-up areas. Low-canopy vegetation exhibited the highest dynamism, reflecting both degradation from canopy loss and natural regeneration from fallowed cropland. Household surveys (n = 141) identified agricultural expansion, charcoal production, fuelwood extraction, and population growth as the dominant perceived drivers, with significant variation across tenure categories. The population in Taita Taveta County increased from 205,334 in 2009 to 340,671 in 2019, reinforcing documented pressures on land resources and woody biomass. As part of the Eastern Arc biodiversity hotspot, the landscape’s diminishing high-canopy patches underscore the importance of conserving undisturbed vegetation remnants as ecological baselines and biodiversity refuges. The findings highlight the need for tenure-sensitive, landscape-scale planning that integrates private landowners, regulates subdivision, promotes agroforestry and alternative energy options, and safeguards remaining high-canopy vegetation to enhance ecological resilience while supporting local livelihoods. Full article

The Transformer architecture has achieved remarkable success across numerous computer vision tasks due to its superior capability for global dependency modeling. However, the high computational complexity and hardware-unfriendly operations such as Layer Normalization (LN), Softmax, and GELU severely hinder its deployment on resource-constrained platforms. To address these challenges, this paper proposes a hardware-friendly CNN-Transformer hybrid pyramid architecture that effectively balances accuracy, efficiency, and deployability. The proposed model integrates convolutional bottlenecks with Transformer encoders to capture both local and global contextual information while maintaining low computational cost. A pyramid feature extraction structure is further introduced to enhance multi-scale semantic representation. To improve hardware efficiency, we redesign key nonlinear components by introducing hardware-friendly activation, normalization, and Softmax approximations. Specifically, GELU and LN are replaced by ReLU and Batch Normalization (BN), and a simplified logarithmic-exponential formulation termed Softmax2 is proposed, which eliminates complex exponential and division operations, significantly reducing hardware implementation cost. Extensive experiments demonstrate the effectiveness of the proposed framework. The experimental results validate that the proposed architecture offers a promising and practical solution for real-time and embedded vision applications. Full article

The Jiulongnao W–Sn ore field in the eastern Nanling Range is characterized by large-scale early Yanshanian magmatic activity and W–Sn mineralization. In recent years, increasing attention has been given to the close relationship between Indosinian magmatic activity and Sn mineralization. The Aotou quartz vein-type Sn deposit is unique for only Sn mineralization without W during the Indosinian period. Seventeen thin-to-thick cassiterite–quartz veins are densely distributed in Ordovician metasandstone and slate, and these veins extend down to the top of the concealed granite. However, both the diagenetic age and the petrological characteristics of the concealed granite remain unclear. This contribution shows that the Aotou muscovite intrusion is a highly fractionated S-type pluton, characterized by a peraluminous, high-K composition, enrichment in LREEs, and depletion of Ba, Sr, Ti, and Eu. In this study, LA–ICP–MS zircon U–Pb dating of the concealed muscovite granite yields emplacement ages of 238.7 ± 1.0 Ma and 225.4 ± 0.9 Ma, indicating that at least two stages of magmatic intrusion occurred in the Triassic, with the diagenetic environment transitioning from a compressional setting to an extensional setting. The ε_Hf(t) values during the two stages are −0.98 to −0.95 and −0.98 to −0.96, and the T_DM2 values are 1.78–2.08 Ga and 1.78–2.06 Ga, indicating that two-stage magma was derived from the late Paleoproterozoic lower crustal materials. Comprehensive analysis reveals that the second stage of Indosinian magma intrusion (232–225 Ma) in the Jiulongnao ore field is closely related to Sn mineralization, and the northern Wenying pluton has good prospecting potential for quartz vein-type Sn(–W) deposits. Full article