Search Results (12,938)

The optimization of low-Reynolds-number airfoils for micro unmanned aerial vehicles (UAVs) is challenging due to strong geometric nonlinearities, tight endurance requirements, and the need to maintain performance across multiple operating conditions. Classical surrogate-assisted optimization (SAO) methods combined with genetic algorithms become increasingly expensive and less reliable when class–shape transformation (CST)-based geometries are coupled with several flight conditions. Although deep learning surrogates have higher expressive power, their use in this context is often limited by insufficient local feature extraction, weak adaptation to changes in operating conditions, and a lack of robustness analysis. In this study, we construct a task-specific convolutional neural network–long short-term memory (CNN–LSTM) surrogate that jointly predicts the power factor, lift, and drag coefficients at three representative operating conditions (cruise, forward flight, and maneuver) for the same CST-parameterized airfoil and integrate it into an Non-dominated Sorting Genetic Algorithm II (NSGA-II)-based three-objective optimization framework. The CNN encoder captures local geometric sensitivities, while the LSTM aggregates dependencies across operating conditions, forming a compact encoder–aggregator tailored to low-Re micro-UAV design. Trained on a computational fluid dynamics (CFD) dataset from a validated SD7032-based pipeline, the proposed surrogate achieves substantially lower prediction errors than several fully connected and single-condition baselines and maintains more favorable error distributions on CST-family parameter-range extrapolation samples (±40%, geometry-valid) under the same CFD setup, while being about three orders of magnitude faster than conventional CFD during inference. When embedded in NSGA-II under thickness and pitching-moment constraints, the surrogate enables efficient exploration of the design space and yields an optimized airfoil that simultaneously improves power factor, reduces drag, and increases lift compared with the baseline SD7032. This work therefore contributes a three-condition surrogate–optimizer workflow and physically interpretable low-Re micro-UAV design insights, rather than introducing a new generic learning or optimization algorithm. Full article

The authentication of wood species is of paramount significance to market regulation and product quality control in the construction industry. While classification based on microscopic wood cell structures serves as a critical reference for this task, the high inter-class similarity of cell structures and the inherent biological anisotropy of fine textures pose significant challenges to existing methods. Due to their generic design, standard deep learning models often struggle to capture these fine-grained directional features and suffer from catastrophic feature loss during global pooling, particularly under limited sample conditions. To bridge this gap between general-purpose architectures and the specific demands of wood texture analysis, this paper proposes CM-EffNet, a lightweight fine-grained classification framework based on an improved EfficientNetV2 architecture. Firstly, a data augmentation strategy is employed to expand a collected dataset of 226 wood species from 3673 to 29,384 images, effectively mitigating overfitting caused by small sample sizes. Secondly, a Coordinate Attention (CA) mechanism is integrated to embed positional information into channel attention. This allows the network to precisely capture long-range dependencies between cell walls and vessels cavities, successfully decoding the challenge of textural anisotropy. Thirdly, a MixPooling strategy is introduced to replace traditional global average pooling, enabling the collaborative extraction of background context and salient fine-grained details to prevent the loss of critical micro-features. Systematic experiments demonstrate that CM-EffNet achieves a recognition accuracy of 96.72% and a training accuracy of 98.18%. Comparative results confirm that the proposed model offers superior learning efficiency and classification precision with a compact parameter size. This makes it highly suitable for deployment on mobile terminals connected to portable microscopes, providing a rapid and accurate solution for on-site timber market regulation and industrial quality control. Full article

To address the application requirements of Visual Simultaneous Localization and Mapping (VSLAM) in intracavitary environments and the scarcity of gold-standard datasets for deep learning methods, this study proposes a hybrid attention mechanism combined with U-Net for vascular branch point extraction in endoluminal images (SuperVessel). The network is initialized via transfer learning with pre-trained SuperRetina model parameters and integrated with a vascular feature detection and matching method based on dual branch fusion and structure enhancement, generating a pseudo-gold-standard vascular branch point dataset. The framework employs a dual-decoder architecture, incorporates a dynamic up-sampling module (CBAM-Dysample) to refine local vessel features through hybrid attention mechanisms, designs a Dice-Det loss function weighted by branching features to prioritize vessel junctions, and introduces a dynamically weighted Triplet-Des loss function optimized for descriptor discrimination. Experiments on the Vivo test set demonstrate that the proposed method achieves an average Area Under Curve (AUC) of 0.760, with mean feature points, accuracy, and repeatability scores of 42,795, 0.5294, and 0.46, respectively. Compared to SuperRetina, the method maintains matching stability while exhibiting superior repeatability, feature point density, and robustness in low-texture/deformation scenarios. Ablation studies confirm the CBAM-Dysample module’s efficacy in enhancing feature expression and convergence speed, offering a robust solution for intracavitary SLAM systems. Full article

Background: Arthroscopic anterior cruciate ligament (ACL) reconstruction is widely recognised as the primary treatment for ACL injuries. However, with the increasing incidence of sports-related injuries and growing demand for rehabilitation services, conventional rehabilitation models—largely reliant on therapists’ experience and subjective assessment—are increasingly insufficient to meet the clinical need for precise and individualised rehabilitation programmes. This study aimed to evaluate the effectiveness of a rehabilitation protocol incorporating an artificial intelligence (AI)-based assessment and correction system on functional recovery following ACL reconstruction. Methods: Using convenience sampling, 80 patients undergoing ACL reconstruction between June to December 2024 were recruited for this randomised controlled trial. Participants were randomly assigned to either a control group (n = 40), which received conventional functional exercise training, or a trial group (n = 40), which received rehabilitation intervention guided by an AI-based assessment and correction system. Knee function scores (Lysholm score, IKDC score), Berg Balance Scale (BBS) scores, joint range of motion (ROM), and rehabilitation exercise compliance scores were collected and analysed 1, 2, 3, and 4 months postoperatively. Results: Compared with the control group, the trial group demonstrated significantly greater improvements in Lysholm score, IKDC score, BBS score, and active knee joint ROM (p < 0.05) at postoperative assessment points. Additionally, rehabilitation exercise adherence was significantly higher in the trial group compared to the control group (p < 0.05). Conclusions: Rehabilitation protocols integrating AI-based assessment and correction systems effectively enhance knee function recovery, joint mobility and balance ability following ACL reconstruction. Moreover, these protocols significantly improve rehabilitation exercise adherence, demonstrating superior efficacy compared to conventional rehabilitation approaches. This digital rehabilitation model represents an efficient and promising intervention for postoperative ACL rehabilitation. Full article

Exposure to plant protection products (pesticides) is common among agricultural workers and may represent an underrecognized cause of mucocutaneous disease. We conducted a descriptive cross-sectional survey in agricultural communities in southern Thailand (August–November 2025) to estimate the prevalence, clinical characteristics, and dermatology-specific quality-of-life impact of pesticide-attributed symptoms. Agricultural workers with pesticide use or exposure within the preceding 12 months were recruited via convenience sampling; participants provided consent and completed standardized interviewer-administered questionnaires assessing demographics, pesticide exposure history and application practices, personal protective equipment (PPE) use, self-reported cutaneous and mucosal symptoms (ocular and oral/nasal), and the Dermatology Life Quality Index (DLQI). Of the 354 eligible individuals, 228 participated in the study, and 226 were included in the analyses. The median age was 54 years (interquartile range [IQR], 15), and 82.7% were male. Overall, 14.6% reported pesticide-attributed cutaneous symptoms, 5.3% reported ocular mucosal symptoms, and 0.4% reported oral/nasal mucosal symptoms. Cutaneous manifestations were predominantly symptoms occurring after exposure, with pruritic, erythematous eruptions affecting the arms and hands that typically resolved within 1–7 days after cessation of exposure. Among symptomatic participants, the median DLQI was 0.5 (IQR 3.0); however, DLQI scores were significantly higher among participants who reported pesticide-attributed cutaneous symptoms (p < 0.001) and ocular symptoms (p < 0.001). These findings suggest that pesticide-associated mucocutaneous effects are generally mild yet clinically meaningful, underscoring the need to strengthen PPE training, risk communication, and occupational health surveillance in agricultural settings. Full article

To address the challenge of bearing load monitoring in elastically supported marine shafting systems, a multi-data fusion-based prediction model is constructed. In view of the small-sample nature of measured bearing load data, transfer learning is adopted to migrate the physical relationships embedded in finite element simulations to the measurement domain. A limited number of actual samples are used to correct the simulation data, forming a high-fidelity hybrid training set. The system—supported by air-spring isolators mounted on the raft—is divided into multiple sub-regions according to their spatial layout, establishing local mappings from air-spring pressure variations to bearing load increments to reduce model complexity. On this basis, a Stacking ensemble learning framework is further incorporated into the prediction model to integrate multi-source information such as air-spring pressure and raft strain, thereby enriching the model’s information acquisition and improving prediction accuracy. Experimental results show that the proposed transfer learning-based multi-sub-region bearing load prediction model performs significantly better than the full-parameter input model. Furthermore, the strain-enhanced Stacking-based multi-data fusion bearing load prediction model improves the characterization of shafting features and reduces the maximum prediction error. The proposed multi-data fusion modeling strategy offers a viable approach for condition monitoring and intelligent maintenance of marine shafting systems. Full article

Farmland shelterbelts are crucial elements in safeguarding agricultural ecological security and sustainable development, with their precise extraction being vital for regional ecological monitoring and precision agriculture management. However, constrained by their narrow linear distribution, complex farmland backgrounds, and spectral confusion issues, traditional remote sensing methods encounter significant challenges in terms of accuracy and generalization capability. In this study, six representative deep learning semantic segmentation models—U-Net, Attention U-Net (AttU_Net), ResU-Net, U²-Net, SwinUNet, and TransUNet—were systematically evaluated for farmland shelterbelt extraction using high-resolution Gaofen-6 imagery. Model performance was assessed through four-fold cross-validation and independent test set validation. The results indicate that convolutional neural network (CNN)-based models show overall better performance than Transformer-based architectures; on the independent test set, the best-performing CNN model (U-Net) achieved a Dice Similarity Coefficient (DSC) of 91.45%, while the lowest DSC (88.86%) was obtained by the Transformer-based TransUNet model. Among the evaluated models, U-Net demonstrated a favorable balance between accuracy, stability, and computational efficiency. The trained U-Net was applied to large-scale farmland shelterbelt mapping in the study area (Alar City, Xinjiang), achieving a belt-level visual accuracy of 95.58% based on 385 manually interpreted samples. Qualitative demonstrations in Aksu City and Shaya County illustrated model transferability. This study provides empirical guidance for model selection in high-resolution agricultural remote sensing and offers a feasible technical solution for large-scale and precise farmland shelterbelt extraction. Full article

Pavement roughness is a critical indicator for road maintenance decisions and driving safety assessment. Existing methods primarily rely on multi-source explicit features, which have limited capability in capturing implicit information such as spatial topology between road segments. Furthermore, their accuracy and stability remain insufficient in cross-regional and small-sample prediction scenarios. To address these limitations, we propose a Graph-Fused Natural Gradient Boosting framework (GF-NGB), which combines the spatial topology modeling capability of graph neural networks with the small-sample robustness of natural gradient boosting for high-precision cross-regional roughness prediction. The method first extracts an 18-dimensional set of multi-source features from the U.S. Long-Term Pavement Performance (LTPP) database and derives an 8-dimensional set of implicit spatial features using a graph neural network. These features are then concatenated and fed into a natural gradient boosting model, which is optimized by Optuna, to predict the dual objectives of left and right wheel-track roughness. To evaluate the generalization capability of the proposed method, we employ a spatially partitioned data split: the training set includes 1648 segments from Arizona, California, Florida, Ontario, and Missouri, while the test set comprises 330 segments from Manitoba and Nevada with distinct geographic and climatic conditions. Experimental results show that GF-NGB achieves the best performance on cross-regional tests, with average prediction accuracy improved by 1.7% and 3.6% compared to Natural Gradient Boosting (NGBoost) and a Graph Neural Network–Multilayer Perceptron hybrid model (GNN-MLP), respectively. This study reveals the synergistic effect of multi-source texture features and spatial topology information, providing a generalizable framework and technical pathway for cross-regional, small-sample intelligent pavement monitoring and smart maintenance. Full article

Background and Objectives: This study compared the effects of two posterior-chain strength training strategies on eccentric hamstring strength, jump and sprint performance, and hamstring injury incidence in elite youth soccer players. Materials and Methods: Twenty-three players were randomly allocated to either a Nordic Hamstring Exercise Group (NHEG; n = 11) or a Deadlift + Leg Curl Slides Group (D + LCSG; n = 12). Both groups completed a 9-week in-season resistance training program consisting of one strength-oriented session (MD-4) and one power-oriented session (MD-2) per week, in addition to regular soccer training. Pre- and post-intervention assessments included eccentric hamstring strength (NordBord), countermovement jump (CMJ), and 10 m and 30 m linear sprint performance. Results: Eccentric hamstring strength increased significantly only in the NHEG (p ≤ 0.05), though this improvement did not transfer to enhancements in jump or sprint performance (p > 0.05). No significant changes were observed in the D + LCSG for any variable (p > 0.05), and no between-group differences were found across all performance outcomes. During the 12-week monitoring period, one hamstring injury was recorded, occurring in the NHEG. Conclusions: These findings suggest that, while the NHE elicited greater exercise-specific eccentric strength gains, neither posterior-chain strategy produced improvements in sprint or jump performance. However, given the small sample size and low number of injury events, these trends cannot be attributed with certainty to the implemented protocols, and both programs reported a low incidence of hamstring injuries per 1000 h of exposure with no statistically protective effect associated with the use of the NHE. Full article

Aiming at the problem that the current traditional apple yield estimation methods rely on manual investigation and do not make full use of multi-source information, this paper proposes an apple orchard yield prediction framework combining multispectral remote sensing features and ground features. The framework is oriented to the demand of yield prediction at different scales. It can not only realize the prediction of apple yield at the district and county scales, but also modify the prediction results of small-scale orchards based on the acquisition of orchard features. The framework consists of three parts, namely, apple orchard planting area extraction, district and county large-scale yield prediction and small-scale orchard yield prediction correction. (1) During apple orchard planting area extraction, the samples of some apple planting areas in the study area were obtained through field investigation, and the orchard and non-orchard areas were classified and discriminated, providing a spatial basis for the collection of subsequent yield prediction-related data. (2) In the large-scale yield prediction of districts and counties, based on the obtained orchard-planting areas, the corresponding multispectral remote sensing features and environmental features were obtained using Google Earth engine platform. In order to avoid the noise interference caused by local pixel differences, the obtained data were median synthesized, and the feature set was constructed by combining the yield and other information. On this basis, the feature set was divided and sent to Apple Orchard Yield Prediction Network (APYieldNet) for training and testing, and the district and county large-scale yield prediction model was obtained. (3) During the part of small-scale orchard yield prediction correction, the optimal model for large-scale yield prediction at the district and county levels is utilized to forecast the yield of the entire planting area and the internal local sampling areas of the small-scale orchard. Within the local sampling areas, the number of fruits is identified through the YOLO-A model, and the actual yield is estimated based on the empirical single fruit weight as a ground feature, which is used to calculate the correction factor. Finally, the proportional correction method is employed to correct the error in the prediction results of the entire small-scale orchard area, thus obtaining a more accurate yield prediction for the small-scale orchard. The experiment showed that (1) the yield prediction model APYieldNet (MAE = 152.68 kg/mu, RMSE = 203.92 kg/mu) proposed in this paper achieved better results than other methods; (2) the proposed YOLO-A model achieves superior detection performance for apple fruits and flowers in complex orchard environments compared to existing methods; (3) in this paper, through the method of proportional correction, the prediction results of APYieldNet for small-scale orchard are closer to the real yield. Full article

Infectious diseases caused by bacterial and viral pathogens remain a major global threat, particularly in areas with limited diagnostic resources. Conventional optical techniques are time-consuming, prone to operator errors, and require sophisticated instruments. Colorimetric biosensors, which convert biorecognitive processes into visible color changes, enable simple and low-cost point-of-care testing. Artificial intelligence (AI) enhances decision-making by enabling learning, training, and pattern recognition. Machine learning (ML) and deep learning (DL) improve diagnostic accuracy, but they do not autonomously adapt and are pre-trained on complex color variation, whereas traditional computer-based methods lack analysis ability. This review summarizes major pathogens in terms of their types, toxicity, and infection-related mortality, while highlighting research gaps between conventional optical biosensors and emerging AI-assisted colorimetric approaches. Recent advances in AI models, such as ML and DL algorithms, are discussed with a focus on their applications to clinical samples over the past five years. Finally, we propose a prospective direction for developing robust, explainable, and smartphone-compatible AI-assisted assays to support rapid, accurate, and user-friendly pathogen detection for health and clinical applications. This review provides a comprehensive overview of the AI models available to assist physicians and researchers in selecting the most effective method for pathogen detection. Full article

Acute lymphoblastic leukemia (ALL) is a genetically heterogeneous disease where current clinical practice guidelines remain focused on traditional cytogenetic markers. Despite recent advances demonstrating excellent diagnostic accuracy for gene expression signatures, a discontinuity exists between biomarker validation and clinical implementation. This study aimed to develop and validate a multiparametric gene expression signature using digital PCR (dPCR) to accurately diagnose pediatric ALL, with potential utility for monitoring measurable residual disease (MRD). We analyzed 130 bone marrow aspirates from pediatric patients from four clinical groups: non-leukemia, MRD-negative, MRD-positive and leukemia characterized by immunophenotype. Gene expression of an 8-gene panel (JUP, MYC, NT5C3B, GATA3, PTK7, CNP, ICOSLG, and SNAI1) was quantified by dPCR. The diagnostic performance of individual markers was assessed, and a Random Forest machine learning model was trained to classify active disease. The model was validated using a 5-fold stratified cross-validation approach. Individual markers, particularly JUP, MYC, and NT5C3B, showed good diagnostic accuracy for distinguishing leukemia from non-leukemia. However, integrating all eight markers into a multivariate Random Forest model significantly enhanced performance. The model achieved a mean cross-validated area under the curve (AUC) of 0.908 (±0.041) on receiver operator characteristic (ROC) analysis and 0.961 (±0.019) on Precision–Recall (PR) analysis, demonstrating high reliability and a favorable balance between sensitivity and precision. The integrated model achieved high sensitivity (88.9%) for detecting active disease, particularly at initial diagnosis. Although specificity was moderate (65.0%), the high positive predictive value (PPV 85.1%) and accuracy (81.5%) confirm the clinical utility of a positive result. While the panel showed promising performance for distinguishing MRD-positive from MRD-negative samples, the limited MRD-positive cohort size (n = 11) indicates that validation in larger MRD-focused studies is required before clinical implementation for treatment monitoring. This dPCR-based platform provides accessible, quantitative detection without requiring knowledge of clonal shifts or specific genomic landscape, offering potential advantages for resource-limited settings such as those represented in our Mexican pediatric cohort. Full article

Understanding surface textures through visual cues is crucial for applications in haptic rendering and virtual reality. However, accurately translating visual information into tactile feedback remains a challenging problem. To address this challenge, this paper presents a class-conditional Generative Adversarial Network (cGAN) for cross-modal translation from texture images to vibrotactile spectrograms, using samples from the LMT-108 dataset. The generator is adapted from pix2pix and enhanced with Conditional Batch Normalization (CBN) at the bottleneck to incorporate texture class semantics. A dedicated label predictor, based on a DenseNet-201 and trained separately prior to cGAN training, provides the conditioning label. The discriminator is derived from pix2pixHD and uses a multi-scale architecture with three discriminators, each comprising three downsampling layers. A grid search over multi-scale discriminator configurations shows that this setup yields optimal perceptual similarity measured by Learned Perceptual Image Patch Similarity (LPIPS). The generator is trained using a hybrid loss that combines adversarial, $L_1$, and feature matching losses derived from intermediate discriminator features, while the discriminators are trained using standard adversarial loss. Quantitative evaluation with LPIPS and Fréchet Inception Distance (FID) confirms superior similarity to real spectrograms. GradCAM visualizations highlight the benefit of class conditioning. The proposed model outperforms pix2pix, pix2pixHD, Residue-Fusion GAN, and several ablated versions. The generated spectrograms can be converted into vibrotactile signals using the Griffin–Lim algorithm, enabling applications in haptic feedback and virtual material simulation. Full article

Background: Maternal health behaviors during pregnancy are crucial for maternal and fetal outcomes. While global research has explored that demographic, clinical, and psychosocial determinants significantly influence these behaviors, evidence from low- and middle-income countries (LMICs), including Albania, remains limited. This study aims to evaluate psychosocial perceptions and health behaviors related to lifestyle among pregnant women in a local Albanian community in order to identify which are higher risk subgroups that need targeted and tailored antenatal care interventions. Methods: This multicenter cross-sectional study included 200 pregnant women attending antenatal clinics from May to August 2024 in Vlora city, Albania. Participants were selected using consecutive sampling based on inclusion criteria. Data were collected through a validated questionnaire composed of five sections: demographic/obstetric data; maternal health behaviors; dietary diversity; physical activity, perceived stress; and social support. Clinical and anthropometric measurements were assessed by trained health professionals during antenatal visits. SPSS version 23.0 and binary logistic regression with p-value ≤ 0.05 statistically significant were used for data analysis. Results: Mean age was 28.3 ± 6.4 years, 71% employed and 83.5% urban residents. Key unhealthy behaviors included tobacco use (25.5%), alcohol consumption (10.5%), exposure to toxins (15%), and low dietary diversity (32%). We found significant correlations between low dietary diversity and rural residence (Adj OR = 2.48), hypertension (Adj OR = 6.88), and overweight/obesity (Adj OR = 2.33). Tobacco use was associated with unemployment and alcohol use with unemployment and hypertension variables. Low/moderate social support and high perceived stress were significantly related with multiple unhealthy behaviors, such as low dietary diversity, inadequate physical activity and antenatal care. Conclusions: Unhealthy nutritional behaviors, tobacco and alcohol use and low physical activity are more prevalent risk factors among pregnant women in Vlora city. Priority should be given to vulnerable groups, including rural residents, pregnant women with low social support, high perceived stress and those with hypertension and obesity. Interventions that integrate psychosocial support and health education into antenatal care services are urgently needed to enhance pregnancy outcomes in Albanian communities. Full article

Sensory evaluation is essential for analyzing fish quality, as it describes its organoleptic profile and reflects consumer perception. Attributes such as appearance, smell, taste, and texture can vary depending on the origin of the fish, its diet, and thermal processing. In order to obtain reproducible results, it is necessary to control factors such as temperature, cooking time, and portion thickness during fish sample preparation for testing. This study develops a standardized guide for the sensory evaluation of cooked fish, particularly Sparus aurata. The guide includes detailed preparation protocols, a structured descriptive method, and a tasting sheet to ensure objective, reproducible evaluations that are applicable in research, industry, training, and quality control. Full article