Search Results (30,007)

Illegal mining poses significant environmental and economic challenges, and effective monitoring is essential for regulatory enforcement. This study evaluates the potential of the U-Net deep learning model for detecting mining activities using Sentinel-2 satellite imagery over the Strzegom region in Poland. We prepared annotated datasets representing various land cover classes, including active and inactive mineral extraction sites, agricultural areas, and urban zones. U-Net was trained and tested on these data, and its classification accuracy was assessed against common Google Earth Engine (GEE) classifiers such as Random Forest, CART, and SVM. Accuracy metrics, including Overall Accuracy, Producer’s Accuracy, and F1-score, were computed. Additional analyses compared model performance for detecting licensed versus potentially illegal mining areas, supported by integration with publicly available geospatial datasets (MOEK, MIDAS, CORINE). The results show that U-Net achieved higher detection accuracy for mineral extraction sites than the GEE classifiers, particularly for small and spatially heterogeneous areas. This approach demonstrates the feasibility of combining deep learning with open geospatial data for supporting mining activity monitoring and identifying potential cases of unlicensed extraction. Full article

Potato (Solanum tuberosum L.) is an important global food crop, being greatly valued for its high carbohydrate content and nutritional profile. In response to the world population’s rapid growth and the increasing need for nutritionally enhanced food quality, potato biofortification has become a key focus of agronomic research. This study investigated the effect of calcium (Ca) biofortification on two potato cultivars (Picasso and Rossi) cultivated in Portugal, assessing its impact on the photosynthetic functioning and the Ca content and distribution of tubers. At the beginning of the tuberization stage, seven foliar applications of CaCl₂ or Ca-EDTA at 12 kg ha⁻¹ were performed. The application of Ca-EDTA led to an increased Ca content in peeled tubers of Picasso (37%) and Rossi (16%), and 88% and 79% in unpeeled tubers, in the same cv. order and as compared to their controls, with Ca predominantly accumulating in the epidermis/peel region. Photosynthetic performance was negatively impacted by the Ca-EDTA treatment in Picasso but not in Rossi, which was reflected in the significant declines in net photosynthesis (P_n) and maximal (F_v/F_m) and actual (F_v′/F_m) photochemical efficiency of photosystem II. Additionally, both genotypes showed negative impacts (greater in Picasso) on the quantum yield of non-cyclic electron transport (Y_(II)) and photochemical quenching (q_L) after five foliar applications. This contrasted with the absence of negative impacts under the use of CaCl₂, which resulted in 17.1% (Picasso) and 29.5% (RFossi) increase in Ca content in peeled tubers, without any significant differences between the unpeeled tubers of both cvs. Moreover, only with CaCl₂, the tuber weight and yield were not negatively impacted. These findings pointed out that, although with a lower Ca increase in the tubers, CaCl₂ was the best suitable option for the Ca biofortification of these cvs. at the applied doses. Full article

Background: The prevalence of stress-related health issues is becoming increasingly significant. This study aimed to examine the relationships between work stress, home stress, overall stress, and individual behavioral and perceptual characteristics among middle-aged employees in the service sector. Methods: Physical activity, diet, and perceptions were assessed using the WellBeNet application (2.10.2, INRAE, Clermont-Ferrand, France) while perceived stress levels were evaluated through an online questionnaire during a one-week period. The associations between stress levels and individual and behavioral characteristics were examined using multiple linear regressions and analyses of variance. Results: General stress was significantly influenced by both work and home stress. Home stress was positively correlated with the perception of one’s silhouette in red, the increasing consumption of dairy products, and the decreasing consumption of vegetables. Work stress was inversely correlated with age and positively correlated with body shape. Conclusions: Our study identified various context markers of stress—including age, body shape, food intake, and color of the silhouette. These markers could be used in subsequent intervention studies to demonstrate causal links. Full article

Addressing the carbon footprint in pig production is a fundamental technical basis for achieving carbon neutrality and peak carbon emissions. Only by systematically studying the carbon footprint can the goals of carbon neutrality and peak carbon emissions be effectively realized. This study aims to reduce the carbon footprint through optimized feeding strategies based on minimizing carbon emissions. To this end, this study conducted a full-lifecycle monitoring of the carbon footprint during pig growth from December 2024 to May 2025, optimizing feeding strategies using a real-time pig weight estimation model driven by deep learning to reduce resource consumption and the carbon footprint. We introduce EcoSegLite, a lightweight deep learning model designed for non-contact real-time pig weight estimation. By incorporating ShuffleNetV2, Linear Deformable Convolution (LDConv), and ACmix modules, it achieves high precision in resource-constrained environments with only 1.6 M parameters, attaining a 96.7% mAP50. Based on full-lifecycle weight monitoring of 63 pigs at the Pianguan farm from December 2024 to May 2025, the EcoSegLite model was integrated with a life cycle assessment (LCA) framework to optimize feeding management. This approach achieved a 7.8% reduction in feed intake, an 11.9% reduction in manure output, and a 5.1% reduction in carbon footprint. The resulting growth curves further validated the effectiveness of the optimized feeding strategy, while the reduction in feed and manure also potentially reduced water consumption and nitrogen runoff. This study offers a data-driven solution that enhances resource efficiency and reduces environmental impact, paving new pathways for precision agriculture and sustainable livestock production. Full article

Background/Objectives: Wet age-related macular degeneration (AMD) is a progressive retinal disease characterized by macular neovascularization (MNV). Currently, the standard treatment for wet AMD is intravitreal anti-VEGF administration, which aims to control disease activity by suppressing neovascularization. In clinical practice, the decision to continue or discontinue treatment is largely based on the presence of fluid on optical coherence tomography (OCT) and changes in visual acuity. However, discrepancies between anatomic and functional responses can occur during these assessments. Methods: This article presents an artificial intelligence (AI)-based classification model developed to objectively assess the response to anti-VEGF treatment in patients with AMD at 3 months. This retrospective study included 120 patients (144 eyes) who received intravitreal bevacizumab treatment. After bevacizumab loading treatment, the presence of subretinal/intraretinal fluid (SRF/IRF) on OCT images and changes in visual acuity (logMAR) were evaluated. Patients were divided into three groups: Class 0, active disease (persistent SRF/IRF); Class 1, good response (no SRF/IRF and ≥0.1 logMAR improvement); and Class 2, limited response (no SRF/IRF but with <0.1 logMAR improvement). Pre-treatment and 3-month post-treatment OCT image pairs were used for training and testing the artificial intelligence model. Based on this grouping, classification was performed with a Siamese neural network (ResNet-18-based) model. Results: The model achieved 95.4% accuracy. The macro precision, macro recall, and macro F1 scores for the classes were 0.948, 0.949, and 0.948, respectively. Layer Class Activation Map (LayerCAM) heat maps and Shapley Additive Explanations (SHAP) overlays confirmed that the model focused on pathology-related regions. Conclusions: In conclusion, the model classifies post-loading response by predicting both anatomic disease activity and visual prognosis from OCT images. Full article

Immune checkpoint inhibitors (ICIs) display efficacy in non-small cell lung cancers (NSCLCs) with brain metastases (BMs) and studies suggest potential synergy with cranial radiation (CR). However, population-based evaluations of optimal time between ICI-CR combinations are limited in the US. Using SEER-Medicare database (2010–2019), we analyzed patients aged ≥65 years with NSCLC and BM receiving ICI-CR within 6 months of diagnosis, excluding those receiving targeted therapies. First treatment after diagnosis (ICI or CR) was defined as index treatment; followed by subsequent treatment. Findings were validated using an independent cohort from the TriNetX LIVE™ Platform. Patients were grouped by interval between the end of the index treatment and the start of the subsequent treatment: ≤15 days (n = 117), 16–30 days (n = 42), and >30 days (n = 77). Overall survival (OS) was measured from the start of the subsequent treatment until death, end of insurance coverage, or study end. Kaplan–Meier survival curves and multivariable Cox proportional hazards models estimated differences between groups. Among 236 patients, median OS was 134 days, 92 days, and 209 days, respectively. No significant OS differences were found across intervals. However, a survival benefit emerged approximately 300 days after follow-up when ICI was administered within 15 days of CR. These findings offer insight into treatment sequencing in NSCLC with BM and support further investigation in larger cohorts. Full article

Understanding genetic variation in commercially valuable tree species is essential for improving breeding and conservation efforts. This study investigates genetic variation, heritability, and trait relationships in Pterocarpus macrocarpus, a vital hardwood species for Thailand’s reforestation initiatives. We evaluated growth (height and diameter), morphology (biomass dry weight and specific leaf weight), and physiological traits (net photosynthesis [A], transpiration rate [E], and water-use efficiency [WUE]) across 112 open-pollinated families from six natural populations under controlled nursery conditions over 30 weeks. Using a randomised complete block design, variance and covariance analyses were conducted to estimate genetic parameters. Seedling survival reached 95%, confirming favourable conditions for genetic expression. There were significant differences among populations and families within populations in growth and biomass. In contrast, physiological traits showed notable family-level variation (A, E, WUE) and only population effects for WUE. Residual variance was predominant across traits, indicating considerable within-family variation. Growth and biomass exhibited moderate to high heritability (individual: 0.39–1.00; family: 0.61–0.90), while specific leaf weight and shoot-to-root ratio had lower heritability at the individual level. Physiological traits showed low to moderate heritabilities (individual: 0.26–0.43; family: 0.47–0.62), with maternal effects via seed weight significantly influencing early growth. The heritability of height decreased over time, whereas the heritability of diameter remained stable. Strong genetic correlations among growth and biomass suggest the potential for combined selection gains. However, physiological traits show weak or no correlations with growth, highlighting their independent genetic control. Variation at the population level in growth and WUE may reflect adaptive responses to seed-source environments. Our findings support the use of nursery-based screening as a cost-effective method for the early identification of high-quality families. WUE is a promising focus for breeding programs targeting drought-prone regions. This study provides key insights for advancing the genetic improvement and conservation of P. macrocarpus, emphasizing the importance of incorporating physiological traits into breeding and conservation strategies. Full article

Mediterranean rocky reef habitats are ecologically valuable yet increasingly degraded due to cumulative human pressures, necessitating efficient, large-scale ecological status assessments to inform management. Macroalgal communities are widely used as indicators of rocky reef conditions and are typically assessed via photoquadrat sampling. However, the manual annotation of benthic images remains time-consuming and costly. This study evaluates the performance of CoralNet (version 1.0), an AI-assisted image annotation platform, using a pre-annotated dataset of 2537 photoquadrat images from 89 rocky reef sites in the Aegean Sea, Greece, classified into 23 taxonomic and morphofunctional groups. Half of the dataset was used to iteratively train CoralNet classifiers, while the remainder was used to compute the reef-EBQI index and compare ecological status estimates with those derived from manual annotations. The classifier accuracy improved with training volume, reaching 67% using the entire dataset. Reef-EBQI scores derived from CoralNet showed 87% agreement with the manual classifications. Despite challenges and limitations, AI-assisted annotation proved effective in regional-scale ecological assessments based on broad taxonomic and morphofunctional categories. Automated tools like CoralNet can reduce post-processing bottlenecks and enable scalable, cost-effective monitoring, especially when integrated with standardized protocols and citizen science initiatives. Full article

Texture, a significant visual attribute in images, plays an important role in many pattern recognition tasks. While Convolutional Neural Networks (CNNs) have been among the most effective methods for texture analysis, alternative architectures such as Vision Transformers (ViTs) have recently demonstrated superior performance on a range of visual recognition problems. However, the suitability of ViTs for texture recognition remains underexplored. In this work, we investigate the capabilities and limitations of ViTs for texture recognition by analyzing 25 different ViT variants as feature extractors and comparing them to CNN-based and hand-engineered approaches. Our evaluation encompasses both accuracy and efficiency, aiming to assess the trade-offs involved in applying ViTs to texture analysis. Our results indicate that ViTs generally outperform CNN-based and hand-engineered models, particularly when using strong pre-training and in-the-wild texture datasets. Notably, BeiTv2-B/16 achieves the highest average accuracy (85.7%), followed by ViT-B/16-DINO (84.1%) and Swin-B (80.8%), outperforming the ResNet50 baseline (75.5%) and the hand-engineered baseline (73.4%). As a lightweight alternative, EfficientFormer-L3 attains a competitive average accuracy of 78.9%. In terms of efficiency, although ViT-B and BeiT(v2) have a higher number of GFLOPs and parameters, they achieve significantly faster feature extraction on GPUs compared to ResNet50. These findings highlight the potential of ViTs as a powerful tool for texture analysis while also pointing to areas for future exploration, such as efficiency improvements and domain-specific adaptations. Full article

Drowning constitutes the leading cause of injury-related fatalities among adolescents. In swimming pool environments, traditional manual surveillance exhibits limitations, while existing technologies suffer from poor adaptability of wearable devices. Vision models based on YOLO still face challenges in edge deployment efficiency, robustness in complex water conditions, and multi-scale object detection. To address these issues, we propose YOLO11-LiB, a drowning object detection model based on YOLO11n, featuring three key enhancements. First, we design the Lightweight Feature Extraction Module (LGCBlock), which integrates the Lightweight Attention Encoding Block (LAE) and effectively combines Ghost Convolution (GhostConv) with dynamic convolution (DynamicConv). This optimizes the downsampling structure and the C3k2 module in the YOLO11n backbone network, significantly reducing model parameters and computational complexity. Second, we introduce the Cross-Channel Position-aware Spatial Attention Inverted Residual with Spatial–Channel Separate Attention module (C2PSAiSCSA) into the backbone. This module embeds the Spatial–Channel Separate Attention (SCSA) mechanism within the Inverted Residual Mobile Block (iRMB) framework, enabling more comprehensive and efficient feature extraction. Finally, we redesign the neck structure as the Bidirectional Feature Fusion Network (BiFF-Net), which integrates the Bidirectional Feature Pyramid Network (BiFPN) and Frequency-Aware Feature Fusion (FreqFusion). The enhanced YOLO11-LiB model was validated against mainstream algorithms through comparative experiments, and ablation studies were conducted. Experimental results demonstrate that YOLO11-LiB achieves a drowning class mean average precision (DmAP50) of 94.1%, with merely 2.02 M parameters and a model size of 4.25 MB. This represents an effective balance between accuracy and efficiency, providing a high-performance solution for real-time drowning detection in swimming pool scenarios. Full article

Background: Ischemic stroke (IS) is a significant cause of global mortality and disability. Yiqi Huoxue Jieyu granules (YHJGs) show therapeutic potential for IS, but their mechanisms remain unclear. This study investigated YHJGs’ effects through network pharmacology, molecular docking, and experimental validation. Methods: Active YHJG components and IS targets were identified from TCMSP, GeneCards, and DisGeNET databases. Network analysis and molecular docking (AutoDock Vina) were performed. In vivo studies used 72 male Sprague-Dawley rats (MCAO model) divided into sham, model, nimodipine (10.8 mg/kg), and three YHJG dose groups (0.72, 1.44, 2.88 g/kg). Assessments included neurological scores, TTC staining, histopathology, and molecular analyses (qPCR/Western blot). Results: Network analysis identified 256 shared targets between YHJG and IS, with PI3K-AKT and MAPK as key pathways. Molecular docking showed strong binding between YHJG compounds (e.g., quercetin) and core targets (AKT1, ERK1/2). YHJG treatment significantly improved neurological function (p < 0.01), reduced infarct volume (p < 0.01), and attenuated neuronal damage. The expression of IL-1β, TNF-α, IL-6, AKT1, and pERK1/2/ERK1/2 significantly increased in the MCAO group (p < 0.01), while YHJG treatment significantly reduced their expression (p < 0.01). PPAR-γ expression significantly increased in the YHJG-H group (p < 0.01). Conclusions: The expression of IL-1β, TNF-α, IL-6, AKT1, and pERK1/2/ERK1/2 significantly increased in the MCAO group, while YHJG treatment significantly reduced their expression. PPAR-γ expression significantly increased in the YHJG-H group. YHJGs could treat IS through diverse ingredients, targets, and pathways by inhibiting inflammatory indices and AKT1 expression, and reducing ERK1/2 phosphorylation. Full article

Lithology identification is a critical technology for geological resource exploration and engineering safety assessment. However, traditional methods suffer from insufficient feature representation and low classification accuracy due to challenges such as weathering, vegetation cover, and spectral overlap in complex sedimentary rock regions. This study proposes a hierarchical multi-feature random forest algorithm based on Feature-Preserved Compressive Sampling (FPCS). Using 3D laser point cloud data from the Manas River outcrop in the southern margin of the Junggar Basin as the test area, we integrate graph signal processing and multi-scale feature fusion to construct a high-precision lithology identification model. The FPCS method establishes a geologically adaptive graph model constrained by geodesic distance and gradient-sensitive weighting, employing a three-tier graph filter bank (low-pass, band-pass, and high-pass) to extract macroscopic morphology, interface gradients, and microscopic fracture features of rock layers. A dynamic gated fusion mechanism optimizes multi-level feature weights, significantly improving identification accuracy in lithological transition zones. Experimental results on five million test samples demonstrate an overall accuracy (OA) of 95.6% and a mean accuracy (mAcc) of 94.3%, representing improvements of 36.1% and 20.5%, respectively, over the PointNet model. These findings confirm the robust engineering applicability of the FPCS-based hierarchical multi-feature approach for point cloud lithology identification. Full article

Sign language recognition technology serves as a crucial bridge, fostering meaningful connections between deaf individuals and hearing individuals. This technological innovation plays a substantial role in promoting social inclusivity. Conventional sign language recognition methodologies that rely on static images are inadequate for capturing the dynamic characteristics and temporal information inherent in sign language. This limitation restricts their practical applicability in real-world scenarios. The proposed framework, called SSTA-ResT, integrates ResNet, soft spatiotemporal attention, and Transformer encoders to achieve this objective. The framework utilizes ResNet to extract robust spatial feature representations, employs the lightweight SSTA module for dual-path complementary representation enhancement to strengthen spatiotemporal associations, and leverages the Transformer encoder to capture long-range temporal dependencies. Experimental results on the LSA64 Argentine Sign Language (ASL) dataset demonstrate that the proposed method achieves an accuracy of 96.25%, a precision of 97.18%, and an F1 score of 0.9671. These results surpass the performance of existing methods across all metrics while maintaining a relatively low model parameter count of 11.66 M. This demonstrates the framework’s effectiveness and practicality for sign language video recognition tasks. Full article

Additive manufacturing of metal components is predominantly based on fusion-welding processes involving melting and solidification. However, processing high-strength aluminium alloys presents challenges, including reduced mechanical properties and increased susceptibility to hot cracking. To address these issues, alternative solid-state processing methods for aluminium are being explored worldwide. One such method is wire-based friction-stir additive manufacturing, which builds on the principles of friction-stir welding. This study focused on assessing a range of pin tool designs to promote improved mixing between the filler material and substrate. The best results were achieved using a two-stirring-probe configuration, which was then employed to fabricate a multilayer wall made of EN AW-6063 aluminium alloy. The resulting structure showed significant grain refinement, with the deposited layers having an average grain size approximately four times smaller than that of the substrate, indicating dynamic recrystallisation. Tensile testing of the intermediate layer revealed a strength of 147 MPa and 10% elongation, corresponding to 77% of the filler wire strength. These findings highlight the potential of the W-FSAM process for producing near-net-shape, high-quality lightweight metal components with refined microstructures and reliable mechanical performance. Full article

Coniferous forests in Canada play a vital role in carbon sequestration, wildlife conservation, climate change mitigation, and long-term sustainability. Traditional methods for classifying and segmenting coniferous trees have primarily relied on the direct use of spectral or LiDAR-based data. In 2024, we introduced a novel data representation method, pseudo tree crown (PTC), which provides a pseudo-3D pixel-value view that enhances the informational richness of images and significantly improves classification performance. While our original implementation was successfully tested on urban and deciduous trees, this study extends the application of PTC to Canadian conifer species, including jack pine, Douglas fir, spruce, and aspen. We address key challenges such as snow-covered backgrounds and evaluate the impact of training dataset size on classification results. Classification was performed using Random Forest, PyTorch (ResNet50), and YOLO versions v10, v11, and v12. The results demonstrate that PTC can substantially improve individual tree classification accuracy by up to 13%, reaching the high 90% range. Full article