Search Results (5,242)

Cardiovascular diseases remain a major global health burden, making the development of accurate and interpretable prediction models important for clinical decision support. In this study, the metaheuristic Ivy Algorithm was employed to perform two-stage hyperparameter optimization for five machine learning classifiers, namely ID3, SVM, RF, XGBoost, and LightGBM. The proposed framework was evaluated on the publicly available Cleveland and Statlog heart disease datasets using outer stratified 10-fold cross-validation. The results showed that IVYA-based optimization improved the predictive performance of all five classifiers to varying degrees. Among them, IVYA-LightGBM achieved the best overall performance, with mean AUC, Accuracy, Precision, Recall, and F1-score values of 0.945, 0.907, 0.931, 0.864, and 0.893, respectively. Paired Wilcoxon signed-rank tests based on the fold-wise results indicated that the improvements in AUC were statistically significant in most model–dataset comparisons. In addition, under consistent experimental settings, IVYA was compared with five widely used metaheuristic optimization algorithms and achieved the highest AUC, Recall, and F1-score, while requiring the shortest average runtime. To enhance model interpretability, SHAP analysis was further incorporated to quantify the contributions of different clinical features to the model predictions and improve the transparency of the prediction process. Overall, IVYA-LightGBM achieved a favorable balance among predictive performance, computational efficiency, and interpretability. Nevertheless, further validation on larger and more diverse clinical datasets is required before practical clinical application. Full article

The escalating global burden of neurodegenerative and neuropsychiatric disorders is deeply intertwined with cumulative environmental toxicant exposure. To accurately capture the prodromal and subclinical impacts of this multi-chemical “pollutome”, research is shifting from symptom-based assessments toward the use of objective, quantifiable biological indicators. Guided by the PRISMA-ScR framework, this scoping review systematically analyzed literature across PubMed, Scopus, Web of Science, and Embase databases without temporal restrictions. The final selection included 52 human observational studies that evaluated the relationship between environmental stressors and objective neurobiological markers across diverse geographical populations and life stages. The synthesized evidence reveals that varied environmental insults—including ambient air pollution (e.g., fine particulate matter, PM_2.5), heavy metals, agrochemicals, and persistent organic contaminants—frequently correlate with measurable alterations in fluid biomarkers. These exposures are primarily associated with variations in markers of axonal damage (neurofilament light chain), astrocytic reactivity (glial fibrillary acidic protein), potential microglial dysfunction (soluble triggering receptor expressed on myeloid cells 2), and cytostructural alterations (Tau proteins and amyloid-beta). The literature highlights distinct windows of vulnerability, spanning from early-life epigenetic modifications (DNA methylation) to adult neurovascular injury. Mechanistically, despite their chemical heterogeneity, the available evidence suggests that these pollutants may converge on shared pathophysiological pathways defined by blood–brain barrier disruption and chronic, self-perpetuating neuroinflammation. Understanding environmental neurotoxicity may benefit from moving beyond traditional single-pollutant approaches toward a broader exposome framework. Future epidemiological research should integrate high-dimensional human biomonitoring with artificial intelligence and machine learning architectures. This computational integration would be essential to decode non-linear multi-pollutant interactions and accelerate the deployment of targeted, early-stage public health interventions. Full article

Background: Whole-blood transcriptomic profiling can capture systemic immune-response alterations associated with COVID-19 and may support host-response-based classification. However, evidence regarding the discriminatory value of targeted immune-gene panels remains limited, and in small, high-dimensional datasets, the timing of feature selection may substantially affect model performance and interpretation. Aim: This study aimed to evaluate whether NanoString Human Immunology Panel profiles could distinguish COVID-19 from healthy-control measurements and to compare full-dataset feature selection (FDFS) with leave-one-out cross-validation (LOOCV)-embedded feature selection (LEFS). Methods: Publicly available E-MTAB-8871 data comprising 579 genes and 32 whole-blood transcriptomic profiles were analyzed. The dataset included 22 longitudinal COVID-19 measurements obtained from three participants and 10 measurements obtained from 10 healthy controls. Elastic Net regularization was used for feature selection. Random Forest, XGBoost, and LightGBM classifiers were evaluated using sample-level LOOCV. Model performance was assessed using threshold-dependent, discrimination, and probability-based metrics. A separate exploratory LightGBM model was analyzed using SHapley Additive exPlanations (SHAP) to characterize feature contributions. Results: FDFS identified a fixed 40-gene set, whereas LEFS selected a mean of 42 genes per fold (range: 40–47). LightGBM correctly classified all 32 measurement-level profiles (derived from 13 unique participants: 10 healthy controls and three longitudinally sampled COVID-19 participants) in both frameworks, achieving area under the receiver operating characteristic curve (ROC-AUC) and area under the precision–recall curve (PR-AUC) values of 1.000 and Brier scores of 0.005 and 0.006 in the FDFS and LEFS frameworks, respectively. Random Forest achieved accuracies of 0.969 and 1.000, whereas XGBoost achieved an accuracy of 0.969 in both frameworks. SHAP analyses consistently identified AICDA as the dominant contributor to model predictions, followed by ARHGDIB. Conclusions: This exploratory analysis showed that targeted NanoString immune-response profiles contained a compact transcriptomic signal capable of distinguishing COVID-19 from healthy-control measurements within the analyzed dataset. These findings provide proof-of-concept evidence of internal measurement-level discrimination. However, because the COVID-19 profiles consisted of repeated measurements from only three participants, sample-level LOOCV did not constitute independent participant-level validation. External validation in larger cohorts comprising independently sampled participants is required. Given that the COVID-19 arm comprised only three independent participants, these biological findings should be regarded as hypothesis-generating and require validation in substantially larger independent cohorts. Full article

Accurate petrographic identification from thin-section images is important for geoscientific analysis but remains strongly dependent on expert interpretation. In this study, we organized 2634 polarized-light images from 324 distinct thin sections representing 108 lithologies and paired them with Chinese petrographic descriptions. Four conventional vision models were trained as closed-set lithology–classification baselines, and five CN_CLIP variants were fine-tuned for image–text alignment. To prevent overlap of images from the same thin section across outer folds, thin-section-wise three-fold evaluation was performed after excluding eight lithologies represented by fewer than three distinct thin sections; this protocol therefore comprised 2530 images from 311 thin sections and 100 eligible lithologies. At the fixed 20-epoch endpoint, CN_CLIP-ViT-L/14@336px and CN_CLIP-ViT-H/14 showed bidirectional image–text matching accuracies of 81.39 ± 1.97% and 81.31 ± 0.61%, respectively, whereas the conventional classifiers achieved Top-1 accuracies of 50.37%–58.20% under their separate closed-set task. These metrics correspond to different task formulations and are not directly comparable. Under the complementary image-level 8:1:1 protocol, CN_CLIP-ViT-H/14 achieved image-to-text and text-to-image matching accuracies of 93.75% and 93.38%, respectively, on the hold-out test subset. In an inference-stage ablation using the same checkpoint, hold-out test batches, and evaluation protocol, removing only the explicit lithology name reduced these accuracies to 82.72% and 77.21%, indicating that the rock-name token is informative but not the sole source of cross-modal discrimination. Without task-specific retraining, the same CN_CLIP framework also supported constrained lithology, major-mineral, and polarization-mode matching, with best task-specific accuracies of 90.57%, 66.51%, and 86.42%, respectively. Overall, the results support vision-language learning as a flexible framework for expert-assisted petrographic image–text matching on held-out thin sections within the available dataset, while emphasizing source-aware partitioning and cautious interpretation of task-specific accuracies. Full article

Underwater object detection faces significant challenges including uneven lighting, low contrast, and scattering-induced distortions. Existing underwater datasets are limited in scale, class diversity, and annotation consistency, which hinders robust model development. This work addresses these limitations by creating the UWOD dataset through the integration of seven publicly available underwater datasets, comprising over 107 K images with approximately 374 K annotations across 39 classes. We employ a semi-automatic annotation pipeline that combines manual labeling with iterative model-in-the-loop training to ensure high-quality ground truth, As a final verification step, all auto-generated labels were manually inspected and corrected as needed. We benchmark state-of-the-art object detectors—including YOLOv8, YOLOv7, YOLOv5, FCOS, EfficientDet, YOLOX, RT-DETR, SSD, and Faster R-CNN—establishing comprehensive performance baselines; YOLOv8 and YOLOv7 achieve the best accuracy–efficiency trade-off. Our transfer learning analysis shows that domain-specific pretraining substantially often outperforms pretraining on general-purpose datasets, yielding up to more than 50% improvement in low-data regimes versus training from scratch, with markedly lower seed-to-seed variance than scratch training. Sequential pretraining on COCO followed by UWOD achieves the strongest results on our most challenging dataset. Due to upstream licensing constraints, we release trained model weights and an automated annotation pipeline that encapsulate the learned underwater-domain knowledge, enabling immediate application to new imagery while respecting intellectual property. Full article

Agrivoltaic systems (AVSs) present a promising solution to land-use conflicts between solar photovoltaics and agriculture. Nevertheless, their electricity-side cost performance is strongly influenced by site-specific conditions and agricultural constraints, and a comprehensive nationwide geospatial assessment of optimal designs under agronomic light constraints remains lacking. To address this gap, we developed a GIS-based techno-economic framework for Japan to identify site-specific optimal designs by minimizing the Levelized Cost of Electricity (LCOE) while satisfying crop light requirements represented by target Daily Light Integral (DLI) values. The analysis indicates a national average LCOE of 0.1021 EUR kWh⁻¹ for optimized AVSs, with lower electricity-side costs in central to southwestern regions characterized by high solar irradiance. The optimal design, particularly the projected ground coverage ratio (average: 30.38%), varied considerably across the country to balance power generation and crop light availability. Scenario and sensitivity analyses revealed that LCOE is highly sensitive to crop light requirements, whereas PV system efficiency and capital expenditure are the most influential cost drivers. This study provides the first high-resolution, nation-scale LCOE map for AVS in Japan derived through grid-level design optimization under explicit DLI-based constraints, offering a quantitative screening-level foundation for region-specific deployment planning. Full article

Background/Objectives: Neopyropia yezoensis is an economically important intertidal red alga that frequently experiences fluctuations in temperature, light intensity, and water availability. Formins are key regulators of actin nucleation and cytoskeletal dynamics, but their functions and stress-responsive roles in red algae remain poorly understood. This study aimed to systematically identify and characterize the Formin gene family in N. yezoensis and investigate their expression responses to different environmental stresses. Methods: Formin family members were identified from the N. yezoensis genome using HMMER and BLAST (2.17.0) searches based on the conserved FH2 domain, followed by domain validation. Gene structures, conserved motifs, physicochemical properties, predicted subcellular localization, protein structures, chromosomal distribution, phylogenetic relationships, and cis-acting elements in the upstream regions were analyzed. The expression patterns of the identified Formin genes were further examined by qRT-PCR under different temperature (4, 10, and 24 °C), light intensity (20, 60, and 100 μmol photons m⁻² s⁻¹), and desiccation/rehydration conditions. Results: Three Formin genes, designated NpyFormin01–03, were identified in N. yezoensis. All three encoded proteins contained the conserved FH2 domain but differed in motif composition, domain architecture, predicted subcellular localization, and structural features. NpyFormin01 contained additional PTEN_C2 and PTP_DSP_cys domains, whereas NpyFormin02 and NpyFormin03 contained only the FH2 domain. Phylogenetic analysis showed that the N. yezoensis Formins clustered with Formins from other red algae. Promoter analysis identified multiple predicted cis-acting elements associated with light, temperature, environmental, and phytohormone responses. Expression analysis revealed distinct responses among the three genes under the tested environmental conditions. Notably, NpyFormin01 was significantly upregulated under high-temperature treatment (24 °C), whereas NpyFormin02 and NpyFormin03 showed no statistically significant expression changes under the tested conditions. Conclusions: This study provides a systematic characterization of the Formin gene family in N. yezoensis. The differences in protein architecture, structural features, promoter cis-acting elements, and environmental-responsive expression patterns suggest potential functional divergence among NpyFormins. In particular, NpyFormin01 represents a potential heat-responsive candidate gene and may contribute to cytoskeletal regulation during environmental stress adaptation in N. yezoensis. These findings provide a basis for further investigation of the molecular functions of Formins in red algae. Full article

The tympanic membrane (TM) is a thin, semi-transparent structure whose morphology and optical appearance provide important diagnostic cues. In the early stages of middle-ear pathology, subtle shape and compliance alterations may precede overt clinical signs, making them valuable early indicators of disease. Such structural changes are difficult to assess reliably using conventional (micro-)otoscopy, which lacks quantitative depth information and is operator-dependent. Data-driven monocular image analysis could enable quantitative assessment of TM geometry and compliance, but the limited availability of annotated three-dimensional datasets constrains the development of these methods. At the same time, realistic simulations of TM appearance remain challenging due to its complex reflectance and transmission behavior. The present study focuses on physiologically healthy tympanic membranes, which provide the baseline anatomical and optical model required before subtle pathological changes can be investigated. This work introduces a parametric physically grounded rendering model of the structures visible during otoscopy: the tympanic membrane, ear canal, and malleus–incus complex. Implemented in the open-source software Blender™ using procedural geometry nodes and physically motivated shaders, the framework generates anatomically plausible three-dimensional geometries via statistical parameter sampling and controlled mesh deformation. Optical appearance is simulated using a computationally efficient layered shading model based on literature-derived tissue reflectance, transmission, and scattering properties. A camera–projector setup models both conventional white-light otoscopy and structured-light imaging, enabling the generation of paired intensity images and corresponding depth maps. The proposed framework establishes a physically grounded representation of the human ear and enables a controllable, extensible modeling pipeline for virtual training, biomechanical finite element analysis, and synthetic data generation for supervised learning. Full article

Light Detection and Ranging (LiDAR)-based forest inventory increasingly relies on diverse platforms, ranging from proximal systems including BackPack, All-Terrain Vehicle (ATV), and terrestrial laser scanning (TLS) to near-proximal systems such as uncrewed aerial vehicles (UAVs). However, differences in point density, viewing geometry, and occlusions among these acquisition systems pose challenges for processing heterogeneous LiDAR datasets using a common workflow. Traditional geometric approaches often rely on parameter tuning. On the other hand, deep learning (DL) approaches can be constrained by domain shift when applied to different sensors or forest environments. This study proposes a forest inventory pipeline for individual tree segmentation and the derivation of key forest biometrics including tree location and diameter at breast height (DBH) across heterogeneous LiDAR datasets. The pipeline uses a confidence-guided, multi-stage quality control framework that evaluates agreement between complementary tree location estimates to reduce common segmentation errors. In addition, a semi-automated procedure is developed to generate reference data for datasets lacking field measurements. The proposed workflow was evaluated using eight diverse datasets representing different platforms, sensors, acquisition patterns, and forest environments and was compared with 3DFIN, TreeLearn, and ForestFormer3D. Field reference measurements were available for a natural forest site, while the remaining datasets were evaluated using semi-automatically generated and manually refined reference data. The proposed tree detection pipeline achieved Precision ranging from 86.44% to 100%, Recall from 74.17% to 100%, and F1-scores from 81.82% to 100% across the evaluated datasets. For the Martell–BackPack dataset with independent field reference measurements, Precision, Recall, and F1-score were 97.55%, 96.95%, and 97.25%, respectively. For correctly detected trees by the proposed approach in the natural forest dataset with field measurements, DBH estimates achieved an RMSE of 2.5 cm with the total basal area underestimated by 1.88%, compared with DBH RMSE and reduction in basal area of 4.0 cm and 3.67%, respectively, for 3DFIN. Although the proposed pipeline did not achieve the highest performance in every test case, it maintained strong and generally consistent tree detection performance for the evaluated datasets. The main limitation of the proposed pipeline is its dependence on sufficient lower-stem visibility, which reduced tree detection accuracy in sparsely sampled areas. The proposed framework provides a practical workflow for LiDAR-based individual tree segmentation and DBH estimation using a fixed parameter configuration for all datasets captured by a given acquisition system. Full article

Accurate iris segmentation remains a fundamental challenge in iris biometric recognition and medical image analysis, particularly in challenging scenarios such as non-cooperative acquisition conditions involving variable illumination, partial occlusions, degraded image quality, and diverse unconstrained environments. Prevailing segmentation algorithms exhibit limited robustness when confronted with such challenges, and the disparity between near-infrared (NIR) and visible-light imaging modalities further compounds the complexity of achieving a robust segmentation outcome. To address these challenges, this paper introduces SwinIrisNet, a hybrid deep learning architecture that integrates Swin Transformer and convolutional neural network (CNN) branches within a U-Net framework for robust iris segmentation. The Swin Transformer branch leverages hierarchical window-based self-attention to capture global contextual dependencies, whereas the CNN branch extracts fine-grained local features essential for precise boundary delineation. A memory-efficient cross-attention fusion module combines these complementary feature representations, further enhanced by a Convolutional Block Attention Module (CBAM), Atrous Spatial Pyramid Pooling (ASPP), and attention-gated skip connections for multi-scale context aggregation. An extensive evaluation is conducted across four publicly available benchmark datasets, including UBIRIS.v2, IITD, CASIA-Thousand, and MMU.v1, encompassing both visible-light and NIR imaging environments. The proposed architecture yields F1 values of 0.9612–0.9672, Dice coefficients of 0.9489–0.9519, mIoU values of 0.9266–0.9450, precision values of 0.9565–0.9633, recall values of 0.9600–0.9672, and classification accuracies of 99.51–99.53%, with NICE1 error rates of 0.57–0.60% and NICE2 values of 1.82–2.24%, confirming pixel-level segmentation quality. Cross-database generalization experiments further demonstrate that SwinIrisNet learns transferable iris representations and generalizes effectively across heterogeneous imaging sources, with the strongest transfer occurring in the NIR-to-visible direction. A comparative analysis against existing algorithms demonstrates that the proposed architecture attains substantial performance improvements over several existing segmentation networks when evaluated on identical benchmark databases, surpassing them across the majority of qualitative and quantitative metrics while maintaining a marginally lower memory footprint. Full article

Phosphate (P_i) deficiency induces responses that enhance P_i uptake and utilization. Light may influence these responses through photosynthetic carbon supply and signaling. We examined how light affects the early root transcriptional response to P_i deficiency in hydroponically grown soybean. Plants were exposed to phosphate-sufficient (+P) or phosphate-deficient (−P) conditions for 30 h under a light/dark cycle or continuous darkness. Root transcriptomes were analyzed using Oxford Nanopore cDNA sequencing. Principal component analysis (PCA) showed that transcriptomes separated mainly by light, with weaker separation by P_i status. Under light, P_i deficiency induced a broad response involving P_i transport, signaling, transcriptional regulation, lipid remodeling, and metabolism. In darkness, relatively few genes were upregulated under P_i deficiency. This limited response was not caused by a loss of transcriptional responsiveness, because thousands of genes were upregulated by darkness itself. Comparison of fold changes revealed that darkness attenuated, rather than reversed, the −P response. Interaction-ranked gene set enrichment analysis (GSEA) showed that the weaker response in darkness extended across signaling, metabolic, and transport pathways. Light-signaling genes responded strongly to darkness but showed little phosphate-dependent regulation. Root sucrose levels were lower in darkness, consistent with reduced carbon availability as a possible contributor to the attenuated P_i-deficiency response. Full article

Human induced pluripotent stem cell (iPSC)-derived neural organoids have emerged as valuable models for investigating human neurodevelopment and neurological disorders. However, their complex three-dimensional architecture poses significant challenges for conventional histological approaches, which rely on physical sectioning and inevitably disrupt spatial relationships within the tissue. Volumetric imaging of intact organoids typically requires light-sheet fluorescence microscopy, a technology not widely accessible to standard cell biology laboratories. Here, we show that solvent-based tissue clearing, using the iDISCO+ and Visikol^® HISTO protocols, combined with conventional laser-scanning and spinning-disk confocal microscopy, platforms already available in most imaging facilities, is sufficient to resolve neuroepithelial rosette-like structures, neuronal networks, and astroglial components within intact human cortical and retinal organoids, while preserving immunofluorescent labeling and tissue architecture. The workflow was also compatible with commonly used immunofluorescence markers. Although light-sheet fluorescence microscopy remains advantageous for large-scale whole-sample imaging, our results show that cleared human organoids within the size range analyzed here can be effectively visualized using accessible confocal systems. This study provides a practical strategy for three-dimensional imaging of intact human neural organoids, facilitating spatial analysis of developmental organization and disease-relevant phenotypes in laboratories without dedicated light-sheet microscopy infrastructure. Full article

Power transformers are among the most critical components of electric power transmission and distribution systems, and unexpected failures can lead to substantial economic losses and prolonged power outages. Dissolved Gas Analysis (DGA) is the most widely used diagnostic technique for transformer fault diagnosis and involves interpreting gases dissolved in insulating oil. However, conventional interpretation methods, such as the Rogers ratio, the Doernenburg ratio, the IEC 60599 ratio method, and the Duval Triangle, rely heavily on expert knowledge, may produce inconsistent diagnoses for the same oil sample, and may fail to provide a diagnosis in certain cases. In this study, 12 classification models were evaluated using the publicly available Power Transformers Fault Detection and Diagnosis (FDD) and Remaining Useful Life (RUL) dataset and a unified evaluation protocol. Model performance was assessed using Accuracy, Balanced Accuracy, and Macro-F1 score, while model interpretability was investigated through Shapley Additive Explanations (SHAP) analysis. The results showed that ensemble tree-based methods achieved the best overall performance. LightGBM and Random Forest both attained an Accuracy of 0.969 and a Macro-F1 score of 0.924, while LightGBM further achieved the highest Balanced Accuracy of 0.940. XGBoost exhibited the most stable performance under cross-validation. SHAP analysis revealed that engineered relative concentration features, particularly the CO/H₂ ratio and the combined gas ratio, were among the most influential features for fault classification. These findings demonstrate that, for datasets of this scale, ensemble tree-based models combined with well-designed features provide strong and interpretable performance for imbalanced DGA-based fault diagnosis, highlighting the effectiveness of feature-based ensemble learning for small- to medium-sized datasets. Full article

Airports handle the final stage of the arriving passenger’s journey at the baggage claim hall, one of the least instrumented terminal operations. At most airports, arriving bags are discharged onto a shared recirculating carousel, a passive, first-come process that lengthens passenger waiting, congests the claim hall, and gives operators no per-party measure of service. This paper formalizes inbound claim as a measurable, controllable terminal service, termed Baggage-Claim-as-a-Service (BCaaS), in which the passenger party rather than the individual bag is the unit of service and the complete bag set is delivered against an observable delivery-time target. The paradigm is evaluated through a per-party, cluster-based reclaim architecture that consolidates each party’s bags, routes the cluster through a multi-level sortation network, releases the complete set at an authenticated collection point, and records per-party telemetry for service monitoring. Using a discrete-event simulation calibrated to King Khalid International Airport Terminal 5 and benchmarked against a conventional carousel across 27 scenarios, per-party reclaim reduces availability-based passenger, cluster, and flight delivery times by 38%, 51%, and 27% on average, largest under light load. The results show that per-party reclaim makes inbound terminal logistics observable and supports operator service-level monitoring, more sustainable urban mobility, and future field validation. Full article

With the increase in global fishing activities, the transparency of fishery data has drawn increasing attention. Data transparency enables stakeholders to play a greater role in ensuring that fisheries are legal, ethical, and sustainable. Improving the transparency of fishing vessel data can effectively support vessel monitoring and enhance fishery safety. This paper presents a systematic review of satellite remote sensing modalities and datasets currently available for fishing vessel identification and monitoring. Conducted in accordance with the PRISMA 2020 guidelines, this study employs a dual-track search strategy to retrieve, screen, and synthesize academic literature and public datasets from mainstream databases, including the Web of Science Core Collection, IEEE Xplore, and CNKI. First, the existing remote sensing modalities were classified into three major categories based on their imaging principles: synthetic aperture radar (SAR), optical remote sensing, and nighttime light (NTL) remote sensing. In addition, the mainstream satellite data sources and their corresponding parameters were summarized for each category. Second, an in-depth comparative analysis of these remote sensing modalities is conducted from core dimensions such as target detection sensitivity, robustness under complex environments and meteorological conditions, and spatiotemporal resolution. This reveals the performance limitations and significant complementarity of different sensor data in fishing vessel detection. Finally, mainstream remote sensing datasets for fishing vessels (such as xView3-SAR, xView, and VBD) are summarized and evaluated, pointing out the gaps in certain types of datasets. In conclusion, this paper suggests that building a “full spatiotemporal and multi-scale” observation framework based on multi-source heterogeneous data fusion is an important trend for the future development of fishing vessel detection using remote sensing, aiming to provide a reference for relevant researchers. Full article