Search Results (5,828)

Accurate color classification plays a critical role across diverse fields, from textile manufacturing and environmental monitoring to biomedical diagnostics. This study introduces a machine-learning-driven approach to spectral color sensing using SENSIPATCH, a compact, wearable sensor system; while SENSIPATCH integrates multiple sensing modalities, including bioimpedance, electrochemical, thermal, humidity, and vibrational sensors, this work specifically utilizes its spectrometer module, which comprises multi-wavelength LEDs and photodiodes. Targeting the classification of 100 standardized PANTONE colors, the proposed framework is evaluated under controlled lighting conditions to ensure repeatable spectral acquisition. The experimental design includes both firm and loose contact scenarios to emulate variability in wearable placement. A structured data-preprocessing pipeline involving baseline correction, bootstrapping, and Z-score normalization was employed to enhance signal quality and improve model generalization. Five machine learning models were evaluated: Random Forest, SVM, MLP, CNN, and LSTM. The MLP demonstrated the strongest classification performance. Notably, the MLP achieved consistent accuracy across both contact conditions, indicating robustness against sensor placement variations. These results highlight the feasibility of compact LED-based wearable spectroscopy for reliable color classification under controlled measurement conditions, providing a baseline for future extensions to more diverse lighting conditions. Full article

Background/Objectives: In addition to the clinical gold standard, tympanometry, several alternatives for middle ear diagnostics have evolved over the past decades. With the so-called pressureless acoustic impedance test, the Neuranix Medwave, another device, came into play. Methods: Using a retrospective, anonymous study design, descriptive data were reported, and the correlation between Medwave’s results and tympanometry types was evaluated. Also, the correlation between the patients’ age and the Medwave resulting parameters was evaluated. We were able to show changes in the measurement results over time in the case of paracentesis and tube insertion. Results: The analyzed data show that it is possible to differentiate between tympanometry result type A and type B using the Medwave resulting parameter resonance frequency (“fR”), but not when using peak admittance (“P”). Between all other types, it was not possible to differentiate using the Medwave resulting parameters, nor fR nor P. Due to the low statistical power, this may be due to a type II error. Regarding age, a correlation was found only for the tympanometry result type A. The case over time showed a clear difference in the affected ear between the time before and after the ear surgeries, as well as the contralateral healthy ear. Conclusions: While this study indicates the potential use of the PLAI technology, especially as a tool in situations where traditional tympanometry is not feasible, the results need to be interpreted with caution. Further validation with larger and more balanced groups of participants is necessary to confirm these initial findings and to more clearly define the clinical utility of this technology. Full article

Circulating long non-coding RNAs (lncRNAs) have emerged as promising non-invasive biomarkers for colorectal cancer (CRC) detection; however, data in Vietnamese populations remain limited. In this study, a total of 218 participants (106 CRC, 80 adenomas, and 32 healthy controls) were included. Relative expression levels and diagnostic performance of three circulating lncRNAs—CCAT1, HOTAIR, and NEAT1—were quantified using RT-qPCR and analyzed by the 2^−ΔΔCt method. Receiver operating characteristic (ROC) curve analysis was performed to assess the diagnostic accuracy of individual lncRNAs and their combinations. CCAT1, HOTAIR, and NEAT1 were significantly upregulated in CRC patients compared with adenoma patients and healthy controls (all p < 0.001). Expression levels were higher in advanced-stage (TNM III–IV) CRC than in early-stage disease. Among individual markers, HOTAIR demonstrated the highest diagnostic accuracy (AUC = 0.918), followed by CCAT1 (AUC = 0.908) and NEAT1 (AUC = 0.890). Combined biomarker models showed improved performance, with the CCAT1 + HOTAIR combination achieving the highest AUC (0.944). Overall, circulating CCAT1, HOTAIR, and NEAT1 demonstrated favorable diagnostic performance in a Vietnamese population and outperformed conventional markers (CEA and CA19-9). These findings support the potential utility of multi-lncRNA panels as non-invasive biomarkers for CRC detection, warranting further validation in larger, independent cohorts. Full article

Accurate distinction between non-muscle-invasive bladder cancer (NMIBC) and muscle-invasive bladder cancer (MIBC) remains the key local staging problem in bladder cancer because treatment intensity, timing of radical therapy, and suitability for bladder-preserving strategies all depend on it. Multiparametric magnetic resonance imaging (mpMRI) and the Vesical Imaging-Reporting and Data System (VI-RADS) now provide a standardized imaging framework for local staging and increasingly support MRI-first clinical pathways. Artificial intelligence (AI) has emerged as an additional decision-support layer, but the evidence base remains methodologically uneven. In this structured narrative review, we synthesized peer-reviewed literature from January 2020 to March 2026, while retaining foundational VI-RADS studies from 2018 to 2019, and prioritized guideline documents, meta-analyses, prospective cohorts, multicenter and externally validated AI studies, response-assessment studies, and papers addressing implementation and reporting quality. Current evidence shows that radiomics and deep learning models can achieve high discrimination for MIBC detection on MRI, and that the most plausible incremental value of AI lies in equivocal VI-RADS lesions, reader support outside high-volume expert settings, and multimodal risk stratification. However, most studies remain retrospective, highly selected, segmentation-dependent, and vulnerable to reference-standard bias, domain shift, and poor calibration. This review therefore emphasizes several translational issues that are often underreported: lesion-level versus patient-level inference, the distortive effect of TURBT-based labels, the need to evaluate false-negative consequences in VI-RADS 3 tumors, and the distinction between diagnostic support and broader pathway redesign. We also discuss response assessment, nacVI-RADS, segmentation automation, multicenter and federated infrastructure, workflow ownership, and the limits of imaging-only models in a biologically heterogeneous disease. The most credible near-term role of AI is not autonomous diagnosis, but augmentation of standardized mpMRI and VI-RADS within multidisciplinary care. Future progress will depend on prospective utility studies, site-held-out validation, transparent reporting, and the integration of imaging with molecular and cellular heterogeneity through radiogenomic and multi-omics approaches. Full article

The primary aim of this study is to validate the utility of Social Media Data (SMD) as a scientifically grounded tool for quantifying the spatial mismatch between objective ecological supply and subjective social demand. Assessing the spatial coupling and mismatch between Habitat Quality (HQ)—representing objective ecological supply—and Ecological Perception (EP)—representing subjective social demand—is essential for developing targeted urban management and development strategies. Focusing on the core urban area of Hangzhou, this study quantified ecological supply using the InVEST HQ model. To reflect social demand, 4958 geolocated Weibo posts were processed using contextual sentiment analysis. A Coupling Coordination Degree model served as a diagnostic tool to evaluate the synergy between these two dimensions. Additionally, a Geodetector model was employed to investigate the factors driving spatial differentiation in this coupling. The findings indicate that: (1) The regional average HQ is 0.56, reflecting a moderate overall level of degradation, while EP shows a preference for natural environments and exhibits a distinct “strip-like” spatial distribution. (2) The overall CCD value is 0.384; high-coupling areas are primarily concentrated in regions with superior natural conditions and dense vegetation, whereas low-coupling areas correspond to zones with intensive urban functions. (3) Driving factor analysis reveals that land-use type exerts the most significant influence on the overall degree of coupling. This study demonstrates that the HQ-EP coupling framework provides a reliable spatial diagnostic tool for urban planners to identify socio-ecological vulnerabilities. The results suggest that an appropriate integration of natural elements enhances coupling outcomes, with the highest synergy observed in environments characterized by high HQ and minimal anthropogenic disturbance. Full article

The first influenza A(H5N1) human case associated with the A(H5N1) dairy cattle outbreak in the United States was identified in April 2024. The U.S. CDC response to this outbreak was activated days later and remained active until July 2025. During this time, 70 human cases of influenza A(H5N1) were detected with a range of epidemiological links to sources of exposure. Next-generation sequencing (NGS) of human samples was an effectual mechanism for tracking and analyzing the outbreak evolution throughout the response. Due to the specimens’ importance and their variable physical quality, an assortment of laboratory methods was utilized including influenza segment-specific amplification, enrichment capture, short-read, and long-read sequencing. Combining these methods allowed for high-quality genomic data production with rapid turnaround times—typically 2 days from sample receipt to public database submission. By leveraging replicate sequencing, enrichment capture, and sequencing of diagnostic amplicons, valuable genomic data could be produced directly from human clinical specimens that would have normally been considered too weak for routine virologic surveillance sequencing. The resulting assemblies were characterized and analyzed by CDC and shared with local and state public health authorities to facilitate case investigations and risk assessment. These data were further used for phylogenetic analyses of viruses from human cases to investigate likely animal-to-human transmission events and identify clusters within the outbreak that might indicate trends in the types of exposures. Through the adaptable laboratory workflow and the rapid release of viral genomic data, the public health risk mitigation strategies could be evaluated and adjusted in real time. Full article

In intensive care units, large-scale clinical time-series data are continuously accumulated through electronic medical records and bedside monitoring systems. However, direct utilization of such data for clinical decision-making remains challenging due to irregular sampling, pervasive missingness, unstructured diagnostic information, and incomplete ICD labeling. Automated ICD coding constitutes an extreme multi-class classification problem with thousands of long-tailed categories, while intervention prediction tasks, such as mechanical ventilation management, involve rare transition events and severe class imbalance. To address these challenges, we propose CAGE, a hierarchical Clinical Decision Support System framework that integrates diagnosis, time-series signals, and intervention prediction. The framework first infers admission-level diagnostic context using a partial-label Automated ICD Coding module that combines DCNv2 with an Adaptive CLPL loss, producing probability-weighted diagnostic embeddings. These embeddings are subsequently fused with ICU time-series tensors and processed by a multi-branch Temporal Convolutional Network equipped with an ICD-conditioned gating mechanism to predict future ventilation state transitions. The experimental results demonstrate that DCNv2 achieves consistent superiority across all hit@k and probability concentration metrics for ICD coding. For intervention prediction, the proposed method substantially outperforms existing baselines, achieving a Macro-AUC of 98.2, Macro-AUPRC of 77.4, and F1-score of 79.4. These findings indicate that reinjecting diagnostic context as a conditioning variable, together with imbalance-aware loss design, effectively enhances rare-event detection and improves the practical applicability of clinical decision support systems. Full article

Background: Population-level evidence on delivery of neuro-oncology care is essential for evaluating access, equity, and quality of treatment pathways. However, real-world data describing how patients with central nervous system (CNS) tumors, especially with glioblastoma, are managed across healthcare systems remain limited. This study aimed to characterize treatment pathways using linked registry and administrative data within a regional care network. Methods: All adult CNS tumors diagnosed between 2016 and 2020 were identified in the Veneto Cancer Registry. Tumor grading was derived using a validated text-mining algorithm, and surgical, radiotherapy, and systemic treatments were captured through linkage with regional healthcare utilization databases. Patterns of care were evaluated by tumor subtype, grade, and diagnostic pathway. Results: Among 1634 histologically confirmed tumors, glioblastoma represented the largest group. Surgical intervention was widely implemented, with high resection rates in glioblastoma and meningioma. Combined chemoradiotherapy constituted the primary adjuvant approach for glioblastoma and high-grade diffuse gliomas, whereas management of lower-grade tumors showed greater variability. Approximately one-third of patients received no oncologic therapy, primarily associated with older age or diagnostic uncertainty. Analysis of recurrent glioblastoma showed heterogeneous systemic treatment use, reflecting evolving therapeutic practice. Conclusions: Linking population-based registry and administrative data provides actionable insight into real-world delivery of neuro-oncology care, in particular for glioblastoma patients. This approach enables monitoring of treatment variability, identification of potential access gaps, and evaluation of system-level performance, supporting data-driven planning of multidisciplinary services and future quality improvement initiatives. Full article

Agrivoltaic systems co-locate photovoltaic (PV) arrays and crops, offering land-use efficiency and potential microclimate benefits, yet they introduce new challenges for computer-vision-based crop monitoring. PV structures produce strong, spatially varying shadows, specular reflections, and periodic occlusions that confound visual cues for diagnosing crop diseases and abiotic stresses. Meanwhile, agrivoltaic deployments are often distributed across farms and operators, making centralized data collection impractical due to privacy, ownership, and regulatory concerns. This paper proposes PrivAgriVolt, a novel privacy-preserving learning framework for agrivoltaic crop issue recognition that explicitly models PV-induced illumination and enables collaborative training without sharing raw images. The core algorithm integrates (i) a PV-geometry-conditioned shadow normalization module that fuses estimated array layout and sun-angle priors into a shadow-aware appearance canonization network, reducing illumination-induced domain shift across times and sites; (ii) a federated contrastive stress learner that aligns stress semantics across farms via prototype-based contrastive objectives while remaining robust to heterogeneous sensors and crop stages; and (iii) an adaptive privacy layer that combines secure aggregation with budget-aware gradient perturbation and client-level clipping to provide formal privacy guarantees while preserving fine-grained diagnostic performance. Extensive experiments on real agricultural vision benchmarks and agrivoltaic shadow variants demonstrate that PrivAgriVolt improves stress recognition and segmentation under PV shading while maintaining strong privacy–utility trade-offs. Full article

Background: The elderly are vulnerable to chronic diseases and altered lipid metabolism, leading to poor outcomes, including mortality. We investigated the association between Korean Medicine (KM) utilization, blood lipid levels, and health outcomes using the National Health Insurance Service Sample Cohort (NHIS-NSC) database. Methods: This retrospective cohort study included elderly participants who underwent health examinations (2009–2010). Participants were divided into KM and non-KM groups and matched 1:1 using propensity score matching (PSM) for age, sex, income, and comorbidities. Primary outcomes were mortality and disease diagnosis; secondary outcomes included medical spending and utilization. Results: After PSM, 13,044 subjects were analyzed. KM utilization was associated with a significantly lower risk of all-cause mortality (HR 0.93; 95% CI 0.87–1.00; p = 0.048). However, the hypolipidemia subgroup showed no significant differences in all-cause mortality and medical expenses compared to other lipid status subgroups. While the KM group showed a higher incidence of disease diagnosis (HR 1.09; 95% CI 1.04–1.14; p < 0.001), this may reflect increased healthcare engagement and proactive health-seeking behavior. Subgroup analysis revealed that statin users in the KM group had a significantly reduced mortality risk (HR 0.91; 95% CI 0.84–0.99; p = 0.022). Medical expenses and utilization were higher in the KM group. Being underweight or aged over 85 was associated with higher mortality. Conclusions: KM utilization is associated with reduced all-cause mortality after propensity score matching, particularly among statin users. Although KM users had a higher cumulative incidence of disease diagnosis, this potentially reflects increased diagnostic opportunities from prolonged survival. Hypolipidemia, underweight, and late-elderly status remain significant risk factors associated with frailty. KM may support improved survival in the elderly, warranting further prospective studies. Full article

To address the challenges in intuitive feature discrimination and precise quantitative evaluation of cable defects, this paper proposes a diagnostic methodology utilizing the Symmetrized Dot Pattern (SDP) transform and reflection coefficient spectra. The Dung Beetle Optimizer (DBO) is introduced to adaptively optimize the SDP transform parameters, employing the Structural Similarity Index Measure (SSIM) as a fitness function to maximize discriminability between deterioration states. Three quantitative features, including the number of effective pixels, the degree of red–blue aliasing, and radial dispersion, are extracted to characterize the physical degradation processes of signal energy accumulation, angular evolution, and path divergence. By incorporating a self-reference calibration mechanism for structural differences, features are fused into a Comprehensive Deterioration Index (CDI). Experimental results on coaxial cables simulating shielding damage and thermal aging demonstrate that SDP images reveal continuous evolution patterns corresponding to defect severity. A regression model based on these patterns effectively characterizes deterioration trends. Compared to complex models, this study achieves intuitive fault identification and preliminary quantitative description of degradation trends through image feature fusion. Although the current sample size is limited, the results validate the feasibility of this method in evaluating cable deterioration severity, offering an efficient new data-processing perspective for cable condition monitoring. Full article

Background/Objectives: The manual analysis of sperm morphology, crucial for male infertility diagnosis, is subjective and time-consuming. Automated methods using deep learning, offer a promising alternative; however, standard deep models are prone to overfitting when applied to small, heavily unbalanced clinical datasets, limiting their generalization capability. This study proposes a knowledge distillation approach that functions as a strong regularizer, improving the robustness of automated sperm morphology analysis. Methods: We utilize soft distillation to transfer knowledge from a set of high-capacity teacher models to a smaller student model (SwinV2-base). The teacher architectures include SwinV2-large, EfficientNetV2-m, and ConvNeXtV2-large. To maximize performance, we investigated two distillation strategies: a single-teacher approach, where the student learns from one specific architecture, and a multi-teacher approach, where the student learns from an averaged response of multiple teachers. The models were trained on the imbalanced Hi-LabSpermMorpho dataset, which comprises 18 different sperm morphology categories derived from three differently stained (BesLab, Histoplus, GBL) sample sets. We adopted a cross-dataset training approach in which the teacher models were fine-tuned using the combination of two stained datasets, and the student model was trained on the third, distinct stained dataset. The global loss function combined cross-entropy loss with Kullback–Leibler divergence, employing the teacher’s soft probabilities to prevent the student from over-confidence. Results: The experimental results demonstrate that the student model trained in a multi-teacher setup with augmentation and soft distillation attains higher accuracies (70.94% on BesLab, 73.61% on Histoplus, 71.63% on GBL) than the baseline models. Conclusions: This approach mitigates challenges associated with data scarcity and heavily unbalanced sperm morphology datasets, providing consistent improvements and offering a highly generalizable solution for clinical diagnostics. Full article

Ground-level ozone (GLO₃) poses a critical threat to public health and the success of the Saudi Green Initiative, yet its long-term spatiotemporal evolution across the Arabian Peninsula remains poorly constrained. Utilizing CAMS-derived mixing ratios (1000–850 hPa) from 2003 to 2024, this study identifies a major systemic regime shift occurring in 2016–2017, marking a transition toward a more O₃-enriched atmospheric state across Saudi Arabia. While the early study period was characterized by pronounced spatial heterogeneity, post-2017 diagnostics reveal a synchronized intensification of GLO₃, particularly within the urban industrial belts of the Eastern and Western Provinces. Statistical trend metrics, including Mann–Kendall and regime-shift detection, show a persistent upward trend in GLO₃ concentrations, most significantly during winter and over the southwestern highlands. These trends are robustly coupled with increasing boundary-layer height, temperature, and UV-B radiation, alongside shifting precursor stoichiometry (CO, VOCs, NO_x) that separates titration-dominated from production-dominated regimes. Our results suggest that this mid-decade intensification reflects a convergence of anthropogenic forcing under Saudi Vision 2030 and shifting regional climatic drivers. By uncovering the transition from localized variability to kingdom-wide synchronization, this research provides a process-based foundation for targeted air quality management and the safeguarding of regional sustainability frameworks. Full article

Background/Objectives: Echocardiography reports are essential diagnostic tools, but their complexity and specialized English terminology frequently hinder comprehension for non-specialists and patients. This study addresses these accessibility gaps by developing a resource-efficient large language model (LLM) system designed to translate and summarize English echocardiography results into Traditional Chinese. Methods: To overcome significant hardware constraints, we utilized Quantized Low-Rank Adapter (QLoRA) techniques and the Unsloth acceleration framework to fine-tune LLaMA-3.2-1B and LLaMA-3.2-3B-Instruct models on a single mid-tier GPU. The system employs a dual-stage inference architecture: the first stage provides technical medical translation for clinicians, while the second stage generates simplified, patient-centric educational summaries to enhance health literacy. Results: Evaluation across multiple metrics, including BLEU, ROUGE, METEOR, and Perplexity, demonstrated that the LLaMA-3.2-3B-Instruct model with the AdamW 8-bit optimizer achieved the most stable validation performance, excelling in semantic coherence and structural consistency. A preliminary qualitative error analysis conducted in the Discussion section further identified clinical nuances, such as terminology simplification and minor hallucinations, underscoring the critical necessity of a Human-in-the-Loop verification procedure. Conclusions: These findings validate the feasibility of deploying cutting-edge medical AI in resource-limited clinical environments. While the results reflect validation-only performance on a specialized dataset, the platform offers a scalable foundation for enhancing clinical decision support and health literacy through accessible, automated medical text processing. Full article

This study investigates the integration of ABC/XYZ (value-based classification/ demand variability classification) inventory classification with demand forecasting models (ETS - Error, Trend, Seasonality, ARIMA - AutoRegressive Integrated Moving Average, Prophet - type of additive model) in a manufacturing enterprise to support sustainable resource management. The research aims to evaluate the inventory structure, demand variability, and forecasting accuracy across different material categories. The results confirm a strong concentration of inventory value in A-class items and significant differences in forecast accuracy across ABC/XYZ segments. While AX items generally exhibit lower forecast errors, notable exceptions highlight the need for additional diagnostic analysis. The findings demonstrate that integrating classification and forecasting improves inventory decision-making, reduces excess stock, and supports sustainable resource utilization. The proposed approach provides practical guidance for optimizing inventory management in industrial environments. Full article