Search Results (4,806)

Objectives: The objective of this study was to explore differences in short-term clinical response, renal outcomes, and healthcare resource utilization between two non-equivalent outpatient furosemide strategies: high-dose oral solution and subcutaneous administration. Methods: This retrospective observational study included 108 consecutive outpatients with advanced heart failure (HF) and diuretic resistance treated between January 2018 and December 2024. Patients received furosemide for 5 days as oral solution (250 mg/day; n = 81) or subcutaneous infusion via elastomeric pump (100 mg/day; n = 27). Primary endpoints were changes in functional status (NYHA [New York Heart Association] class), body weight, and renal function, assessed by changes in serum creatinine and estimated glomerular filtration rate and by clinically relevant worsening renal function criteria. Secondary endpoints included NT-proBNP changes and unplanned healthcare utilization at 30 days. Results: Both strategies were associated with short-term decongestion. Weight decreased similarly in both groups, with no difference in the proportion of patients achieving weight loss (oral 90% vs. subcutaneous 89%; p = 0.854). Functional improvement was more frequently observed in the oral-solution group (99% vs. 89%; p = 0.019). Renal outcomes were comparable between strategies, with no significant changes in eGFR and a similar proportion of patients experiencing a serum creatinine increase ≥0.3 mg/dL (25.9% vs. 22.2%; p = 0.85). No significant differences were observed in NT-proBNP changes. HF-related healthcare encounters decreased during the 30 days following treatment compared with the preceding 30-day period in the oral-solution group, whereas no significant within-group change was observed in the subcutaneous group. Conclusions: In this exploratory real-world cohort of patients with advanced HF and refractory congestion, both outpatient strategies were associated with short-term decongestion, with differences observed in symptomatic response. These findings are hypothesis-generating and warrant confirmation in prospective studies using comparable dosing strategies. Full article

Cardiovascular disease diagnosis requires accurate and timely analysis of electrocardiogram (ECG) signals to support reliable clinical decision-making. However, ECG signals are inherently non-stationary, exhibit substantial inter-patient variability, and may share similar morphological patterns across different cardiac disorders, making automated multi-class diagnosis challenging. This study proposes a multi-domain machine learning framework for automated ECG-based cardiac disease classification, integrating signal preprocessing, heartbeat segmentation, Variational Mode Decomposition (VMD), multi-domain feature extraction, minimum Redundancy Maximum Relevance (mRMR) feature selection, and hybrid CNN–Transformer learning. Experiments are conducted on the PTB-XL database using five diagnostic superclasses: NORM, MI, STTC, CD, and HYP. First, a fourth-order Butterworth band-pass filter (0.5–40 Hz) is applied to remove baseline wander and high-frequency noise, followed by adaptive Symlet-8 wavelet denoising with soft thresholding to suppress residual high-frequency fluctuations while preserving the P-wave, QRS complex, and T-wave morphology, after which the signal is z-score normalized. R-peaks are subsequently detected to segment standardized cardiac cycles. VMD is then employed to decompose the heartbeat signals into intrinsic modes, from which the most informative modes are retained using correlation-based mode selection. Temporal, statistical, spectral, and nonlinear features are extracted to capture complementary characteristics of cardiac electrical activity, while mRMR selects the eight most informative features by maximizing feature relevance and minimizing redundancy. The resulting representation is processed through a hybrid CNN–Transformer architecture, in which convolutional layers learn local morphological patterns and Transformer-based attention captures long-range dependencies within the cardiac feature representation. The proposed framework achieves 93.60% accuracy, 93.61% macro precision, 93.60% macro recall, 93.60% macro F1-score, and 98.40% macro specificity across the five diagnostic classes. Confusion-matrix analysis, receiver operating characteristic (ROC) analysis, comparative evaluation, and ablation experiments further demonstrate the discriminative capability and robustness of the proposed approach. These findings indicate that multi-domain biomedical feature learning combined with attention-based deep learning can provide an effective and robust strategy for automated ECG-based cardiac disease classification, highlighting the potential of machine learning for intelligent biomedical signal analysis and computer-aided clinical diagnosis. Full article

Radio frequency (RF)-based drone identification enables passive low-altitude sensing, but its performance degrades under low signal-to-noise ratio (SNR) and long-range propagation. Existing deep models mainly learn spectrogram amplitude textures and underuse the second-order structure and cross-channel correlations of multi-channel RF signals. We propose the Multi-resolution Riemannian-Spherical Network (MRS-Net), a robust identification framework centered on Riemannian Structure Preservation (RSP). RSP derives noise-referenced Riemannian distance, local geometric variation, log-determinant, and multi-scale statistics from local symmetric positive definite (SPD) covariance matrices. Pairwise similarity alignment transfers these structural relationships into the convolutional neural network (CNN) embedding space as a training-stage teacher signal. RSP is removed at inference and therefore adds no online manifold computation. Multi-resolution short-time Fourier transform (STFT) representations provide complementary weak-signal observations, while CosFace enlarges inter-class angular margins. Experiments on DroneRFa and Noisy Drone RF evaluate propagation attenuation and additive noise, respectively. With four-channel DroneRFa input, MRS-Net achieves 90.30% overall balanced accuracy. Compared with SR-CNN-4ch, it improves the 80–150 m result from 61.30% to 78.30%; compared with MR-Spherical-4ch, it reduces the 5-fold standard deviation from 16.78% to 1.46%. Ablations show that RSP is most effective when cross-channel covariance is informative and the backbone preserves local time-frequency structure. Full article

Attention-Deficit/Hyperactivity Disorder (ADHD) symptoms show substantial developmental heterogeneity, yet less is known about how co-occurring psychopathology is associated with membership in different longitudinal symptom trajectories. This study examined developmental trajectories of parent-reported ADHD symptoms across seven annual assessment waves from late childhood through adolescence using data from the Adolescent Brain Cognitive Development (ABCD) Study. ADHD symptoms were assessed using the Child Behavior Checklist DSM-5-Oriented ADHD Problems scale and therefore represent dimensional symptom scores rather than clinical ADHD diagnoses. Growth mixture modelling was conducted in an analytic sample of 9718 participants and identified three trajectory classes: low/stable (79.87%), high/stable (10.77%), and remitting (9.36%). Baseline associations with class membership were subsequently examined using the R3STEP procedure in 9716 participants, preserving the identified class structure while accounting for classification uncertainty. Relative to the low/stable class, male sex and higher baseline affective, anxiety, conduct, and oppositional defiant problem scores were associated with greater odds of membership in both the remitting and high/stable classes, whereas age and somatic problems showed no clear associations. Comparisons between the remitting and high/stable classes were largely similar, although male sex was associated with greater odds of remitting-class membership. Overall, the findings demonstrate substantial heterogeneity in the developmental course of ADHD symptoms and suggest that baseline co-occurring psychopathology is associated primarily with differentiation between lower- and higher-symptom trajectories rather than clearly distinguishing persistent from remitting elevated-symptom pathways. Full article

Turmeric (Curcuma longa L.) is one of the most important spice crops and a valuable medicinal plant, but it is seriously affected by various types of diseases such as fungal, bacterial, nematode and viral diseases. In this paper, a complete mathematical model of the turmeric plant disease dynamics is developed under a fractal-fractional model in this context, encompassing all four types of pathogens and associated treatment classes. The fractal-fractional Caputo derivative operator captures memory effects and, through its fractal exponent, a genuine deformation of the classical memory kernel, allowing the underlying biological dynamics to be represented more flexibly than under the classical integer-order derivative; we do not, however, claim that this kernel deformation corresponds to demonstrated self-similarity or spatial heterogeneity in the turmeric plant–pathogen system. We show the positivity and boundedness of the solutions, calculate the next-generation matrix approach-based basic reproduction number ${\mathcal{R}}_0$ and investigate the local and global stability of both disease-free and endemic equilibria by Lyapunov functionals. A sensitivity analysis of ${\mathcal{R}}_0$ is conducted to determine the most important parameters influencing disease transmission and control. The existence and uniqueness of solutions and Ulam-Hyers stability of solutions are established by fixed point theory. For the associated integer-order system, we formulate an optimal control problem is formulated with three time-dependent controls: the prevention effort ($u_1$), the enhancement of treatment ($u_2$), and the care management ($u_3$), and the optimality conditions are derived via Pontryagin’s maximum principle. Numerical simulations are conducted with three different fractal-fractional operators, namely Caputo, Caputo-Fabrizio and Atangana-Baleanu. A deep neural network is developed and trained to approximate the solution of the integer-order system. The third-layer deep neural network consists of neurons of sizes 80, 32, and 24, with activation functions of logistic sigmoid, radial basis and hyperbolic tangent, respectively, and is trained to approximate the system dynamics with the fourth-order Runge-Kutta method as a reference. The DNN is found to be very accurate in predicting the values with Nash-Sutcliffe Efficiency between 0.79 and 0.99 and Theil Inequality Coefficient around ${10}^{-2}$ in all 11 compartments, and hence proved capable of being a good surrogate modelling tool for the ODE systems. The present work contributes towards SDG 2 (Zero Hunger) and SDG 3 (Good Health and Well-being) by laying a mathematical basis for integrated disease management in turmeric cultivation for sustainable agriculture and food security. Full article

Background/Objectives: This randomized, split-mouth clinical trial compared the 72-month clinical performance of a giomer-based restorative material (BEAUTIFIL II LS) and a nanohybrid restorative composite (Filtek Supreme Universal) in Class V non-carious cervical lesions (NCCLs). Methods: Forty-nine participants, each with two paired NCCLs, were enrolled (98 restorations) and received both materials under a standardized total-etch adhesive protocol. Clinical performance was assessed at 72 months using modified Hickel/FDI criteria across esthetic, functional, and biological domains, with between-material comparisons performed by paired nonparametric analysis (Wilcoxon signed-rank test) for complete paired observations. Results: Of the 49 participants enrolled, 26 attended the 72-month follow-up visit and 25 contributed evaluable data, yielding 44 evaluable restorations. No restoration was lost, and no statistically significant differences were detected between BEAUTIFIL II LS and Filtek Supreme Universal for any evaluated esthetic, functional, or biological criterion. Both materials showed acceptable clinical performance, with minor changes related mainly to marginal staining and marginal adaptation over time. Conclusions: BEAUTIFIL II LS and Filtek Supreme Universal showed similar long-term clinical performance in Class V non-carious cervical lesion restorations over 72 months. Within the limitations imposed by long-term attrition, material selection may be guided by clinical requirements, handling characteristics, and clinician preference. Full article

Pathology foundation models have recently emerged as powerful pretrained representations for computational pathology, yet whether complex downstream modeling is still necessary once frozen representations are evaluated under a common downstream framework remains insufficiently understood. We address this question for microsatellite-instability-high (MSI-H) prediction in colorectal cancer by benchmarking nine frozen encoders—five pathology-specific (CONCH, CONCH v1.5, UNI, Virchow2, Phikon) and four conventional vision backbones (ResNet18, ResNet50, ViT-B/16, ConvNeXt-Tiny)—for colorectal cancer histopathology under a patient-level framework, using TCGA-COAD/READ as the development cohort and CPTAC-COAD as an independent external cohort. For each encoder, H&E tiles were embedded without fine-tuning, mean-pooled to slide and patient representations, and evaluated with the same logistic regression linear probe, alongside nonlinear classifiers and a gated-attention multiple-instance learning (MIL) comparator. The complete MANTIS-defined 206-patient cohort (82 MSI-H, 124 MSS) is reported as the primary development benchmark. A label provenance audit identified 15 of 206 development cohort patients (7.3%) with cross-source MSI provenance discordance, and the concordant-label 191-patient cohort (68 MSI-H, 123 MSS) is reported as a secondary sensitivity analysis. In the primary cohort, the selected pathology-specific encoders had a higher mean pooled out-of-fold AUROC than the selected conventional vision backbones (0.827 vs. 0.773; paired-bootstrap difference, +0.054; 95% interval, +0.015 to +0.094), which is reported as a descriptive benchmark rather than a formal inference about the model classes. Virchow2 and UNI produced the strongest linear probe discrimination (AUROC 0.861 and 0.855, respectively), with no consistent gain from the tested nonlinear classifiers or attention–MIL configurations, and representational similarity analysis (linear centered kernel alignment) confirmed that encoders occupy distinct feature geometries rather than converging to a shared representation. In external validation on CPTAC-COAD (105 patients; 24 MSI-H, 81 MSS), selected TCGA-trained models retained variable discrimination, including CONCH attention–MIL AUROC 0.880 and UNI linear probe AUROC 0.859, but external calibration and threshold behavior varied substantially; for CONCH attention–MIL, the development-derived threshold did not transport under the external ensemble implementation (specificity 0.000), whereas threshold-only local adaptation on a small external subset improved median specificity to 0.686 without updating model weights. These findings indicate that modern pathology foundation models can encode MSI-associated morphology in frozen representations under this benchmark, while decision threshold transportability and multimodal or explainability extensions remain open questions for future work. Full article

To address the challenges of crop–weed instance segmentation and image-plane operation-point localization under visual similarity, background interference, leaf occlusion, and irregular plant morphology in cotton fields, a lightweight instance-segmentation model, GDMM-YOLO11, was developed. Based on YOLO11n-seg, the four backbone stage-transition downsampling convolutions at P2/4, P3/8, P4/16, and P5/32 were replaced with GhostConv to reduce redundant computation, a C3k2_DySnakeConv_Mona module was introduced to strengthen structural and multi-scale feature representation, and the original post-SPPF C2PSA block was replaced with mixed local channel attention (MLCA) to recalibrate high-level feature responses. Experiments were conducted on 2177 images containing 2856 annotated plant instances using a stratified 65%/15%/20% training–validation–test split. Across three independent runs with random seeds 3407, 3408, and 3409, GDMM-YOLO11 achieved a mask precision of 89.82 ± 2.60%, mask recall of 84.50 ± 2.49%, mask mAP@0.5 of 89.76 ± 0.66%, and mask mAP@0.5:0.95 of 68.50 ± 0.34%. Relative to YOLO11n-seg, the corresponding three-run mean values were numerically higher by 3.52, 1.07, 1.67, and 2.33 percentage points, respectively, while the parameter count decreased from 2.836 M to 2.533 M and the computational cost decreased from 9.6 to 8.8 GFLOPs. For image-plane operation-point localization, 505 of 528 ground-truth weed instances obtained valid same-class mask matches. The proposed skeleton-constrained fused-center method achieved a ground-truth-mask inclusion rate of 99.41%, a conditional Success@0.15 of 91.29%, and an end-to-end Success@0.15 of 87.31%. Paired comparisons with the mask-centroid baseline showed statistically significant improvements in ground-truth-mask inclusion and weed-boundary clearance, whereas differences in localization error and Success@0.10/0.15 were not statistically significant. TensorRT FP16 deployment on an NVIDIA Jetson AGX Orin achieved a model-only inference latency of 2.796 ± 0.115 ms, corresponding to 357.67 FPS. These results show that GDMM-YOLO11 provides a favorable accuracy–complexity trade-off while supporting image-plane operation-point generation and high-throughput model-only edge inference. Full article

We revisit the Hunter–Saxton equation through its classical first-order Lagrangian. The determining system for all first-order multipliers is reduced to one linear equation in two variables. Its solutions include both point-symmetry characteristics and genuinely generalised variational characteristics; in the analytic category, the remaining freedom is locally parameterised by two arbitrary analytic functions. Every member of the known infinite-dimensional point-symmetry ideal is shown to preserve the action up to a total divergence and produces an arbitrary-function family of conserved currents, whereas one finite point symmetry is excluded from the Noether point-symmetry algebra of the Lagrangian. We also correct an omission in the previously reported associations between finite currents and Lie point symmetries. For the scaling symmetry, the radial component of an associated current vanishes identically after transformation. Alignment and multiplier criteria explain this degeneracy and show when it is a property of the conservation-law class. Thus association guarantees invariance of the transformed component, but not that the component retains differential content. Reducing the Lagrangian instead recovers the known similarity equations and their first integrals for two representative symmetries. In the scaling reduction, the reduced Noether symmetry is induced by a commuting member of the infinite-dimensional ideal. Full article

This study presents an inappropriate-content classification framework for digital-violence detection in Ecuadorian Spanish, addressing extreme class imbalance and dialectal variation. Data were collected in a post-API setting through a Selenium-based scraping pipeline that reconstructs conversational context using a window of ? = 3 prior interventions. An initial zero-shot labeling attempt with LLMs (Hermes) revealed severe cultural misinterpretation, overestimating the Violence class by 50 times (97.5% false alerts), which motivated full human validation and targeted data engineering. To correct imbalance without contaminating evaluation, the corpus was split before augmentation (70/15/15), and minority classes were selectively leveled via few-shot generation with LLaMA 3.1, followed by strict deduplication and cosine-similarity filtering (? = 0.85) to preserve semantic diversity. Model selection compared BETO and mBERT, with BETO outperforming. Across four training scenarios, naïve oversampling produced artificially inflated metrics indicative of overfitting, whereas the proposed cost-sensitive and regularized configuration (BETO with semantic deduplication and weighted loss) achieved 94.39% accuracy, 0.9429 weighted F1, and 0.9022 macro F1, significantly improving recovery of critical classes. Results highlight that hybrid data are effective only when carefully curated and paired with leakage-free evaluation protocols. This work demonstrates how machine learning innovation and knowledge extraction from heterogeneous data can be combined to build robust models for digital-violence detection in low-resource, culturally specific contexts. Full article

Active responsive façades are commonly classified by function or technology, while their visual organization across component and whole-façade scales remains underexplored. This study asks whether image features can organize cross-scale differences and whether cluster-derived semantics can improve generative steering. It introduces a traceable dual-scale workflow linking visual features, clustering, architectural semantics, and image generation. The dataset comprises 75 matched pairs of component close-ups (Set A) and whole-façade views (Set B). Each image was encoded by 43 features: SSIM, mean RGB values, a 32-bin gradient orientation histogram, and seven Hu moments. In two-dimensional PCA space, K-means yielded six scale-specific clusters, interpreted through feature evidence, metadata, and researcher-reviewed GPT-assisted descriptions. Cross-scale agreement was weak but nonrandom (ARI = 0.033; NMI = 0.194; p = 0.0065); however, this association became nonsignificant after feature-group equal weighting (p = 0.829). Tectonic type was associated with clustering at both scales. In a 144-image experiment, semantic prompts increased visual–prototype similarity by 0.0339 (p = 0.0215) and the own-prototype–strongest-competitor margin by 0.0306 (p = 0.0093). Top-1 identification rose from 13.9% to 44.4%, although the difference was nonsignificant (p = 0.0586). The contribution is a traceable visual–semantic process for exploratory generative steering, rather than a fixed taxonomy or stable class-specific control. Full article

Screening for §103 propensity at the idea stage is inherently difficult: obviousness is a context-sensitive legal determination over prior art combinations that surface-level text cannot fully capture. Yet because §103 concerns the inventive step of the underlying idea rather than only the wording of the claims, an application as filed may already carry a weak signal of §103 propensity—motivating a screening tool at the idea stage. We present TriageRAG (T-RAG), a confidence-based decision-support framework. A fine-tuned ModernBERT-large classifier produces a prediction together with a confidence score; high-confidence cases are delivered directly, while only low-confidence cases are escalated to a large language model (LLM), which is supplied with the classifier’s own prediction as the primary signal together with retrieved similar prior applications, and is instructed to verify the classifier rather than replace it. We evaluate under deliberately leakage-free conditions—a same-era corpus, a temporal hold-out, and a contamination-free label set in which the §103 label follows the USPTO Office Action Research Dataset and the two classes are disjoint by construction. Under these strict conditions the system remains useful: the classifier confidence rank-orders correctness well enough to support high-precision automatic decisions at low coverage, and classifier-primary escalation improves accuracy precisely on the uncertain cases where the classifier is weakest, without degrading it overall. We position the contribution as a triage architecture—turning a deliberately commodity classifier into an auditable decision-support tool—rather than as a new classifier. Ablation studies isolate the roles of confidence routing, retrieval design, and the escalation prompt, and characterize the accuracy–cost trade-off across the escalation threshold. Full article

Future industrial human activity recognition (HAR) may require discrimination among many operational elements, including movements with partially overlapping kinematics. This controlled pilot study characterized recognition across 22 upper-body movement classes with varying structural similarity and examined classifier, signal scaling, and temporal window length. Sixteen adults performed the movements with an XSENS motion-capture system; eight upper-limb inertial measurement units provided 80 synchronous time-series channels. Thirteen participants were used for model development, and three whose observed execution patterns differed comparatively from the remainder were deliberately reserved as a challenge-oriented holdout. Support vector classifier (SVC), random forest (RF), Gaussian naive Bayes (NB), and long short-term memory (LSTM) models were evaluated with min–max or maximum-absolute scaling and 62-, 93-, or 124-frame windows. RF with maximum-absolute scaling and a 124-frame window achieved the best aggregate holdout performance (accuracy = 0.950; F1 = 0.939). Importantly, this performance was obtained on three entirely unseen participants who were deliberately reserved because their observed execution patterns and fluency differed from those of the model-development participants, providing a controlled, challenge-oriented test of transfer across inter-individual execution variability. Nevertheless, strong aggregate performance did not translate into uniform class-level reliability, and prominent errors remained concentrated in specific movement pairs. These findings provide empirical evidence for both the participant-independent transfer capability and the class-specific limitations of motion-only recognition, supporting its role as a methodological precursor to future AI-assisted work study and human–robot collaboration rather than as evidence of end-to-end industrial HAR. Full article

Manual sorting is still the mainstream scheme for component identification in down feather quality evaluation, which suffers from low detection efficiency and poor stability. To address these drawbacks, this paper proposes a fine-grained component detection algorithm based on YOLO12 for down feather quality classification. Five typical down feather components are selected as detection targets, including discolored feathers, down filaments, immature down, feathers and pure down. A dedicated down feather object detection dataset consisting of 1140 images is established accordingly. To improve the feature representation capacity of the model for tiny objects, faint boundary features and subtle distinctions between analogous categories, this work integrates the Convolutional Gated Linear Unit (CGLU) into the A2C2f module of YOLO12. While maintaining the original feature aggregation pathways and residual architecture, the conventional MLP feed-forward branch within ABlock is replaced with convolutional gated transformation. Experimental results demonstrate that the proposed A2C2f-CGLU model achieves precision of 96.50%, recall of 94.49%, mAP50 of 98.05% and mAP50-95 of 57.89% with the optimal weights on the validation set. Compared with the original YOLO12, the mAP50-95 metric is elevated by 3.04 percentage points, and the overall performance surpasses two comparative variants, A2C2f-DFFN and A2C2f-KAN. Visualizations of PR curves, confusion matrices and real test samples validate that the proposed method effectively enhances the recognition stability of tiny down feather targets and similar classes. This research provides a visual inspection foundation for subsequent component proportion calculation, quality grade discrimination and the development of intelligent detection systems. Full article

In developing countries, railway networks often create operational bottlenecks by combining older single-track sections with modern, double-track, signaled infrastructure. This study addresses train scheduling in such hybrid networks using a segment-based rather than a train-based approach. A mixed integer linear programming (MILP) model is developed that minimizes total waiting time across all journey segments, incorporating directional safety intervals separating single-track and double-track sections and co-directional and counter-directional movements. Train-type priority is included in the objective function as a class-dependent, age-adjusted weight, rather than a rigid priority rule; thus preventing the unlimited delay that a rigid priority ranking might otherwise impose on lower-priority trains in heavily shared resources. Since an exact MILP solution is impractical on a national scale, a segment-based heuristic method is developed; the age-weighted heuristic method, validated against exact solutions on smaller subnetworks, is found to require 35% more time to reach full optimum. This approach, applied to Turkey’s national railway network (109 routes, 997 stations or siding sections, 2281 transit segment between two stations or siding sections, 1539 daily services), has reduced daily waiting times from 2508 h to 609 h. This corresponds to an annual saving of approximately $13.2 million in fuel and labor costs and a reduction of 24,260 tons/year in CO₂ emissions. Since diesel-fueled services account for more than 99% of the cost savings, sensitivity analyses show that these estimates are robust against fuel price assumptions. The study offers a low-cost solution for improving railway capacity before costly infrastructure investments and is expected to be generalized to similar heterogeneous national networks. Full article