Search Results (4,786)

Early detection of small-scale fires is crucial for minimizing damage and enabling rapid emergency response. While recent deep learning-based fire detection systems have achieved high accuracy, they still face three key challenges: (1) limited deployability in resource-constrained edge environments due to high computational costs, (2) performance degradation caused by feature interference when jointly learning flame and smoke features in a single backbone, and (3) low sensitivity to small flames and thin smoke in the initial stages. To address these issues, we propose a lightweight dual-stream fire detection architecture based on YOLOv5n, which learns flame and smoke features separately to improve both accuracy and efficiency under strict edge constraints. The proposed method integrates two specialized attention modules: ESCFM++, which enhances spatial and channel discrimination for sharp boundaries and local flame structures (flame), and ESCFM-RS, which captures low-contrast, diffuse smoke patterns through depthwise convolutions and residual scaling (smoke). On the D-Fire dataset, the flame detector achieved 74.5% mAP@50 with only 1.89 M parameters, while the smoke detector achieved 89.2% mAP@50. When deployed on an NVIDIA Jetson Xavier NX(NVIDIA Corporation, Santa Clara, CA, USA)., the system achieved 59.7 FPS (single-stream) and 28.3 FPS (dual-tream) with GPU utilization below 90% and power consumption under 17 W. Under identical on-device conditions, it outperforms YOLOv9t and YOLOv12n by 36–62% in FPS and 0.7–2.0% in detection accuracy. We further validate deployment via outdoor day/night long-range live-stream tests on Jetson using our flame detector , showing reliable capture of small, distant flames that appear as tiny cues on the screen, particularly in challenging daytime scenes. These results demonstrate overall that modality-specific stream specialization and ESCFM attention reduce feature interference while improving detection accuracy and computational efficiency for real-time edge-device fire monitoring. Full article

Introduction: Low vision affects more than visual acuity; it substantially disrupts daily functioning and may contribute to long-term cognitive, emotional, and social consequences. When medical or surgical treatment options are no longer effective, structured low-vision rehabilitation becomes essential, providing strategies and tools that support functional adaptation and promote independence. This review aims to map the current outcomes of rehabilitation services, identify gaps in existing research, and highlight opportunities for further study. Methods: An article search was conducted via PubMed, Scopus, PsycInfo, and Google Scholar. Then, title, abstract, and full-text screenings for inclusion were performed by all the authors independently, and disagreements were resolved through discussion. The relevant outcomes from the eligible publications were extracted by four authors and then cross-checked by the other authors. The results are presented via the Preferred Reporting Items for Systematic Reviews and Meta-analysis extension for Scoping Reviews checklist. Results: A total of 13 studies met the inclusion criteria. Most were randomized controlled trials (n = 10,77%), with the majority conducted in the United States and the United Kingdom. Study populations consisted of adults aged 18 years and older. Across the included studies, low-vision rehabilitation interventions particularly visual training, magnification-based programs, and multidisciplinary approaches, were associated with significant improvements in visual function, activities of daily living, and vision-related quality of life. Conclusions: Low vision rehabilitation interventions demonstrate clear benefits for visual acuity, contrast sensitivity, reading speed, and functional independence. However, substantial gaps remain, including limited evidence on long-term outcomes, inconsistent assessment of psychosocial influences, and underrepresentation of diverse populations. Standardized outcome measures and long-term, inclusive research designs are needed to better understand the sustained and equitable impact of low-vision rehabilitation. Full article

Stone cultural relics are primarily composed of sandstone, a water-sensitive rock that is highly susceptible to deterioration from environmental solutions and dry-wet cycles. Sandstone pagodas are often directly exposed to natural elements, posing significant risks to their preservation. Therefore, it is crucial to investigate the performance of sandstone towers in complex solution environments and understand the degradation mechanisms influenced by multiple environmental factors. This paper focuses on the twin towers of the Huachi Stone Statue in Qingyang City, Gansu Province, China, analyzing the changes in chemical composition, surface/microstructure, physical properties, and mechanical characteristics of sandstone under the combined effects of various solutions and dry-wet cycles. The results indicate that distilled water has the least effect on the mineral composition of sandstone, while a 5% Na₂SO₄ solution can induce the formation of gypsum (CaSO₄·2H₂O). An acidic solution, such as sulfuric acid, significantly dissolves calcite and diopside, leading to an increase in gypsum diffraction peaks. Additionally, an alkaline solution (sodium hydroxide) slightly hydrolyzes quartz and albite, promoting calcite precipitation. The composite solution demonstrates a synergistic ion effect when mixed with various single solutions. Microstructural examinations reveal that sandstone experiences only minor pulverization in distilled water. In contrast, the acidic solution causes micro-cracks and particle shedding, while the alkaline solution results in layered spalling of the sandstone surface. A salt solution leads to salt frost formation and pore crystallization, with the composite solution of sodium hydroxide and 5% Na₂SO₄ demonstrating the most severe deterioration. The sandstone is covered with salt frost and spalling, exhibiting honeycomb pores and interlaced crystal structures. From a physical and mechanical perspective, as dry-wet cycles increase, the water absorption and porosity of the sandstone initially decrease slightly before increasing, while the longitudinal wave velocity and uniaxial compressive strength continually decline. In summary, the composite solution of NaOH and 5% Na₂SO₄ results in the most significant deterioration of sandstone, whereas distilled water has the least impact. The combined effects of acidic/alkaline and salt solutions generally exacerbate sandstone damage more than individual solutions. This study offers insights into the regional deterioration characteristics of the Huachi Stone Statue Twin Towers and lays the groundwork for disease control and preventive preservation of sandstone cultural relics in similar climatic and geological contexts. Full article

Wildfires are increasing globally and pose major challenges for assessing post-fire vegetation recovery and ecosystem resilience. We analyzed long-term Landsat time series in two contrasting fire-prone ecosystems in the United States and Australia. Vegetation area was extracted using the Enhanced Vegetation Index (EVI) with Otsu thresholding. Recovery to pre-fire baseline levels was modeled using linear, logistic, locally estimated scatterplot smoothing (LOESS), and generalized additive models (GAM), and their performance was compared using multiple metrics. The results indicated rapid recovery of Australian forests to baseline levels, whereas this was not the case for forests in the United States. Among climatic factors, temperature was the dominant parameter in Australia (Spearman ρ = 0.513, p < 10⁻⁸), while no climatic variable significantly influenced recovery in California. Methodologically, GAM consistently performed best in both regions due to its success in capturing multiphase and heterogeneous recovery patterns, yielding the lowest values of AIC (United States: 142.89; Australia: 46.70) and RMSE_cv (United States: 112.86; Australia: 2.26). Linear and logistic models failed to capture complex recovery dynamics, whereas LOESS was highly sensitive to noise and unstable for long-term prediction. These findings indicate that post-fire recovery is inherently nonlinear and ecosystem-specific and that simple models are insufficient for accurate estimation, with GAM emerging as an appropriate method for assessing vegetation recovery using remote sensing data. This study provides a transferable approach using remote sensing and GAM to monitor forest resilience under accelerating global fire regimes. Full article

Escherichia coli (E. coli) is a well-established indicator of faecal pollution and a potent pathogen linked to numerous gastrointestinal and systemic illnesses. Ensuring public safety requires rapid and sensitive detection methods capable of real-time, on-site deployment. Many conventional techniques are either laborious, time-intensive, costly, or require complex infrastructure, limiting their applicability in field settings. Raman spectroscopy offers label-free molecular fingerprinting; however, its inherently weak scattering signals restrict its effectiveness as a standalone technique. Surface-Enhanced Raman Spectroscopy (SERS) overcomes this limitation by exploiting plasmonic enhancement from nanostructured metallic substrates—most commonly gold, silver, copper, and aluminium. Despite the commercial availability of SERS-active substrates, challenges remain in achieving high reproducibility, long-term stability, and true field applicability, necessitating the development of integrated lab-on-chip platforms and portable, handheld Raman devices. This review critically examines recent advances in SERS-based E. coli detection across water and perishable food products with particular emphasis on the evolution of SERS substrate design, the incorporation of biosensing elements, and the integration of electrochemical and microfluidic systems. By contrasting conventional SERS approaches with next-generation biosensing strategies, this paper outlines pathways toward robust, real-time pathogen detection technologies suitable for both laboratory and field applications. Full article

Surface deformation induced by the extraction of natural resources constitutes a non-stationary spatiotemporal process. Modeling surface deformation time series obtained through Interferometric Synthetic Aperture Radar (InSAR) technology using deep learning methods is crucial for disaster prevention and mitigation. However, the complexity of model hyperparameter configuration and the lack of interpretability in the resulting predictions constrain its engineering applications. To enhance the reliability of model outputs and their decision-making value for engineering applications, this study presents a workflow that combines a Tree-structured Parzen Estimator (TPE)-based Bayesian optimization approach with ensemble inference. Using the Rhineland coalfield in Germany as a case study, we systematically evaluated six deep learning architectures in conjunction with various spatiotemporal coding strategies. Pairwise comparisons were conducted using a Welch t-test to evaluate the performance differences across each architecture under two parameter-tuning approaches. The Benjamini–Hochberg method was applied to control the false discovery rate (FDR) at 0.05 for multiple comparisons. The results indicate that TPE-optimized models demonstrate significantly improved performance compared to their manually tuned counterparts, with the ResNet+Transformer architecture yielding the most favorable outcomes. A comprehensive analysis of the spatial residuals further revealed that TPE optimization not only enhances average accuracy, but also mitigates the model’s prediction bias in fault zones and mineralize areas by improving the spatial distribution structure of errors. Based on this optimal architecture, we combined the ten highest-performing models from the optimization stage to generate a quantile-based susceptibility map, using the ensemble median as the central predictor. Uncertainty was quantified from three complementary perspectives: ensemble spread, class ambiguity, and classification confidence. Our analysis revealed spatial collinearity between physical uncertainty and absolute residuals. This suggests that uncertainty is more closely related to the physical complexity of geological discontinuities and human-disturbed zones, rather than statistical noise. In the analysis of super-threshold probability, the threshold sensitivity exhibited by the mining area reflects the widespread yet moderate impact of mining activities. By contrast, the fault zone continues to exhibit distinct high-probability zones, even under extreme thresholds. It suggests that fault-controlled deformation is more physically intense and poses a greater risk of disaster than mining activities. Finally, we propose an engineering decision strategy that combines uncertainty and residual spatial patterns. This approach transforms statistical diagnostics into actionable, tiered control measures, thereby increasing the practical value of susceptibility mapping in the planning of natural resource extraction. Full article

Understanding how climatic variability affects growth and water relations of Scots pine (Pinus sylvestris L.) is essential for assessing stand sustainability in hemi-boreal regions. Linear mixed-effects models were used to quantify the effects of climatic variability and tree characteristics on stem volume increment (ZV), sap flow (SF), and water-use efficiency (WUE) of Scots pine growing on highly oligotrophic soils in Curonian Spit National Park. Annual ZV was strongly controlled by tree size and seasonal temperature conditions. Higher temperatures in late winter and mid-summer enhanced growth, whereas elevated temperatures in April–May reduced increment. June moisture availability, expressed by the hydrothermal coefficient, had a positive effect, highlighting the sensitivity of growth to early-summer drought and heat waves. Sap-flow density during May–October was primarily driven by climatic factors, with temperature stimulating and relative humidity reducing SF, while tree size played a minor role. Random-effects analysis showed that unexplained variability in ZV was mainly associated with persistent differences among trees and sites, whereas SF variability occurred largely at the within-tree level. In contrast, WUE was dominated by climatic drivers, with no detectable site- or tree-level random effects. Higher June precipitation increased WUE, while warmer growing-season conditions reduced it. Overall, Scots pine growth and WUE are mainly regulated by intra-annual climatic conditions, particularly summer water availability. Despite rapid climatic change, no critical physiological thresholds or growth collapse were detected during the study period, indicating substantial adaptive capacity of Scots pine even under the observed exceptional conditions. Full article

The background error covariance matrix (BE) is a fundamental component of data assimilation (DA) systems. Its impact on both the DA process and subsequent forecast performance depends on model configuration and the types of observations assimilated. However, few studies have specifically examined BE behavior in the context of satellite DA for regional tropical cyclone (TC) prediction. In this study, we develop the BE and evaluate its structure for a TC forecasting system over the western North Pacific. A total of six BEs are modeled using three control variable (CV) schemes (aligned with the CV5, CV6, and CV7 options available in the Weather Research and Forecasting DA system (WRFDA)) with training data from two distinct periods: the TC season and the winter season. Results demonstrate that the BE structure is sensitive to the training data used. The performance of TC-season BEs derived from different CV schemes is assessed for TC track forecasting through the assimilation of microwave sounder satellite brightness temperature data. The evaluation is based on a set of 14 cases from 2018 that exhibited large official track forecast errors. The CV7 BE, which uses the x- and y-direction wind components as CVs, captures finer small-scale momentum error features and yields greater forecast improvement at shorter lead-times (24 h). In contrast, the CV6 BE, which employs stream function (ψ) and unbalanced velocity potential (χ_u) as CVs, incorporates more large-scale momentum error information. The inherent multivariate couplings among analysis variables in this scheme also allow for closer fits to satellite microwave brightness temperature data, which is particularly crucial for forecasting TCs that primarily develop over oceans where conventional observations are scarce. Consequently, it enhances the large-scale environmental field more effectively and delivers superior forecast skill at longer lead times (48 h and 72 h). Full article

Greywater reuse presents a promising strategy for reducing potable water demand and supporting the irrigation of urban green infrastructure, yet its implementation in early planning phases remains limited by fragmented regulations, data gaps, and the absence of practical decision support tools. This study develops a comprehensive evaluation matrix based on Multi-Criteria Decision Analysis (MCDA) to assess the feasibility of greywater reuse in residential district development. The framework integrates eight domains (legal, technical, infrastructural, ecological, economic, and social factors) and is complemented by automated supporting worksheets for water balance, ecological indicators, and economic parameters. Application of the matrix to two contrasting residential case studies demonstrated its diagnostic value: the new-build district in Dortmund showed a high reuse potential, strongly influenced by favourable infrastructure conditions and ecological indicators, whereas the existing building in Weimar yielded a moderate potential due to infrastructural constraints and lower greywater availability. Sensitivity analyses further revealed that local water tariffs, intended-use scenarios, and stakeholder weightings substantially affect outcomes. Overall, the results show that the matrix supports transparent early-stage decision-making, identifies critical bottlenecks, and strengthens governance-oriented integration of greywater reuse in sustainable urban development. Full article

Background/Objectives: This study examined the clinical utility of diffusion-weighted whole-body magnetic resonance imaging with background body signal suppression (DWIBS) for differentiating infectious from non-infectious aortitis. Methods: The study included 32 patients with suspected inflammatory aortitis who underwent non-contrast computed tomography (NCCT) and magnetic resonance imaging. We evaluated the diagnostic performance of DWIBS using the spinal cord as a reference, NCCT, and their combination. The diagnosis of infectious aortitis was adjudicated based on imaging, clinical, and laboratory findings. We conducted a sensitivity analysis using a stricter definition of infectious aortitis that required both surgical and microbiological confirmation. Results: Fifteen patients were diagnosed with infectious aortitis. The sensitivity, specificity, and areas under the receiver operating characteristic curves were 93.3%, 70.6%, and 0.82, respectively, for NCCT; 93.3%, 76.5%, and 0.85, respectively, for DWIBS; and 86.7%, 94.1%, and 0.90, respectively, for the combination of both modalities. In the sensitivity analysis, the combined DWIBS and NCCT approach demonstrated a specificity of 87.5% and a sensitivity of 70.8%. Conclusions: DWIBS using the spinal cord as a reference appears to be a promising diagnostic tool for differentiating infectious from non-infectious aortitis, especially when combined with NCCT. Full article

This study re-evaluates ROC-AUC for binary classification under severe class imbalance (<3% positives). Despite its widespread use, ROC-AUC can mask operationally salient differences among classifiers when the costs of false positives and false negatives are asymmetric. Using three benchmarks, credit-card fraud detection (0.17%), yeast protein localization (1.35%), and ozone level detection (2.9%), we compare ROC-AUC with Matthews Correlation Coefficient, F2-score, H-measure, and PR-AUC. Our empirical analyses span 20 classifier–sampler configurations per dataset, combined with four classifiers (Logistic Regression, Random Forest, XGBoost, and CatBoost) and four oversampling methods plus a no-resampling baseline (no resampling, SMOTE, Borderline-SMOTE, SVM-SMOTE, ADASYN). ROC-AUC exhibits pronounced ceiling effects, yielding high scores even for underperforming models. In contrast, MCC and F2 align more closely with deployment-relevant costs and achieve the highest Kendall’s τ rank concordance across datasets; PR-AUC provides threshold-independent ranking, and H-measure integrates cost sensitivity. We quantify uncertainty and differences using stratified bootstrap confidence intervals, DeLong’s test for ROC-AUC, and Friedman–Nemenyi critical-difference diagrams, which collectively underscore the limited discriminative value of ROC-AUC in rare-event settings. The findings recommend a shift to a multi-metric evaluation framework: ROC-AUC should not be used as the primary metric in ultra-imbalanced settings; instead, MCC and F₂ are recommended as primary indicators, supplemented by PR-AUC and H-measure where ranking granularity and principled cost integration are required. This evidence encourages researchers and practitioners to move beyond sole reliance on ROC-AUC when evaluating classifiers in highly imbalanced data. Full article

Pediatric pneumonia detection faces the challenge of pathological asymmetry, where immature lung tissues present blurred boundaries and lesions exhibit extreme scale variations (e.g., small viral nodules vs. large bacterial consolidations). Conventional detectors often fail to address these imbalances. In this study, we propose YOLO-SMD, a detection framework built upon a symmetrical design philosophy to enforce balanced feature representation. We introduce three architectural innovations: (1) DySample (Content-Aware Upsampling): To address the blurred boundaries of pediatric lesions, this module replaces static interpolation with dynamic point sampling, effectively sharpening edge details that are typically smoothed out by standard upsamplers; (2) SAC2f (Cross-Dimensional Attention): To counteract background interference, this module enforces a symmetrical interaction between spatial and channel dimensions, allowing the model to suppress structural noise (e.g., rib overlaps) in low-contrast X-rays; (3) SDFM (Adaptive Gated Fusion): To resolve the extreme scale disparity, this unit employs a gated mechanism that symmetrically balances deep semantic features (crucial for large bacterial shapes) and shallow textural features (crucial for viral textures). Extensive experiments on a curated subset of 2611 images derived from the Chest X-ray Pneumonia Dataset demonstrate that YOLO-SMD achieves competitive performance with a focus on high sensitivity, attaining a Recall of 86.1% and an mAP@0.5 of 84.3%, thereby outperforming the state-of-the-art YOLOv12n by 2.4% in Recall under identical experimental conditions. The results validate that incorporating symmetry principles into feature modulation significantly enhances detection robustness in primary healthcare settings. Full article

Background: Metabolic syndrome (MetS) is a multifactorial condition characterized by insulin resistance, dyslipidemia, hypertension, and central obesity, and is strongly influenced by lifestyle factors. Growing evidence highlights the gut microbiota as a key mediator linking diet and physical exercise to cardiometabolic health. Objective: This narrative review aims to qualitatively synthesize current evidence on the effects of physical exercise and major dietary patterns including the Mediterranean diet (MedDiet), Dietary Approaches to Stop Hypertension (DASH), and ketogenic/very-low-calorie ketogenic diets (KD/VLCKD) on gut microbiota composition and function, and their implications for metabolic health in MetS. Methods: A qualitative narrative synthesis of experimental, observational, and interventional human and animal studies was performed. The reviewed literature examined associations between structured physical exercise or dietary interventions and changes in gut microbiota diversity, key bacterial taxa, microbial metabolites, and cardiometabolic outcomes. Considerable heterogeneity across studies was noted, including differences in populations, intervention duration and intensity, dietary composition, and microbiota assessment methodologies. Results: Across human interventional studies, moderate-intensity physical exercise was most consistently associated with increased gut microbial diversity and enrichment of short-chain fatty acid (SCFA)-producing taxa, contributing to improved insulin sensitivity and reduced inflammation. MedDiet and DASH were generally linked to favorable microbiota profiles, including increased abundance of Faecalibacterium prausnitzii, Akkermansia muciniphila, and Bifidobacterium, alongside reductions in pro-inflammatory metabolites such as lipopolysaccharides and trimethylamine N-oxide. In contrast, KD and VLCKD were associated with rapid weight loss and glycemic improvements but frequently accompanied by reductions in SCFA-producing bacteria, depletion of Bifidobacterium, and markers of impaired gut barrier integrity, raising concerns regarding long-term microbiota resilience. Conclusions: Lifestyle-based interventions exert diet- and exercise-specific effects on the gut microbiota–metabolism axis. While MedDiet, DASH, and regular moderate physical activity appear to promote sustainable microbiota-mediated cardiometabolic benefits, ketogenic approaches require careful personalization, limited duration, and medical supervision. These findings support the integration of dietary quality, exercise prescription, and individual microbiota responsiveness into translational lifestyle strategies for MetS prevention and management. Full article

Mangrove ecosystems are extremely sensitive to compounded stress, as evidenced by the widespread degradation and mortality of the pioneer mangrove species Avicennia marina along the Guangxi coast in recent years. However, research on how mangrove ecosystems respond to compound biotic stressors remains limited. Therefore, the present study aimed to systematically examine the ecological response mechanisms of A. marina under dual threats from the burrowing isopod Sphaeroma terebrans and the defoliating moth Hyblaea puera. Two contrasting sites were selected: Guchengling (subject to chronic stem-boring and sudden defoliator outbreaks) and Tieshangang (free from compounded stress). Photosynthetic capacity, metabolic function, and root structural integrity were all compromised considerably by chronic boring stress. During insect outbreaks, 15.33 ha of mangroves were destroyed due to impairments that breached the ecological threshold. In contrast, the healthier Tieshangang community exhibited strong ecological resilience, with rapid green canopy regeneration following defoliation and notable recovery in the normalized difference vegetation index. To enable early identification and precise intervention in mangrove decline, a comprehensive health index model was developed that includes root–canopy coordination, root length, and boring density. Field validation results, showing 100% agreement with expert evaluations across 19 validation sites (Cohen’s κ = 1.0), confirmed the high accuracy of the model. This study highlights the importance of identifying sensitive zones and undertaking timely ecological restoration, thereby providing a scientific basis and a practical tool that could facilitate early warning and timely management of mangrove degradation events. Full article

Immunotherapy has become a promising treatment for gastric cancer. However, its effectiveness varies significantly across subtypes because of heterogeneous immune microenvironments and genomic alterations. Here, we established Immune&Driver molecular subtypes CS1 and CS2 by systematically integrating multi-omics data for immune-related and driver genes. CS1 was linked to a better prognosis, while CS2 represented a poorer prognostic phenotype. CS1 displayed enhanced genomic instability, marked by higher mutation frequency and chromosomal alterations. In contrast, CS2 exhibited higher immune activity, with a higher density of immune cell infiltration and increased expression of chemokines and immune checkpoint genes. Among FDA-approved anti-cancer agents included in a pan-cancer drug sensitivity prediction framework, CS1 was predicted to be more sensitive to conventional chemotherapeutic agents, whereas CS2 was predicted to be more responsive to immune-related agents. In melanoma datasets, a CS2-like transcriptomic pattern was associated with improved response to anti-PD-1 therapy, with the combination of anti-PD-1 and anti-CTLA-4 showing more favorable response patterns compared to anti-PD-1 monotherapy. Additionally, we developed an immunotherapy response prediction model using PCA-based logistic regression according to the transcriptional expression of CS biomarkers. The model was trained in melanoma immunotherapy cohorts and validated across independent melanoma datasets, and it further achieved a higher AUC in an external gastric cancer cohort treated with anti-PD-1 therapy. Collectively, this study highlights immune and genomic heterogeneity in gastric cancer and provides a hypothesis-generating framework for exploring immunotherapy response. Full article