Search Results (5,465)

Object detection in unmanned aerial vehicle (UAV) imagery remains a crucial yet challenging task due to complex backgrounds, large scale variations, and the prevalence of small objects. Visible-spectrum images lack robustness under all-weather and all-illumination conditions; by contrast, multispectral sensing provides complementary cues (e.g., thermal signatures) that improve detection robustness. However, existing multispectral solutions often incur high computational costs and are therefore difficult to deploy on resource-constrained UAV platforms. To address these issues, SG-YOLO is proposed, a lightweight and efficient multispectral object detection framework that aims to balance accuracy and efficiency. First, a Spectral Gated Downsampling Stem (SGDS) is designed, in which grouped convolutions and a gating mechanism are employed at the early stage of the network to extract band-specific features, thereby maximizing spectral complementarity while minimizing redundancy. Second, a Spectral–Spatial Iterative Attention Fusion (SSIAF) module is introduced, in which spectral-wise (channel) attention and spatial-wise attention are iteratively coupled and cascaded in a multi-scale manner to jointly model cross-band dependencies and spatial saliency, thereby aggregating high-level semantic information while suppressing redundant spectral responses. Finally, a Spatial–Channel Synergistic Fusion (SCSF) module is designed to enhance multi-scale and cross-channel feature integration in the neck. Experiments on the MODA dataset show that SG-YOLOs achieves 72.4% mAP₅₀, outperforming the baseline by 3.2%. Moreover, compared with a range of mainstream one-stage detectors and multispectral detection methods, SG-YOLO delivers the best overall performance, providing an effective solution for UAV object detection while maintaining a favorable trade-off between model size and detection accuracy. Full article

This study provides a comparative evaluation of two Salvia species, the widely cultivated Salvia sclarea L. and the comparatively underexplored wild species Salvia pratensis L., integrating phytochemical profiling, chemical safety assessment, and biological activity investigation. Dried hydroethanolic extracts and essential oils obtained from aerial parts were analysed. HPLC–PDA analysis revealed distinct phenolic acid profiles, with S. sclarea characterized by higher levels of rosmarinic and protocatechuic acids, whereas S. pratensis contained greater amounts of hydroxycinnamic acids such as caffeic, p-coumaric, and ferulic acids. The total phenolic content was higher in S. pratensis (79.22 mg GAE/g dry extract) than in S. sclarea (52.50 mg GAE/g). GC–MS analysis showed that the essential oil of S. sclarea was dominated by oxygenated monoterpenes, mainly linalyl acetate and linalool, while S. pratensis exhibited a linalool-rich profile accompanied by sesquiterpene derivatives. Chemical safety assessment indicated minimal contamination, with pesticide residues detected only in S. sclarea at levels below regulatory limits and low concentrations of cadmium and lead in both species. The extracts showed strong antioxidant activity (DPPH IC₅₀ values of 6.67 µg/mL for S. sclarea and 3.16 µg/mL for S. pratensis) and moderate broad-spectrum antimicrobial activity (MIC 312.5–2500 µg/mL). In vitro assays on HEK 293 and HaCaT cells confirmed low cytotoxicity, with no evidence of membrane damage or pro-inflammatory effects. Overall, the results highlight the significant bioactive potential of the less studied S. pratensis, demonstrating that this wild species represents a promising alternative source of natural antioxidant and antimicrobial compounds comparable to the widely cultivated S. sclarea. Full article

Background/Objectives: Antimicrobial agents have revolutionized disease management in humans and animals; however, their misuse and overuse have accelerated the emergence and spread of antimicrobial resistance (AMR) and antimicrobial resistance genes (ARGs). Dairy farms are recognized as potential hotspots for ARG dissemination, particularly through Escherichia coli, which acts as a reservoir and vector of ARGs, enabling their horizontal transfer via plasmids and other mobile genetic elements. This study aimed to characterize the genomic diversity, ARG profiles, plasmid content, and phylogenetic relationships of E. coli isolated from dairy farm environments and milk using whole-genome sequencing. Methods: A total of 31 E. coli isolates recovered from soil, effluent, cow dung, and milk samples underwent deoxyribonucleic acid extraction, library preparation, and sequencing on the Illumina MiSeq platform, followed by comprehensive bioinformatic analysis. Results: The E. coli isolates exhibited 20 distinct sequence types, including one novel sequence type. Plasmids were detected in 71% of the isolates, with the IncF plasmid family being the most predominant. Furthermore, 12 ARG groups were identified, with β-lactam resistance genes detected in 67.7% of isolates. Notably, blaCTX-M genes were identified in all phenotypically confirmed extended-spectrum β-lactamase-producing isolates. Additional ARGs, including those conferring resistance to tetracyclines (tet(A), tetX4), quinolones (qnrS1), aminoglycosides (aph, aad, ant), and folate pathway inhibitors (dfr and sul), were widely distributed throughout the samples. Phylogenetic analysis revealed clustering of isolates from different sample types, particularly among ST58 isolates, suggesting cross-environmental transmission. Conclusions: This study demonstrates that E. coli from dairy farm environments harbor diverse ARGs and plasmids, confirming their role as reservoirs of AMR. These findings underscore the importance of prudent antimicrobial use, routine genomic surveillance, and enhanced biosecurity measures to limit cross-environmental transmission. Full article

Karst water systems are highly vulnerable to land use pressures, requiring integrated assessments to support conservation and management. This study evaluated the physicochemical, microbiological, and pesticide-related water quality in the Environmental Protection Area Nascentes do Rio Vermelho (APANRV), a karst conservation unit in the Brazilian Cerrado. Sixteen sampling sites (rivers, springs, and cave waters) were monitored during the dry (May 2024) and rainy (October 2024) seasons. Analyses included nutrients, major ions, Escherichia coli, and a broad spectrum of pesticides. The results showed marked spatial and seasonal variability, with elevated hardness and conductivity in karst areas due to carbonate dissolution. Nitrate and total phosphorus reached peak values of 13.59 and 0.132 mg L⁻¹, respectively, indicating localized nutrient enrichment. E. coli concentrations reached ≥2419.6 MPN 100 mL⁻¹, exceeding regulatory limits, particularly during the rainy season at recreational cave sites. Pesticides were detected in both seasons, with 11 compounds in the dry season and 8 in the rainy season, including atrazine degradation products, and maximum quantified concentrations up to 1.8 µg L⁻¹ (acephate). These findings highlight the combined influence of geology, seasonality, and land use on karst water quality and reinforce the need for continuous monitoring and targeted management strategies. Full article

This paper aims to present a non-destructive, optimized variational mode decomposition (VMD)-based ground-penetrating radar (GPR) method developed for identifying void defects in reinforced concrete (RC) structures. This study also presents an enhanced framework for defect detection in RC by integrating advanced spectrum analysis with deep learning techniques. A GPR investigation was conducted on an RC bridge deck with known structural defects to generate a representative dataset reflecting both intact and void-defective conditions. In addition to conventional spectral techniques such as fast Fourier transform (FFT), spectrogram, and scalogram, an optimized variational mode decomposition (VMD) method was implemented. The VMD approach decomposes GPR signals into intrinsic mode functions, enabling refined feature extraction beyond traditional spectral methods and allowing clear differentiation between intact and defective signals. The limited availability and quality of GPR small datasets have restricted the application of a functional 1D-CNN which generally requires at least several hundred datasets. To address this challenge, a data augmentation strategy is adopted. FFT-based features were successfully utilized to train a one-dimensional convolutional neural network (1D-CNN) for automated defect identification. The results demonstrate that both the advanced spectrum-based approach and the hybrid framework combining spectral analysis with deep learning significantly improve defect detection performance. Overall, the proposed methodology provides an effective and intelligent solution to support timely, data-driven decision-making for maintenance and safety assurance of bridge infrastructure. Full article

Background/Objectives: Antimicrobial resistance (AMR) in companion animals is an emerging public health threat due to zoonotic potential and limited therapeutic options. Dogs with otitis externa may harbor multidrug-resistant (MDR) bacteria, including Escherichia coli (E. coli), Proteus mirabilis (P. mirabilis), and Enterobacter cloacae complex (E. cloacae complex), some producing extended-spectrum beta-lactamase (ESBL) or AmpC β-lactamases. This study aimed to assess the prevalence, AMR patterns, MDR occurrence, β-lactamase production, and co-infection profiles of these pathogens in canine otitis externa. Methods: Ear canal samples were collected from 592 dogs presenting clinical signs of otitis externa, with one sample per dog included in the analysis. Samples were collected from veterinary clinics in Timiș County, Romania, from 2022 to 2025. Samples were cultured on blood agar and MacConkey agar, followed by biochemical testing and MALDI-TOF mass spectrometry for bacterial identification. Antimicrobial susceptibility testing against 15 agents across six classes was performed using the VITEK^® 2 system. MDR and β-lactamase production (ESBL, AmpC) were determined according to CLSI 2018 veterinary guidelines. Co-isolation with bacterial and fungal species were recorded. Results: E. coli, P. mirabilis, and E. cloacae complex were isolated in 9.12%, 6.25%, and 1.2% of cases, respectively. E. coli exhibited the highest resistance to aminoglycosides (tobramycin 72.2%, gentamicin 61.1%) and full susceptibility to carbapenems. P. mirabilis showed the highest resistance to ampicillin (54%) and trimethoprim + sulfamethoxazole (46%), with complete susceptibility to carbapenems and fluoroquinolones. E. cloacae complex displayed universal resistance to cephalosporins but remained susceptible to non-cephalosporin β-lactams (piperacillin–tazobactam), carbapenems and aminoglycosides. MDR prevalence was 35.2% for E. coli, 18.9% for P. mirabilis, and 14.3% for the E. cloacae complex. ESBL production was detected in 13% of E. coli and 8.1% of P. mirabilis isolates, while all E. cloacae complex isolates were AmpC-positive. Co-isolations were common, primarily involving Staphylococcus pseudintermedius (S. pseudintermedius) and Malassezia pachydermatis (M. pachydermatis). Conclusions: MDR and β-lactamase-producing bacteria were identified in dogs with otitis externa, emphasizing the importance of routine antimicrobial susceptibility testing, targeted therapy based on local resistance profiles, and continuous AMR surveillance to prevent treatment failure and mitigate zoonotic risk. Full article

Dynamic Spectrum Access (DSA) has emerged as a viable solution to address spectrum scarcity in shared-spectrum networks. In response, the FCC established the Citizens Broadband Radio Service (CBRS) to manage and facilitate shared use of the federal and non-federal spectrum in a three-tiered access and authorization framework. However, due to the open nature of spectrum access and the usually limited coverage of the monitoring infrastructure, enforcing access rights in a shared-spectrum network becomes a daunting challenge. In this paper, we stipulate the use of crowdsourcing as a viable approach to engaging volunteers in spectrum monitoring in order to enforce spectrum access rights robustly and reliably. The success of this approach, however, hinges strongly on ensuring that spectrum access enforcement is carried out by reliable and trustworthy volunteers within the monitored area. To this end, a hybrid spectrum monitoring framework is proposed, which relies on opportunistically recruiting volunteers to augment the otherwise limited infrastructure of trusted devices. Although a volunteer’s participation has the potential to enhance monitoring significantly, their mobility may become problematic in ensuring reliable coverage of the monitored spectrum area. To ensure continued monitoring, inspite of volunteer mobility, deep learning-based models are used to predict the likelihood that a volunteer will be available within the monitoring area. Three models, namely LSTM, GRU, and Transformer, are explored to assess their feasibility and viability to predict a volunteer’s availability likelihood over an extended time interval, in a given spectrum monitoring area. Recurrent Neural Networks (RNNs) such as GRU and LSTM are effective for tasks involving sequential data, where both spatial and temporal patterns matter, which is the focus of volunteer availability prediction in spectrum monitoring. Transformers, on the other hand, excel at handling long range dependencies and contextual understanding. Furthermore, their parallel processing capabilities allows faster training and inference compared to RNN-based models like GRU and LSTM. A simulation-based study is developed to assess the performance of these models, and carry out a comparative analysis of their ability to predict volunteers’ availability to monitor the spectrum reliably. To this end, a real-world trace dataset of volunteers’ location, collected over five years, is used. The simulation results show that the three models achieve high prediction accuracy of volunteers’ availability, ranging from 0.82 to 0.92. The results also show that a GRU-based model outperforms LSTM and Transformer-based models, in terms of accuracy, Root Mean Square Error (RMSE), geodesic distance, and execution time. Full article

Original high-definition radar data contains rich environmental information, including distance, Doppler velocity, and azimuth. However, extracting robust features from such sparse and noisy frequency-domain data remains a challenge. To address this issue, this paper proposes an improved multi-task network, the Efficient Feature Aggregation with Balanced Attention Radar Network (EFA-RadNet). This network introduces the VoVNetV2 architecture into the field of raw radar perception and effectively preserves feature diversity across different receptive fields through a One-Shot Aggregation (OSA) module, avoiding signal aliasing. In addition, we propose an attention mechanism module, Balanced effective Squeeze–Excitation (B-eSE), which is better suited for sparse radar processing and effectively addresses the problem of weak target loss in the radar spectrum. Experiments on the RADIal dataset show that our EFA-RadNet achieves excellent target detection performance while also attaining optimal accuracy in free space segmentation. Full article

Background: Antimicrobial-resistant Escherichia coli and Klebsiella pneumoniae represent an increasing challenge in community-acquired pediatric diarrheal infections. Understanding the genomic basis and dissemination of resistance in outpatient settings is essential for guiding antimicrobial use. Methods: Eighteen Gram-negative isolates obtained from pediatric outpatients with acute diarrhea were analyzed using selective culture methods, antimicrobial susceptibility testing, and whole-genome sequencing. Multilocus sequence typing, serotyping, virulence profiling, antimicrobial resistance gene detection, plasmid replicon typing, mobile genetic element analysis, and core genome-based phylogenetic analysis were performed. Phenotypic resistance profiles were correlated with genomic resistance determinants. Results: Klebsiella pneumoniae (55.56%) and Escherichia coli (44.44%) were identified, with all isolates exhibiting putative multidrug resistance-associated genomic profiles. Extended-spectrum β-lactamase genes, particularly blaCTX-M variants, were strongly associated with resistance to third-generation cephalosporins. In contrast, fluoroquinolone resistance correlated with gyrA and parC mutations and plasmid-mediated qnr genes. Phylogenetic analysis revealed diverse lineages harboring resistance determinants. In silico plasmid analysis revealed that key resistance genes co-occurred with IncF-type plasmids and mobile genetic elements, including ISEcp1, IS26, and class 1 integrons, suggesting putative plasmid association rather than confirmed localization. Conclusions: These findings highlight the small scale of plasmid-mediated antimicrobial resistance among E. coli and K. pneumoniae causing pediatric community-acquired diarrhea. The integration of phenotypic and genomic analyses underscores the need for continuous resistance surveillance to support rational antibiotic use in outpatient settings. Full article

Insects underpin key ecosystem services. Yet tree-associated insect communities remain comparatively poorly documented, particularly in temperate forests. This study aimed to characterize the diversity and abundance of insect assemblages associated with the predominantly insect-pollinated forest tree Tilia cordata Mill. in protected areas in Lithuania, and to assess the occurrence of known and putative pollinator groups within these assemblages. We quantified insect assemblages associated with Tilia cordata using two sampling methods but did not directly measure pollination effectiveness (e.g., pollen loads, visitation rates to flowers, or fruit/seed set). Consequently, our inferences refer to the presence and composition of potential pollinators rather than demonstrated pollination function or realized pollination services. Fieldwork was conducted over two years in six protected T. cordata sites in Lithuania using two complementary sampling methods: net sampling and sticky traps. Sampling was structured into three observation periods corresponding to T. cordata phenology: pre-flowering (I), flowering (II) and post-flowering (III). In total, 207 insect taxa from 15 orders were recorded by net sampling and 86 taxa from 11 orders by sticky traps. Net sampling showed significantly higher diversity (Shannon H = 3.81) than sticky traps (H = 2.10). Hemiptera, Coleoptera and Diptera were the most common groups, and most taxa occurred at low to moderate abundances, with only a few species showing local dominance in specific periods or sites. Taxa documented in the literature as significant pollinators were consistently present but at low relative abundances (about 5–10% in total). Insect assemblage composition and species proportions varied among phenological periods and between years, with no clear, consistent peak in overall insect abundance or diversity associated specifically with the T. cordata flowering phase. These findings indicate that T. cordata stands in protected areas harbor diverse insect assemblages typical of temperate deciduous and mixed forest habitats and include a broad spectrum of non-bees and other potential pollinators. Therefore, we did not detect a distinct peak in insect abundance or species richness during the T. cordata flowering period, indicating that flowering did not coincide with a pronounced maximum in pollinator-related insect activity. However, the quantitative patterns observed suggest that, in this context, T. cordata provides only modest support for pollinator guilds, and its role is better interpreted as one component of wider forest insect diversity rather than as a primary driver of pollination services. Full article

Background/Objectives: Autism spectrum disorder (ASD) is a clinically and genetically heterogeneous neurodevelopmental condition, and the diagnostic yield of whole-exome sequencing (WES) varies across settings. This single-center study aimed to determine the molecular diagnostic yield of WES in pediatric ASD and to explore genotype–phenotype overlap using a structured, phenotype-driven reanalysis strategy. Methods: We enrolled 60 children with syndromic and non-syndromic ASD, who underwent detailed clinical and dysmorphology assessment. WES for single-nucleotide and copy-number variant (CNV) detection was performed in an accredited laboratory, followed by clinician-driven reinterpretation, integrating expanded phenotypic data and ACMG/AMP-based variant classification. Genes were considered if they harbored rare, potentially pathogenic variants and were previously reported or curated in established ASD-associated gene resources. Results: The initial external laboratory report identified 5 of 60 patients (8.3%) with a pathogenic (P) or likely pathogenic (LP) variant (positive result), 30 of 60 (50.0%) with a variant of unknown significance (VUS) (inconclusive result), and 25 of 60 (41.7%) with a negative result. Clinician-based variant reinterpretation identified pathogenic or likely pathogenic variants in 9 of 60 patients (15.0%), representing an 80% relative increase in diagnostic yield, as well as 43 VUSs distributed across 34 patients, while 17 patients had no reportable variants (negative result). Overall, reanalysis revealed 11 additional variants of interest (pathogenic, likely pathogenic, or VUS) that had not been reported in the initial assessment. In total, 52 sequence and copy-number variants in 46 genes were detected, most of which were VUSs (83%). Conclusions: In this pediatric ASD cohort, WES with phenotype-driven reinterpretation and CNV assessment yielded a clinically positive result in 15% of patients and uncovered additional candidate variants, highlighting both the value and the current interpretative challenge of comprehensive genomic testing in ASD. Full article

The remarkable evolution of oncological therapies has dramatically improved cancer survival rates but has simultaneously introduced a significant burden of cardiovascular complications. Cardio-oncology has emerged as a critical multidisciplinary field focused on mitigating the “collateral damage” of life-saving anticancer treatments, ranging from traditional chemotherapeutics to novel immunotherapies. This review provides a comprehensive analysis of the pathophysiological mechanisms, clinical phenotypes, and evolving management strategies for cancer therapy-related cardiac dysfunction (CTRCD). An extensive synthesis of the current literature was conducted, focusing on the molecular pathways of cardiotoxicity, including Topoisomerase IIβ inhibition by anthracyclines, HER2 signaling disruption by targeted agents, and immune-mediated myocarditis triggered by checkpoint inhibitors (ICIs). Cardiotoxicity is increasingly recognized as a spectrum of phenotypes. Heart failure with reduced ejection fraction (HFrEF) remains a primary concern with cytotoxic agents, while heart failure with preserved ejection fraction (HFpEF) is emerging as a critical complication of radiation therapy and tyrosine kinase inhibitors (TKIs). The integration of advanced diagnostic tools—specifically Global Longitudinal Strain (GLS) and Cardiac Magnetic Resonance (CMR) mapping—has shifted the clinical focus toward subclinical detection. Furthermore, pivotal clinical trials such as PRADA and SUCCOUR have validated early pharmacological prophylaxis and strain-guided interventions. Emerging challenges, including the management of CAR-T cell-induced cytokine release syndrome and the specific cardiovascular needs of pediatric and geriatric populations, are also explored. The future of cardio-oncology lies in precision medicine, leveraging genomic profiling and artificial intelligence to identify high-risk individuals. A proactive, multidisciplinary approach is essential to ensure that the success of modern oncology is not compromised by irreversible cardiovascular morbidity. Full article

Cardiac sarcoidosis (CS) is a critical and frequently underdiagnosed phenotype of sarcoidosis, characterized by non-caseating granulomatous infiltration of the myocardium. This review synthesizes current knowledge regarding the pathogenesis, diagnosis, and management of CS. The disease manifests with a heterogeneous clinical spectrum ranging from asymptomatic conduction abnormalities to life-threatening ventricular arrhythmias and heart failure. Diagnosis remains challenging due to the patchy distribution of granulomas, which limits the sensitivity of endomyocardial biopsy. Consequently, a multimodal diagnostic approach is essential, integrating advanced imaging modalities such as cardiac magnetic resonance (CMR) with late gadolinium enhancement (LGE) and ¹⁸F-fluorodeoxyglucose positron emission tomography (FDG-PET). These tools not only facilitate detection but also enable the differentiation of active inflammation from chronic fibrosis. Histopathological assessment, supported by specific immunophenotyping and electron microscopy, remains the gold standard for confirming diagnosis and excluding mimics like giant cell myocarditis or infectious granulomatous diseases. Management requires a multidisciplinary strategy combining immunosuppressive therapy, primarily corticosteroids and steroid-sparing agents, with guideline-directed cardiac care, including implantable cardioverter-defibrillators for arrhythmia risk stratification. Emerging biomarkers and artificial intelligence-driven imaging analysis promise to further refine risk stratification and therapeutic monitoring, advancing precision medicine in this complex disorder. Full article

We have used long observations of galaxy clusters obtained with the Rossi X-ray Timing Explorer to search for the 3.55 keV line from sterile neutrino decay. If a lepton-number asymmetry exists in one or more types of active neutrinos in the early Universe, sterile neutrinos can be produced via the Shi–Fuller mechanism. The data consist of 11 clusters observed for a total of 3.1 megaseconds using the Proportional Counter Array. A 2.5$\sigma$ excess of emission over a thermal model is found over the energy span of the 3.55 keV line in the combined spectra of the eight clusters that individually have an excess. These residuals are added to increase the signal to noise ratio of the excess, which is then modeled with a Gaussian to simulate the instrumental spectral response. We find a significant correlation (r = 0.76) for a line centered at 3.6 keV with a model flux of 3.07 × 10⁻⁵ ph cm⁻² s⁻¹. Mixing angle for detected clusters ranges from 2.0 to 21.6 × 10⁻¹⁰. The decay rate inferred from the line flux is strongly correlated (r = 0.87) with cluster temperature, which is due to hotter, more massive clusters having a larger amount of dark matter. Approximately half of the total flux comes from the Coma cluster. The mixing angle for Coma is calculated to be 6.2 × 10⁻¹⁰. We fit the Coma cluster spectrum with two different three-component models. The first includes a Gaussian fixed at 3.55 keV to model soft emission. The flux of the Gaussian is 5.6 × 10⁻¹² ph cm⁻² s⁻¹ or 1.3% of the total flux. The second three-component model uses a second thermal component to model soft emission. This model gives a temperature of 0–17 keV for the second thermal component and a lower temperature for the hot component. This indicates that the second thermal component is modeling high-energy residuals rather than low ones, where the Gaussian is. Though our line fluxes exceed most reported detections and upper limits, they do not overproduce the dark matter. We conclude that some fraction of the marginally detected excess could be attributed to the decay line since low-temperature thermal emission and systematics fail to model it completely. Full article

Sulfur dioxide is a trace constituent of the upper atmosphere of Venus but plays a dominant role in the photochemistry above the cloud tops. Because SO₂ undergoes indirect dissociation to a relatively long-lived excited state, it has a line-type absorption spectrum in the dissociation region (~190–220 nm). This leads to strong isotopic fractionation under optically thick conditions, a process referred to as self-shielding. Here, I use SO₂ cross-sections, shielding functions, and a simple steady-state photochemical model to estimate sulfur isotope ratios in SO₂. The results indicate that large isotope depletion relative to SO₂ in the deep atmosphere is expected in SO₂ below 70 km altitude, with δ³⁴S ~ −100 to −200 permil. This is readily detectable by the VTLS tunable laser spectrometer planned for the NASA DAVINCI mission. Full article