Abstract: This study evaluates the cost-effectiveness of the MAR-HF-Titration telemedicine programme, implemented at Hospital del Mar for patients with heart failure with reduced ejection fraction (HFrEF). A pre-post retrospective cost-effectiveness analysis was conducted using data from 208 patients enrolled between 2021 and 2023. Effectiveness was estimated by constructing individual pre- and post-intervention QALYs based on literature-derived utility values. Cost data were obtained from hospital administrative records at the patient level. Paired t-tests were performed, and costs and QALYs were modelled using gamma and beta distributions for probabilistic sensitivity analysis. A sensitivity analysis excluding newly diagnosed patients was also conducted. The programme was associated with a 52% reduction in total healthcare costs (€4,771.79 per patient), mainly driven by a reduction of the heart failure-related intervention costs, which accounted for 88% of total costs and decreased by 58% in the post-intervention period (€4,384.97 per patient). This reduction was also associated with a substantial decrease in length of stay during the post-intervention period, with a 76.63% reduction in total hospitalisations and an 82.34% reduction in hospitalisations related to heart failure. The effectiveness analysis showed an increase of 0.0298 QALYs per patient during the six-month intervention period. Similar outcomes were observed for non-newly diagnosed patients.

Abstract: Objective: Pupillometry is a noninvasive method which measures changes in the size of the pupil intending to study cognitive processes and identify networks that are related to specific attentional states such as attention, mental effort, working memory, and many others, through monitoring pupil size. This study aimed to investigate hotspots and global research trends in the field of pupillometry by bibliometric analysis.Methods: Publications from 1999 to 2025 in the Web of Science database were extracted and analyzed using bibliometric methods. This study utilized the R Bibliometrix package, along with the web platforms VOS Viewer and CiteSpace.Results: A total of 2,897 peer-reviewed documents were obtained, and the annual research output from 1999 to 2025 has increased significantly. The role of pupillometry in this research has been investigated by over 11,534 authors coming from 3,019 institutions in 89 countries or regions. According to the report, the United States had the highest number of publications, while Psychophysiology was the most productive journal. Amongst the most active academic institutions, the University of Toronto stood out as the most active, and among the most prolific authors, Olson Daiwai emerged as the most prolific. We found that keywords from pupillometry related including locus-coeruleus (304 occurrences), attention (240 occurrences), pupil (239 occurrences), pupillary light reflex (237 occurrences), and retinal ganglion cells (135 occurrences) had high frequency and among the strongest citation burst keywords, “infrared pupillometer”, “in situ keratomileusis”, “retinal ganglion cells”, and “refractive surgery” showed citation bursts from 1999 to 2025, corresponding with the increase in published documents.Conclusions: It can be noted that this study was the first comprehensive bibliometric study to summarize trends and developments in pupillometry research from a bibliometric perspective. By analyzing bibliometric data on development trends in pupillometry development trends, developers or researchers would be able to propose future research directions as well as pursue further collaborations.

Abstract: Object detection technologies form the foundation of real-time performance across a broad spectrum of applications, ranging from autonomous systems to medical imaging. This study analyzes the extensive architectural evolution of the YOLO series, the benchmark for this field, from its initial version to the cur rent YOLO26 model. Throughout the paper, structural transformations in the backbone, neck, and detection head components are examined chronologically. The review focuses on critical technical milestones, including the transition from anchor-based to anchor-free systems, the integration of attention mechanisms, and optimizations to loss functions. Furthermore, by evaluating the density of literature, data labeling tools, and the wide range of applications, the study exam ines the evolutionary logic of object detection architectures and the technological framework of modern models in comprehensive detail.

Abstract: Submarine gas emissions represent a key expression of fluid migration processes in volcanic and hydrothermal marine environments and provide valuable analogues for monitoring strategies relevant to sub-seabed carbon storage. This study investigates the feasibility of using marine Distributed Acoustic Sensing (DAS) to detect natural CO₂ bubble emissions in a shallow-water setting offshore Panarea (Aeolian Islands, Italy). A 1.1 km armored fiber-optic cable was deployed on the seabed and interrogated using two different DAS systems to acquire continuous passive acoustic data. The DAS recordings were complemented by controlled gas releases from scuba tanks to provide reference signals, as well as by independent high-resolution boomer seismic survey and side-scan sonar imaging to characterize the shallow subsurface and seabed morphology. The results show that DAS is sensitive to acoustic signals associated with both artificial and natural bubble emissions, despite the complex acoustic conditions typical of shallow marine environments. The integration of passive DAS monitoring with independent geophysical observations provides a robust framework for interpreting gas-related signals and seabed processes. These findings demonstrate that marine DAS represents a promising geophysical tool for monitoring of submarine volcanic–hydrothermal systems and offers important insights for the development of sub-seabed CO₂ leakage detection in offshore CCS contexts.

Abstract: Background/Objectives: Kidney injury is a frequent complication of multiple myeloma (MM) and monoclonal gammopathies. Podocyte stress markers, such as urinary nephrin and podocin, have been studied in other renal diseases but their utility in paraprotein-related kidney disease remains unclear. This pilot study investigated the association of urinary nephrin and podocin levels with albuminuria and biopsy-proven podocytopathy in patients paraprotein-related diseases. Methods: We retrospectively analyzed 75 patients with plasma cell dyscrasias including MM and MGRS and 11 healthy controls. Urinary podocin and nephrin mRNA levels were measured using qPCR, and urinary podocin protein levels were quantified via ELISA. Associations were assessed between these biomarkers and urinary pro-tein-to-creatinine ratio (uPCR), albumin-to-creatinine ratio (uACR), and histologically confirmed podocytopathia. Diagnostic performance was evaluated using Receiver Op-erating Characteristic (ROC) analysis. Results: Higher urinary podocin protein levels were significantly associated with lower uACR (p=0.007) and uPCR (p=0.026). Neither podocin nor nephrin mRNA showed sig-nificant associations with proteinuria metrics. ROC analysis indicated that podocin ELISA (AUC=0.350) and podocin mRNA (AUC=0.510) lacked diagnostic accuracy for predicting renal involvement. The presence of urinary tract infection (UTI) was a sig-nificant confounder, leading to increased levels of podocin and nephrin mRNA. Conclusions: Urinary podocin shows a trend toward elevation in MM/MGRS patients with histological podocyte injury. The study revealed an unexpected inverse association between urinary podocin and albuminuria, suggesting complex release kinetics or stage mismatches in this population. Given the confounding effect of UTIs and the pilot na-ture of this study, further research is required to validate these podocyte proteins as prognostic biomarkers in paraprotein-related kidney disease.

Abstract: Open-air asphalt milling generates hazardous respirable crystalline silica (RCS), posing severe risks to operators of legacy machines lacking enclosed cabs. This study evaluates a novel, standalone retrofit water spray system designed to intercept fugitive dust. Field validation across 11 road maintenance sites involved particle characterization and paired system-off/on exposure monitoring. Results indicated a Mass Median Aerodynamic Diameter (MMAD) of 6.12 µm, confirming the efficacy of fine-atomizing nozzles (0.3 mm) for capturing respirable fractions. The system achieved RCS suppression efficiencies ranging from 60% to over 85% under low-to-moderate wind conditions (<2.5 m/s). A comparative analysis revealed no significant performance gain from larger 0.5 mm nozzles, supporting the use of smaller orifices for optimal water conservation. However, suppression efficacy degraded significantly when crosswinds exceeded 2.5 m/s, identifying a critical operational threshold. This retrofit solution provides a scientifically validated, cost-effective engineering control for reducing occupational silica exposure in aging road maintenance fleets.

Abstract: Recent advances in artificial intelligence (AI), especially large language models, have accelerated the integration of multimodal data in scientific research. Given that scientific fields involve diverse data types, ranging from text and images to complex biological sequences, graphs, and structures, multimodal large language models (MLLMs) have emerged as powerful tools to bridge these modalities, enabling more comprehensive data analysis and intelligent decision-making. This work, S³-Bench, provides a comprehensive overview of recent advances in MLLMs, focusing on their diverse applications across science. We systematically review the progress of MLLMs in key scientific domains, including drug discovery, molecular & protein design, materials science, and genomics. The work highlights model architectures, domain-specific adaptations, benchmark datasets, and promising future directions. More importantly, we benchmarked open-source MLLMs on a range of critical molecular and protein property prediction tasks. Our work aims to serve as a valuable resource for both researchers and practitioners interested in the rapidly evolving landscape of multimodal AI for science.

Abstract: This study develops a human-centered AI framework enabling rapid ecodesign prioritization for ESPR compliance while demonstrating Large Language Model (LLM) integration in sustainability strategy. Four-stage hybrid methodology combining LLM-assisted action identification (30 ESPR-aligned interventions) with Multi-Criteria Decision Analysis with Analytic Hierarchy Process (MCDA-AHP) is developed. Expert validation addressed LLM-driven interventions limitations with practitioners evaluating AI suggestions based on value chain context. The framework applied to two Italian BIPV SMEs demonstrated strategic differentiation based on Feasibility vs. Desirability vs. Affordability producing systematically different action portfolios within regulation-aligned aggregate structures. Sensitivity analysis showed 100% priority stability under ±10% AHP variations for priority 1, 3 and 4 actions and 82% for priority 2 actions, validating framework robustness. The framework's provide empirical evidence for augmentation-not-automation in AI-assisted strategic planning, contributing replicable methodology for responsible LLM integration across manufacturing sectors. Results demonstrate combining AI synthesis efficiency with human contextual judgment enables regulation-aligned, business-model-specific sustainability strategies.

Abstract: We propose that the transition from reversible quantum correlation to an objective classical record is a thermodynamically irreversible process with a quantifiable heat signature. We formulate a three-stage taxonomy separating reversible premeasurement, irreversible record formation, and memory reset, and derive a conditional record-formation heat bound: under explicit operational conditions (C1 to C6) in the uncontrolled-decoherence regime (no work extraction/coherence-recovery channel), the irreversible record-formation channel must dissipate at least kB·T·ln(2) of heat per bit of classical information created, quantified by the mutual information I(X;Y) between a prepared classical label X and the recorded outcome Y. Using an explicit system/pointer/bath model, we identify the precise stage at which this Landauer cost is paid: not during unitary premeasurement coupling, but during irreversible environmental coupling, when the pointer becomes entangled with N ≫ 1 environmental degrees of freedom and the record is stabilized. We design a circuit-QED differential microcalorimetry experiment using superconducting qubits and nanocalorimeters (TES or SNS nanobolometer class). The protocol employs matched ON/OFF branches that share identical premeasurement pulses and routing losses, differing only in whether an objective record is stabilized. The measurand is the per-shot differential deposited energy ΔQ ≡ Q_ON − Q_OFF, which isolates the record-formation contribution. Four primary controls (ground-state baseline, measurement-strength scaling, reversal-delay timing sweep, and prior-variation) discriminate from systematic effects. Sensitivity analysis using demonstrated nanobolometer performance shows detection is feasible with N ~ 2×10⁹ to 8×10⁹ ON/OFF pairs at 10 mK for Landauer-scale residual tests at SNR ~ 10 (for σ_Q ≈ 0.32 to 0.6 zJ and ΔQ_target ≈ 9.57×10^-5 zJ). The bound is falsified if the observed residual ΔQ − k_B·T·ln(2)·I(X;Y) falls statistically below zero.

Abstract: Flow channels on the snow surface are a common phenomenon frequently reported by field observers. The interpretation of those field observations and an understanding of the underlying physical processes are important for forecasting routines and models of avalanche warning, hydrological, or meteorological services. Flow channels on the snow surface are typically associated with rain-on-snow (ROS) events and are often interpreted as an indicator of the approximate snowfall level. However, recent field observations of flow channels on the snow surface without significant liquid precipitation in the Austrian Alps challenge the assumption that ROS events are the sole cause of flow channel formation. In this study, we quantitatively compare liquid water input into the snowpack from melt processes to the amount of rain during a documented flow channel formation event. Using a combination of field observations, energy balance calculations and model simulations, we demonstrate that, in our case study, meltwater was the predominant driver of flow channel formation. Our results indicate that more than 97 % of the total liquid water input originated from melt, while rain contributed only roughly 2 %. These findings highlight the need for a revised interpretation of flow channel formation, suggesting that meltwater-driven flow channels may be more significant than previously assumed.

Abstract: An influential intuition about the biology of aging is that organisms are born with a fixed amount of “life stuff” that they burn through in the process of living. According to this mindset, the key to living longer is simply to burn through this life stuff as slowly as possible. We instead reinterpret this “life stuff” as Survival Capital: resources allocated to maintenance and growth that can compound over time to facilitate future reproduction. Contrasting with classical evolutionary theories that treat life-history expenditures as linear substitutes, we develop an explicit framework in which investments in survival and reproduction act as synergistic complements. We first show how early-life investments in survival traits generate complementary incentives for later-life investments, thereby increasing the value of survival into older ages. We then add multiplicative damage to survival, capturing biologically plausible risk accumulation without altering the qualitative logic of the first model, thereby reinforcing the effects of complementarity among survival investments. Next, we demonstrate that such complementarities naturally imply bi-stability of life-history strategies when trade-offs between survival and reproduction across discrete life stages are considered. Finally, we show that when investments can be made to affect multiple components of health and interact synergistically, they can produce runaway increases in health investment, driving extreme healthspans. This moves beyond the classical “fast–slow” continuum by making the value of later life endogenous. Linking these allocation trade-offs to existing evidence concerning real world counterparts of Survival Capital generates powerful empirical predictions for both evolutionary biology and synergistic geroscience.

Abstract: Research has identified how heritage elicits emotional responses that significantly affect the tourist destination image. In this vein, this study explores how people emotionally re-act to architectural heritage by examining five representative styles in Spain—Gothic, Re-naissance, Baroque, Modernism, and Contemporary. In an experimental design, partici-pants observed visual stimuli depicting seven types of urban infrastructure, such as buildings, streets, bridges, façades, quarters, squares, and churches, and their emotional responses were recorded using automated facial expression recognition. The system clas-sified eight core emotions: neutral, happiness, sadness, surprise, fear, disgust, anger, and contempt. Gender differences were also analyzed to identify possible variations in emo-tional activation. The results show that architectural forms—whether historical or con-temporary—evoke distinct emotional patterns that help explain how people perceive and value historical heritage. By bringing together concepts from heritage tourism, environ-mental psychology, and user experience studies, this research highlights the relevance of emotions as a key element in understanding the relationship between people and the ar-chitectural spaces they inhabit or visit.

Abstract: Autumn irrigation is a key agricultural practice during the non-growing season in the Hetao Irrigation District, and its spatial distribution and dynamic monitoring are essen-tial for refined water resource management and for salinity control. Taking the Yichang Irrigation District in Hetao as the study area, this study integrated multi-source optical remote sensing data from GF-1/GF-6, HJ-2A/B, Landsat 8/9, and Sentinel-2, acquired dur-ing the autumn irrigation period (October–November 2024). The spectral characteristics of croplands in the non-growing season were analyzed, and a soil moisture index was con-structed to retrieve the extent of autumn irrigation and reveal its spatiotemporal evolution at both the district and irrigation unit scales. The results indicate that cropland spectral responses in the non-growing season can be classified into three types: non-irrigated fields, post-irrigation fields with a surface-water layer, and post-irrigation fields after sur-face-water infiltration. The visible–red-edge–near-infrared (450–890 nm) and shortwave infrared (1565–2290 nm) regions are highly sensitive to soil moisture variations. Among the tested indices, the soil moisture index based on the near-infrared–shortwave infrared band combination exhibited the most consistent performance and highest accuracy in identifying irrigated areas. Newly irrigated areas characterized the irrigation rate, whereas the cumulative irrigated area reflected the actual coverage. Together, these indi-cators delineate the four phases of autumn irrigation: initiation, acceleration, stabilization, and termination. The coefficient of determination (R²) between the remotely sensed cumu-lative irrigated area and statistical records was 0.95. The final cumulative au-tumn-irrigated area was 1040.76 km², compared with 831.4 km² reported in the statistical data, satisfying the accuracy requirement for large-scale irrigation monitoring. At the irri-gation-unit scale, the linear fit for the newly irrigated area yielded R² = 0.75, effectively capturing incremental changes across different phases. The spatiotemporal evolution ex-hibited clear regularity. Temporally, the entire district follows a four-phase pattern, and the similarity between unit- and district-scale processes increases with the ca-nal-controlled area. Spatially, early stage irrigation was concentrated along the Fengji, Shahe, Yihe, and Tongji main canals, forming an east–west gradient with greater intensity in the east. As irrigation progresses, irrigated areas expand along the hierarchy of the main, primary, branch, and lateral canals, gradually forming contiguous patches and ul-timately an integrated district-wide pattern. Irrigation remains relatively sparse along the northern margin and in the southwest, whereas large canal units, such as Fengji, Zaohuo, Tongji, and Yihe, dominate the spatial evolution. This multi-source remote sensing ap-proach provides a scientific basis for optimizing autumn irrigation water allocation, im-proving irrigation scheduling, and enhancing water-use efficiency in the Hetao Irrigation District.

Abstract: With the increasing importance of securing quantum communication networks, practical and robust entity authentication is a critical requirement. Accordingly, we propose and experimentally validate a quantum entity authentication protocol specifically designed for integration with BB84-type QKD workflows and existing terminal architectures. We analyze the protocol's security against intercept–resend man-in-the-middle impersonation, showing that an unauthenticated adversary induces a characteristic 25% correlation error and that the rejection probability approaches unity as the number of detected authentication events increases. For practical realization, the protocol is deployed using weak coherent pulses with decoy-state estimation to bound single-photon contributions and mitigate photon-number-splitting–enabled leakage. The system is demonstrated over a field-deployed fiber link of approximately 20 km with ~8 dB optical loss using signal/decoy intensities of ~0.5/~0.15 and sending probabilities 0.88/0.10/0.02 (signal/decoy/vacuum). Across both verification directions, stable operation is observed with QBER typically fluctuating between 1% and 4% while the sifted key rate remains constant over time. These results provide an experimental basis for integrating physical-layer entity authentication into deployed quantum communication networks.

Abstract: Humanity is facing an unprecedented socio-ecological and climate crisis resulting from human impact on the planet, which requires a profound transformation in how we inhabit and develop our territories. Regenerative development is emerging as a key approach to strengthening living systems and improving environmental health. In this context, UNESCO Biosphere Reserves are consolidating their role as strategic instruments that link biodiversity conservation with sustainable development through integrated and par-ticipatory land management models. Mexico stands out for its regional and global leadership in implementing these areas. Participatory governance, promoted by the MAB program, encourages the active involvement of local communities. This article analyses the application of a regenerative and participatory design methodology in a Biosphere Reserve, evaluating both the process and the tools used. Beyond the fulfilment of sus-tainability objectives, it examines the lessons learned, results and scope from a regen-erative perspective, providing critical reflections on its effectiveness as a strategy for the socio-ecological management of vulnerable territories.

Abstract: The gut microbiota–brain axis has emerged as a key regulator of central nervous system (CNS) development and function. Beyond its established roles in immunity and behavior, growing evidence indicates that microbial-derived metabolites influence glial biology, including oligodendrocyte development and myelination. In this review, we synthesize current findings supporting a role for microbial metabolites—particularly short-chain fatty acids (SCFAs), tryptophan-derived indoles, and secondary bile acids—in regulating oligodendrocyte precursor cell (OPC) proliferation, differentiation, and myelin formation. We examine the molecular and cellular mechanisms involved, including epigenetic regulation through histone deacetylase inhibition, G-protein–coupled receptor signaling, nuclear receptor activation, and metabolic support pathways. Experimental and clinical studies indicate that butyrate modulates oligodendrocyte lineage progression via chromatin remodeling, indole derivatives signal through the aryl hydrocarbon receptor (AhR) to attenuate neuroinflammation and promote glial homeostasis, and secondary bile acids act through receptors such as TGR5 and the vitamin D receptor to support pro-remyelinating environments. Additionally, microbial acetate contributes to acetyl-CoA pools necessary for lipid biosynthesis and epigenetic regulation within oligodendrocytes. Notably, dysbiosis and reduced availability of microbial metabolites have been consistently reported in Multiple Sclerosis (MS), a demyelinating disorder characterized by impaired remyelination, suggesting that microbiome-targeted strategies—including dietary interventions and postbiotic supplementation—may complement existing therapies. Collectively, current evidence supports the concept that microbial metabolites function as systemic modulators of CNS myelination, linking dietary and microbial inputs to oligodendrocyte biology and highlighting promising avenues for translational research in demyelinating diseases.

Abstract: Purpose: This study aims to explore how sports businesses can be reimagined in the digital era by identifying key transformative indicators shaped by virtualization. Design/methodology/approach: The research is of mixed type (qualitative-quantitative). The statistical population of the qualitative section included all IT specialists in the sports industry and university professors with expertise in sports management and digitalization from executive and academic fields. Purposive sampling method was used to select the research samples. The sample size continued until theoretical saturation (15 people) was reached. In the initial stage, the data of this research were collected through semi-structured interviews with the selected samples. To ensure the validity, various methods such as tracking data over time (audit tracking) and consolidating the findings through repeated assessments of data and patterns were used. After the coding processes and identification of main themes and indicators, a questionnaire was developed to validate the extracted components. For qualitative data, content analysis was used, and in the quantitative phase, second-order confirmatory factor analysis was used using SPSS version 25 and Smart PLS version 3 software. Findings: Based on the findings, the foundations for developing digital business included business virtualization contexts, assumptions of the digital workplace in business; digital business processes included the efficiency of the virtual business system and business virtualization processes; digital business functions encompassed the development of digital business and the sustainability of the digital work environment. Also, the results of the validity and reliability analyses and confirmatory factor analysis of the research components indicated the appropriate validity of all items of the questionnaire obtained from the qualitative section. The examination of the indicators of the final research model also indicated the high validity of the model. According to other results, the desired model had appropriate fit indicators. This research provides a scientific and practical basis for policymakers and managers in the sports industry to better focus on advancing the virtualization and digitalization of sport businesses in their decision-making processes. Originality: This study advances a holistic framework for understanding the virtualization of sports businesses. Unlike prior work, it integrates technological, economic, and social dimensions, introducing new indicators for innovation and sustainable value creation.

Abstract: Cytochrome P450 (CYP) 2B7 is currently classified as a pseudogene due to the presence of a premature stop codon in exon 7 that is predicted to result in a truncated protein. We hypothesized that in some individuals, CYP2B7 may code for a full length protein with enzyme activity similar to CYP2B6, as the result of a DNA sequence polymorphism in the CYP2B7 premature stop codon. Genomic DNA was prepared from our human liver bank (2 Hispanic, 4 African-Americans, and 47 Caucasians) and blood obtained from 20 African-Americans undergoing an unrelated clinical study. PCR was used to amplify and sequence the genomic region surrounding the CYP2B7 premature stop codon. All European-American and Hispanic DNA showed no variation of the stop codon sequence (TGA). In 9 of 24 African Americans, the first nucleotide of the stop codon was heterozygous (T/C), and in one of 24 African-American this nucleotide was homozygous (C/C) resulting in an arginine (codon CGA) instead of the stop codon. We next cloned the CYP2B7-R378 full-length cDNA by reverse transcription PCR from human liver mRNA. However, in addition to the CYP2B7-R378 cDNA, we also cloned a chimeric CYP2B6/CYP2B7-R378, which like CYP2B7-R378 was also predicted to code for a full-length protein. Both CYP2B7-R378 and CYP2B6 amino acid sequences could be aligned to the rabbit CYP2B4 crystal structure indicating that CYP2B7-R378 contained no mutations that would disrupt protein structure. Expression of cDNAs encoding CYP2B7-R378, CYP2B6/CYP2B7-R378, and CYP2B6 in 3T3 fibroblast cells showed immunodetectable CYP2B protein for each, but minimal (CYP2B6) or no bupropion hydroxylation activity. However, by use of the baculovirus-insect cell expression system we successfully showed bupropion hydroxylation activity in cells infected with CYP2B7-R378 and CYP2B6, but not in negative controls (including CYP2B7-X378). No activity was found for chimeric CYP2B6/CYP2B7-R378, although this result needs to be substantiated by further study. Expression of CYP2B7 mRNA was also demonstrated in most human tissues that also contained CYP2B6 mRNA, with particularly high expression in lung, liver and intestine. In conclusion, results indicate that CYP2B7 is expressed as an active cytochrome P450 in some individuals, particularly those of African-American ancestry. The presence of this polymorphic enzyme may explain some important differences in drug response and carcinogen activation related to race and/or ethnicity.

Abstract: The railway sector plays a vital role in passenger transportation, economic growth, and large-scale urban development. Despite continuous improvements in signaling and control systems, railway safety remains a major concern due to frequent incidents caused by track obstructions such as unauthorized human intrusions, animals, vehicles, and miscellaneous foreign objects. Existing railway monitoring systems rely heavily on manual inspection, physical barriers, and rule-based alerts, which suffer from delayed response times, limited coverage, high maintenance costs, and reduced reliability under challenging conditions including low illumination, adverse weather, and crowded environments. This paper presents a deep learning-based real-time railway obstruction detection framework utilizing advanced image analysis and object detection techniques. Multiple state-of-the-art neural network architectures, including SSD, Faster R-CNN, RetinaNet, YOLOv3, YOLOv7, YOLOv8, YOLOv5S6, YOLOv5X6, and YOLOv9, are implemented and systematically evaluated for detecting and localizing railway obstructions. The models are trained and tested using a hybrid dataset composed of custom railway images and publicly available datasets to ensure robustness across diverse operational scenarios. Performance is assessed using standard metrics such as Mean Average Precision (mAP), precision, recall, and inference speed. Experimental results demonstrate that YOLO-based architectures, particularly YOLOv8 and YOLOv9, achieve superior detection accuracy while maintaining real-time processing capability. These findings highlight the effectiveness of YOLO-derived models as reliable solutions for enhancing railway safety and supporting intelligent transportation systems.

Abstract:

Between 1993 and 2024, 274 hard ticks (Ixodidae) were collected from 138 underground localities in Croatia. This study represents the most extensive survey of hard tick fauna in Croatian underground habitats to date. The collected hard ticks were classified into three genera and seven species. Ixodes was the most abundant genus, with five species, while Haemaphysalis, and Hyalomma were each represented by one species. The highest number of hard tick species was collected in underground habitats in Dalmatia, followed by north-western Croatia and Slavonia. Ixodes vespertilionis Koch, 1844 was the most abundant species in the collected sample, comprising 81.02%, and was recorded in all studied regions. Only I. vespertilionis was recorded throughout the year, while I. hexagonus was found in nine months, I. frontalis in four months, and the other species in shorter periods. The largest number of I. vespertilionis specimens was collected during the spring months (33.17%), while the fewest were collected in winter (16.58%). The discovery of Ixodes lividus Koch, 1844 in north-western Croatia represents a new country record, increasing the number of hard tick species currently known in Croatia to 28.