Browse Articles

TDD-YOLO: A novel model for precise detection of tomato diseases

Tomato diseases pose a significant threat to global agricultural production, often leading to substantial yield loss and major economic damage. Traditional disease detection methods rely on manual inspection, which is not only time-consuming and labor-intensive but also difficult to implement for real-time monitoring. While deep learning-based object detection techniques offer a potential alternative to manual inspection, existing models still face challenges in extracting subtle disease features, suppressing complex background interference, and in handling multi-scale disease representations in complex agricultural environments, limiting detection performance. To address these limitations, this paper proposes a novel TDD-YOLO model for precise tomato-disease detection (TDD) in complex agricultural settings. The proposed model is based on YOLOv11 with the following three main improvements: (1) a feature enhancement module is added to improve the backbone’s ability to extract disease spot textures; (2) a joint attention mechanism is introduced to explicitly model cross-dimensional dependencies, effectively suppressing background interference; and (3) a feature fusion module is added to retain disease information across different scales while reducing computational costs. Experimental results, obtained on the Tomato-Village dataset (containing field-acquired images of tomato leaves with six diseases, collected in real agricultural environments, featuring complex backgrounds and varying illumination conditions) and Tomato-Disease dataset (emphasizing a greater diversity in tomato disease types along with healthy leaf samples), demonstrate that the proposed TDD-YOLO model outperforms the baseline in detection of tomato diseases (e.g., by improving mAP@50 and mAP@50:95, averaged across disease categories, by 4.1% and 6.0% on Tomato-Village and by 3.6% and 3.9% on Tomato-Disease, respectively) and state-of-the-art models (e.g., by improving the average mAP@50 and mAP@50:95, compared to the first runner-up, by 3.2% and 4.7% on Tomato-Village and by 2.4% and 2.1% on Tomato-Disease, respectively), while maintaining good parameter count and computational complexity, confirming its effectiveness and potential for practical usage in complex agricultural environments. The author-generated code and weight files are publicly available at https://github.com/LingShaQ/TDD-YOLOCode .

Enhancing drought forecasting with CNN-TQWT and metaheuristic hybrids: evidence from Norway

Abstract Droughts are among the most damaging natural hazards, exerting severe impacts on ecosystems, economies, and societies. While typically catastrophic, their effects can vary across regions, including polar areas where outcomes may differ unexpectedly. Accurate forecasting is therefore essential for climate adaptation and resource planning. This study analyzes monthly precipitation data from 1970 to 2025 across four Norwegian cities (Bergen, Kristiansand, Oslo, and Tromsø). Standardized Precipitation Index (SPI12) values were calculated and used as inputs to predictive models based on Convolutional Neural Networks (CNN). To enhance performance, the CNN framework was combined with Random Forest (RF) and also optimization techniques, Genetic Algorithm (GA), and Particle Swarm Optimization (PSO), as well as signal decomposition methods such as Variational Mode Decomposition (VMD) and the Tunable Q-Factor Wavelet Transform (TQWT). Four input configurations were tested, and results show that CNN–TQWT hybrid models consistently outperformed other approaches across all study sites, highlighting their potential for reliable drought prediction in diverse climatic settings. Specifically, the most effective models were identified as: Bergen: CNNTQWTM03 (R = 0.9732, NSE = 0.9231) Kristiansand: CNNTQWTM03 (R = 0.9879, NSE = 0.9568) Oslo: CNNTQWTM02 (R = 0.9727, NSE = 0.9172) Tromsø: CNNTQWTM01 (R = 0.9777, NSE = 0.9395). The findings highlight the superior accuracy of the TQWT decomposition method across all stations, underscoring its potential to support decision-makers in developing effective agricultural and water management policies. Furthermore, the findings of this study will contribute to the formulation of drought preparedness and development plans. By providing a technical foundation for proactive planning, these results will assist in mitigating the impacts of drought through more resilient management strategies.

A qualitative study to inform the design and implementation of AI-driven diagnosis: Challenges, barriers, and clinical insights of physicians

Introduction Acute febrile illnesses are significant clinical challenge in tropical regions due to overlapping symptomatology among infections such as dengue, scrub typhus, and leptospirosis. This diagnostic dilemma leads to delays in appropriate treatment and negatively impacted patient outcome due to limited diagnostic tool availability. Recent advancement in artificial intelligence (AI) application in healthcare help in developing diagnostic aiding tool in such clinical dilemma. Aims Study aimed to explore physicians’ perspectives on diagnostic challenges of tropical Acute-febrile-illnesses and assess perceived utility of AI as a supportive tool in clinical decision-making. Methods A qualitative interview was conducted using validated semi-structured interview guide using literature and expert opinion and validated by IPR method. To understand the physicians perspective on management infectious diseases in critical care, interview guide administered among ten physicians with 10–30 years of experience across different centres of India. Interviews were audio-recorded, transcribed verbatim, and thematically analyzed using ATLAS.ti software. Results Four major themes emerged: clinical experience and practice, diagnostic-challenges, diagnostic-parameters and decision-making, and AI integration in clinical-practice. However, no statistical test can be performed since study mainly focused to understand perception of physicians with respect to managing illness and hence prevalence cannot be determined. Diseases such as dengue, scrub typhus, and leptospirosis were reported as most prevalent, particularly during monsoon seasons. Key diagnostic challenges included overlapping symptoms, diagnostic delays, limited sensitivity/specificity of available tests, and cost constraints. Physicians primarily relied on clinical judgment and laboratory findings, despite the limitations. Most of the participants expressed positive attitudes toward the potential of AI as supportive tool, citing its ability to enhance diagnostic accuracy and streamline clinical workflow with proper validation. Conclusions Growing in clinical dilemma AI has significant role in diagnosis of tropical fever with integration of the clinical data & physician perspective can successfully able to incorporate in clinical application.

Sustainable one-part geopolymeric hybrid composite derived from glauconite, talc, and olive seed waste–based activated carbon for Congo red adsorption

Abstract This study details the synthesis and application of a novel one-part geopolymeric hybrid composite (OP-GPHC) derived from glauconite, talc, and olive seed waste–based activated carbon for the efficient sequestration of Congo red (CR) dye from contaminated water. The hybrid binder was synthesized by impregnating activated carbon-based biogenic waste into a thermally treated glauconite/talc matrix, followed by alkali-activation with NaOH. Comprehensive characterization via XRD, FTIR, BET, TG/DTG, FESEM/EDX, and elemental mapping confirmed the material’s exceptional adsorptive properties. A Box–Behnken design (BBD) optimization established the optimal operational conditions: pH 2.0, adsorbent dosage 0.07 g/ 25 ml, contact time 77.5 min, initial CR concentration 150 mg/L, and temperature 328 K, achieving a removal efficiency of 99.2%. Equilibrium data were best described by the Langmuir isotherm, yielding a maximum adsorption capacity of 367 mg/g at 328 K, while kinetic data followed the pseudo-first-order (PFO) model. Advanced statistical physics modeling revealed a multimolecular adsorption mechanism with a vertical orientation of CR molecules at the active sites, independent of temperature, and binding energies in the range of 19.25–21.46 kJ/mol, consistent with physisorption dominated by hydrogen bonding, π–π interactions, and electrostatic forces. Thermodynamic parameters confirmed the endothermic and spontaneous nature of the process. The OP-GPHC adsorbent exhibited excellent reusability (87.8% after five cycles) and a low production cost of $0.032/g. Based on batch-derived data, treatment of 100 L of CR-contaminated water (50 mg/L) is projected to cost approximately $1.68, highlighting its strong potential for industrial-scale tertiary treatment applications.

Differential game in closed-loop supply chain of innovative products with double regrets of consumers

In this paper, we consider a three-level dynamic closed-loop supply chain differential game model led by manufacturers and recycled by third-party recyclers in order to study the impact on the dynamic closed-loop supply chain of innovative products based on simultaneous occurrence of consumer purchase regret and replacement regret. Then, we obtain the optimal path for wholesale price, retail price and return effort by solving the differential game model. Specifically, we explore the influence of the coupling effect of consumer purchase regret and replacement regret on the supply chain through theoretical analysis and numerical simulation. The results show that consumer purchase regret leads to lower sales in the supply chain, which in turn causes a decrease in long-term profits of manufacturers and retailers, while consumer replacement regret will attract more consumers to increase sales, thereby increasing the long-term profits for manufacturers and retailers accordingly. However, consumer purchase regret does not affect the profit of recyclers. In addition, both types of consumer post-purchase regret prolong the product length of time in the market, and consumer replacement regret has a more significant effect on the length of time of product in the market. Meanwhile, lower consumer purchase regret and higher replacement regret can improve the performance of all members of the entire dynamic supply chain. However, these factors do not influence the recyclers.

Coupling thin film nanofiltration membrane and photocatalyst for the removal of antidepressant drugs from aqueous solutions

High donor hemoglobin interacts with pre-transplant recipient neutropenia to modulate mortality after allogeneic hematopoietic stem cell transplantation: An exploratory, single-center, retrospective, real-world study

Prognostication after allogeneic hematopoietic stem cell transplantation remains a critical challenge, and the complex interplay between recipient vulnerability and donor graft characteristics is poorly understood. The primary objective of this retrospective, hypothesis-generating study was to investigate the interaction between pre-transplant recipient neutropenia and donor hemoglobin levels on long-term survival. We performed a pragmatic, single-center, retrospective cohort study on 94 consecutive patients who underwent transplantation at a reference center in Western Iran. Using multivariable survival models, we assessed the independent and interactive effects of pre-transplant factors on 5-year overall survival, with appropriate handling of missing data. The robustness of our central finding was confirmed via sensitivity analyses. Our adjusted multivariable analysis revealed two main findings. First, higher continuous donor hemoglobin was associated with a trend toward increased mortality (Hazard Ratio per 1 g/dL increase = 1.45; 95% Confidence Interval, 0.87–2.39; p = 0.148). Second, the central finding was a statistically significant, qualitative interaction between recipient neutropenia and donor hemoglobin (adjusted HR = 0.44, p for interaction = 0.013). This interaction reversed the potentially deleterious effect of hemoglobin: in the subgroup of neutropenic recipients, higher donor hemoglobin was associated with a protective trend, mitigating the profoundly poor prognosis observed in patients with isolated neutropenia. In conclusion, our study identified a novel and statistically robust interaction, suggesting that the prognostic impact of donor hemoglobin is context-dependent and fundamentally altered by the recipient’s baseline immune status. While these preliminary findings provide a compelling rationale for future mechanistic studies, they require urgent validation in larger cohorts and should not be used to guide clinical donor selection.

Reporting and data systems for laparoscopic cholecystectomy with findings of a user adoption survey

Crisis leadership and strategic decisions in Swedish maternity care during the COVID-19 pandemic: A deductive analysis from the COPE staff project

Background Healthcare managers played a crucial role during the COVID-19 pandemic, tasked with organizing care for a new medical condition, implementing restrictions to reduce infection spread, and handling unprecedented staff shortage while maintaining operability and ensuring a sustainable working environment for employees. The aim of this study was to generate knowledge about Swedish maternity care managers’ decision-making, by exploring their crisis management to cope and to mitigate the pandemic’s effects in working units. Methods Semi-structured interviews were conducted with 18 managers from various organizational levels at different Swedish maternity care units during the third wave of the pandemic (March – June 2021). A deductive qualitative content analysis informed by the Consolidated Framework for Implementation Research (CFIR), was performed. Results The pandemic compelled managers to employ innovation, reorganization and altered working routines to address external and internal demands. Many decisions were made rapidly, often with limited information and without sufficient time for thorough consideration. Peer support and effective communication were identified as essential for coping with the situation. Managers worked long hours and expressed both challenges in their decision-making and pride in their ability to fulfill their roles. Transparency in the decision-making process, along with continuous reflection and evaluation, were viewed as successful crisis management tools for enhancing workplace sense-making and helping employees maintain motivation. Conclusions Throughout the pandemic, managers had to develop their own methods for making and implementing decisions, aimed at ensuring patient and employee safety and well-being, often without organizational guidance on leading in crisis. It is essential to share knowledge about effective regulatory strategies to mitigate crisis impacts and to incorporate these strategies into crisis management frameworks to strengthen preparedness for future emergencies.

Autonomous RCM-less endoscope control: integrating force-based pivoting with deep learning visual servoing

Range-weighted branch length difference (RWiBaLD), a new method for distinguishing meso-endemism from neo-endemism and paleo-endemism

Our goal is to enhance understanding of phylogenetic endemism (PE), which is studied using a range-weighted phylogenetic tree. We present a new method called R ange W e i ghted B r a nch L ength D ifference (RWiBaLD), which is based on the difference between the length of a branch on a range-weighted observed tree and its length on the corresponding range-weighted comparison tree. The latter is a tree with the same topology as the observed tree but with branches adjusted to be of equal length before range-weighting. The goal of this method is to detect the branches of the phylogeny that contribute most to PE, and to distinguish among neo-endemism, paleo-endemism, and a novel category: meso-endemism. The latter is a heretofore missing category in the study of endemism referring to a highly range-restricted lineage that is not particularly short (neo-) or long (paleo-). Application of this approach is illustrated using the plant genus Acacia across the continent of Australia, using previously published spatial data and phylogeny. We show the properties of RWiBaLD, and how to apply it by: (1) mapping the outputs on a phylogeny, thus for the first time enabling PE to be scored for use in existing phylogenetic comparative methods, and (2) dissecting the composition of concentrations of PE on the landscape that are detected by the Categorical Analysis of Neo- And Paleo-Endemism (CANAPE) method. Whereas CANAPE identifies geographic concentrations of high PE, and gives a summary classification of the type of endemism dominating in a location, RWiBaLD identifies the specific branches that contribute the most to PE, and what type of endemism they represent. Thus, RWiBaLD enables novel studies of the evolutionary and ecological causes of endemism, as well as improved conservation assessment.

The miR4261/PAMM axis in multiple myeloma promotes bone resorption

Distinct early-life mechanisms of quantity discrimination in domestic chicks

In research on animal numerical cognition, newly hatched domestic chicks have been shown to rely on distinct strategies when confronted with quantitative choices. In some conditions, such as after imprinting on a specific set of objects, chicks preferentially approach the larger set of familiar items, indicating sensitivity to magnitude. In other conditions, however, their responses are governed not by magnitude per se, but by the degree of similarity between the test objects and a previously experienced set, where similarity is defined in terms of conformity to specific perceptual constraints, such as the possibility of a symmetrical division into identical subsets (as in composite versus prime sets of items). In the present study, we sought to replicate both phenomena while aligning key methodological features, including the test arena, the comparison (5 vs. 9), and the age at testing. One group of chicks was imprinted on a set of identical objects to test preference for larger familiar set; another group was habituated to even-numbered sets to assess sensitivity to perceptual asymmetry in prime-numbered ones. We successfully replicated both effects: chicks preferred the larger set after imprinting and showed longer inspection of the prime-numbered set after habituation, despite its smaller magnitude. Our results show that different mechanisms supporting quantity discrimination are available from the earliest stages of life and can be triggered by task- or environment-specific factors.

Patient-related diagnostic delay and risk of unfavorable treatment outcomes among pulmonary tuberculosis patients at the antituberculosis center of Brazzaville, Republic of Congo

Abstract Tuberculosis (TB) remains a major public health concern worldwide. Early diagnosis is crucial to reduce TB transmission and improve treatment outcomes. This study assessed patient-related diagnostic delay and their impact on the treatment outcomes among pulmonary tuberculosis patients (PTB) at the Antituberculosis center in Brazzaville. We conducted a prospective cohort study from July 2023 to August 2024. Sociodemographic, clinical characteristics, patient-related diagnostic delays delay (short ≤ 30 days, prolonged > 30 days), and treatment outcomes were recorded. Logistic regression models were used to identify risk factors, reporting crude and adjusted odds ratios (OR, AOR) with 95% confidence intervals (CI). A p-value < 0.05 was considered significant. A total of 313 patients was included (median age: 34 years, range 24–41); 295 (94.2%) were newly diagnosed, and 16 (5.1%) were HIV-positive. Men accounted for 69% of cases, and the age group 24–44 was the most represented (55.9%). The median patient delay was 30 days (IQR 21–62), and 135 (43.1%) experienced prolonged delays. Multivariate analysis showed that residence in Mfilou district was associated with longer delays (OR = 2.77, 95%CI: 1.22–6.30; p  = 0.004), whereas diabetes mellitus was linked to shorter delays (OR = 0.15, 95%CI: 0.02–1.13; p  = 0.008). Although patient-related diagnostic delay was not significantly associated with treatment outcomes, patients with delays > 30 days had higher odds of death (OR = 2.30, 95%CI: 0.7–7.3) and treatment failure (OR = 5.4, 95%CI: 0.8–66.4). A high median patient-related diagnostic delay of 30 days was observed in Brazzaville. Residence in peripheral districts and diabetes mellitus status were significant predictors of delay. Although not statistically significant, prolonged patient delays tended to be associated with higher risks of death and treatment failure. Strengthening early case detection and promoting prompt healthcare-seeking and diagnosis among symptomatic individuals are critical for reducing TB diagnostic delays and improving treatment outcomes.

Trajectory planning method for pipeline installation robots based on AS-DTRRT

Pipeline installation robots install pipelines on both sides of mine roadways using robotic arms; the geometric shape and length of the pipelines affect the movement trajectory of the robotic arms, and the working environment is complex and changeable. Aiming at the issues of the mechanical arm’s long motion trajectory, low installation efficiency, and pipeline collision vulnerability during installation, a trajectory planning method based on the improved dual-tree RRT algorithm (AS-DTRRT) is proposed. Taking the pipeline installation robot with a 6-degree-of-freedom Cartesian coordinate robotic arm as the research object, the Artificial Potential Field method is introduced to guide the growth direction of new nodes in the path of the dual-tree RRT algorithm; aiming at the problem that the Artificial Potential Field method is prone to falling into local optimality, an adaptive step-size dynamic environment bias strategy is proposed, which automatically adjusts the step size according to the environmental conditions of the position where the new node is located to further improve the path search capability. This study analyzes the influence of pipeline geometry and length on the robotic arm’s working space, proposes a pipeline collision avoidance strategy considering the working environment, and applies quintic polynomial interpolation to plan the trajectory of collision-free paths while considering joint speed and acceleration constraints. Simulations and platform experiments comparing AS-DTRRT with RRT*, D-RRT, Bi-RRT*, and A-RRT validate the method’s feasibility, demonstrating that AS-DTRRT outperforms the four algorithms by reducing actual runtime by 20.22%, 15.48%, 4.69%, and 16.47% and shortening average path lengths by 21.17%, 13.78%, 6.85%, and 9.24% respectively. The proposed method enhances search efficiency and reduces path lengths, providing an effective approach for trajectory planning of pipeline installation robots.

Advanced statistical models to handle response styles and uncertainty when modelling emotional intelligence of elite swimmers

Abstract Emotional intelligence is a key factor for success in sporting competitions, arousing great interest in the psychological assessment of athletes. When the evaluation of psychological behaviour relies on Likert-type psychometric scales, individuals could tend to respond to items regardless of their content or by selecting the extremes or the middle part of the response scale, compromising the measurement process. In this vein, the present paper aims to address measurement issues regarding uncertainty and response style during the assessment of emotional intelligence of elite swimmers by exploiting latent trait models. Results provide evidence in favour of models accounting for specific response behaviour compared to simple item response theory models.

Retraction: SPAG6 hypermethylation silences a novel tumor suppressor and inhibits renal cell carcinoma progression via PI3K/AKT/mTOR pathway

Weekend admission and outcomes in cancer-associated pulmonary embolism: a cross-sectional national inpatient sample study, 2016–2022

Abstract The “weekend effect” has been variably associated with outcomes in acute pulmonary embolism (PE), but its relevance in cancer-associated PE remains uncertain. Among U.S. hospitalizations for cancer-associated PE, is weekend admission independently associated with in-hospital mortality, and do PE-related interventions differ by day of admission? Retrospective cross-sectional study using the HCUP National Inpatient Sample (2016–2022). Adults (≥ 18 years) hospitalized with a principal diagnosis of PE and concomitant malignancy were included. Weekend admission was Saturday–Sunday. Survey-weighted analyses in R generated national estimates. Survey-weighted logistic regression estimated adjusted odds ratios (ORs) for in-hospital mortality. We compared PE-related therapies/procedures (mechanical ventilation, vasopressors, systemic thrombolysis, and inferior vena cava [IVC] filter placement), including hospital-day-1 use (0 days to procedure). We identified 37,491 unweighted cancer-associated PE hospitalizations, representing approximately 187,455 weighted national hospitalizations, of which 7,804 unweighted hospitalizations (20.8%) were weekend admissions. Weekend admissions had higher acuity, with the administrative marker of high-acuity PE more frequent on weekends (29.1% vs. 26.3%), and higher unadjusted mortality (6.6% vs. 5.9%). In the fully adjusted model ( N  = 36,408; deaths = 2,210), weekend admission was not associated with mortality (adjusted OR 1.04, 95% CI 0.93–1.16). Mechanical ventilation was more frequent on weekends (4.7% vs. 3.6%). Thrombolysis and overall IVC filter use were similar, but hospital-day-1 IVC filter placement was less common on weekends (14.7% vs. 18.9%). Higher unadjusted weekend mortality in cancer-associated PE appears largely explained by case mix and severity rather than weekend admission itself.

Characteristics of Escherichia coli ST131 strains isolated from dogs and cats with urinary tract infections in a teaching hospital in Taiwan

Escherichia coli sequence type (ST) 131 is a globally disseminated multidrug-resistant clone that poses a substantial public health threat. This study investigated the prevalence, virulence characteristics, and antimicrobial resistance profiles of ST131 among companion-animal E. coli isolates. A total of 400 E. coli isolates obtained from dogs and cats diagnosed with urinary tract infections at the National Taiwan University Veterinary Hospital between 2011 and 2019 were analyzed. Phylogenetic grouping identified 192 isolates (48.0%) belonging to phylogroup B2, which is strongly associated with ST131. Among these, 26 isolates (13.5%) were confirmed as ST131 by single-nucleotide polymorphism screening and multilocus sequence typing, and 19 isolates were selected for detailed characterization. Virulence gene analysis by multiplex PCR showed that fyuA (100.0%), traT (94.7%), iutA (89.5%), and kpsMT II (84.2%) were the most prevalent genes. Antimicrobial susceptibility testing demonstrated universal susceptibility to meropenem, whereas all isolates were resistant to amoxicillin and ampicillin. Six canine-derived isolates produced extended-spectrum β-lactamases (ESBLs) or AmpC β-lactamases. Among ESBL producers, bla gene groups bla CTX-M-1 , bla CTX-M-9 , and bla TEM were detected, whereas bla CMY-178 was the only AmpC gene identified. Conjugation experiments demonstrated transferability for most bla genes, except bla TEM-102 , bla TEM-12 , and bla CTX-M-55 . Plasmid replicon typing revealed IncF (100%) and IncB/O (69%) as the predominant plasmid groups. Only two O-antigen types, O25b and O16, were detected among ST131 isolates, which carried fimH variants fimH22 , fimH27 , fimH30 , or fimH41 subclones. These findings demonstrate the circulation of virulent and multidrug-resistant ST131 strains in companion animals in Taiwan, highlighting their potential relevance within a One Health context.

Serum antithrombin III as an early predictive marker for post-hepatectomy liver failure: a prospective cohort study

Abstract Post-hepatectomy liver failure (PHLF) remains a serious complication following liver resection, yet early prediction is an unmet clinical need. This prospective study enrolled 151 patients at elevated risk of PHLF (major hepatectomy, thrombocytopenia, or hyperbilirubinemia) and evaluated serum antithrombin III (ATIII) activity, measured preoperatively and on postoperative days (PODs) 1, 2, 3, and 5, as an early predictive marker. PHLF, diagnosed according to the International Study Group of Liver Surgery criteria, occurred in 35 patients (23.2%). ATIII activity was evaluated as raw values and as the percentage change from the preoperative baseline. Raw ATIII activity was significantly lower in the PHLF group at all time points ( P  < 0.001). The POD 3 ATIII decrease was significantly greater in the PHLF group (36% vs. 29%, P  = 0.041). Multivariable logistic regression identified ATIII change from baseline ≥ 30% at POD 3 (odds ratio 3.04, P  = 0.021), ALBI grade B (OR 2.77, P  = 0.031), and ICG R-15 ≥ 15% (OR 3.50, P  = 0.034) as independent risk factors for PHLF. The multivariable model demonstrated acceptable discriminative performance (apparent AUC 0.750; bootstrap-corrected AUC 0.730). These findings suggest that early postoperative decline in ATIII activity could serve as an early biomarker for identifying patients at increased risk of PHLF, potentially enabling timely intervention.