Browse Articles

PickAMoo: LIDAR-enhanced mask R-CNN segmentation for precision weight estimation in dairy cattle using smartphone imaging

Abstract Data on body weight, as well as objective measures of body condition and size, are essential for appropriate decision-making on farm level, for example in calculations of nutrient requirements, health monitoring, and breeding-related assessment. Weighing dairy cattle and assessing their body condition are labour-intensive and are therefore often not performed as frequently as desired under practical farm conditions. Although recent research has shown strong potential for computer vision and image analysis in automated estimation of body weight, body condition score, and conformation, many existing workflows still depend on fixed multi-camera or 3D setups that increase hardware and deployment costs. We therefore developed a two-step, smartphone-centred workflow for practical live-weight estimation in dairy cattle. First, a Mask R-CNN segmentation model was trained on 567 manually annotated cow images collected under varied barn conditions and achieved an F1 score of 0.98. Second, body weight was discretized into nine data-driven categories using a Gaussian Mixture Model (BIC-selected), after which the source weight variable was removed to prevent leakage and a leak-safe pipeline (imputation, robust scaling, fold-internal SMOTE, Extra Trees) was trained within the training partition. Primary evaluation used cow-level grouped splitting so that all features based on repeated observations from the same cow remained in a single partition; a PyCaret implementation was used as an independent cross-check. On the 216-image grouped holdout, the tuned Extra Trees model achieved a macro-F1 of 0.936 (95% CI 0.860–0.983) with a 4.2% error rate. These results were obtained on 1080 images collected using the developed camera app and not used for segmentation-model training. The idea is to further streamline the algorithm to allow its downscaling and transition in the form of a smartphone application to be used on-farm as an open-source support tool.

The moderating effect of child abuse on the association between domestic violence and suicidal ideation/self-harm among Lebanese women

Phytoplankton functional response to spatial and temporal differences before and after the regime shift of Caohai Lake, China

A novel immersion chemical mechanical polishing device and tailored slurry for processing heavy duty gas turbine compressor blades

Association between METS-VF and hypertension in U.S.: a cross-sectional analysis of NHANES data

A comparative analysis of large language models for providing oral cavity cancer information

Abstract This study aimed to comparatively evaluate the medical information delivery capacity and content quality of current large language models (LLMs), specifically ChatGPT (GPT-5.2), Gemini (3.1), and DeepSeek (V4), regarding oral cavity cancer (OCC) based on expert opinions. 20 open-ended questions addressing the risk factors, diagnosis, and treatment of OCC were directed to the three models. The responses were evaluated using a blinded method by 31 expert physicians from Oral and Maxillofacial Surgery, Otorhinolaryngology (ENT), and Medical Oncology. The Modified Global Quality Scale (1–5 points) was utilised for evaluation. Statistical analyses were performed using Kruskal–Wallis, ANOVA, and Bonferroni post-hoc tests, while Fleiss’ Kappa coefficient determined inter-expert consistency. The general performance scores of the models were high (3.57–4.15). In the overall assessment, Gemini received statistically significantly higher scores than the DeepSeek model ( p  = 0.036). Significant performance differences were identified across 15 of 20 questions ( p  < 0.05); ChatGPT excelled on clinical and treatment-oriented questions, while Gemini stood out on comprehensive informational items. While no statistically significant difference was found among the specialist groups for the overall evaluation and 19 out of 20 questions ( p  > 0.05), a significant difference was observed solely for Q6 ( p  = 0.042). Although LLMs have the potential to generate high-quality information about OCC, their performance varies by content type and model architecture. While Gemini demonstrated more consistent performance overall, expert supervision remains essential before these tools can be used as reliable sources of clinical information. Clinicians must be aware of the specific strengths and limitations of different LLMs in OCC to better guide patients who increasingly use such tools for medical information.

Reservoir assessment, source rock evaluation, and 1D basin modeling in Azraq, Dead Sea, Sirhan, Jafr, and Risha basins, Jordan

Abstract This study presents a coherent petrophysical analysis of selected wells in the Azraq, Dead Sea, Sirhan, and Risha basins to define reservoir quality and hydrocarbon prospectivity in diverse geological settings. The application of a consistent set of petrophysical cutoffs enabled the systematic identification and delineation of pay intervals in terms of effective porosity, shale volume, fluid saturation, and net pay thickness, providing a robust basis for comparative assessment and prospect ranking. In total, more than 260 m of combined net pay have been assessed and characterized within the Paleozoic to Eocene interval of the studied wells from four different basins. Based on limited thin sections, quick look petrographic characteristics were evaluated. The assessment of source rocks based on Rock Eval pyrolysis data revealed that the potential source rocks are the Ghareb and Wadi Essir formations in the Azraq Basin, the Mudawwara Formation in the Risha, Jafr, and Sirhan basins, and the Dubeidib and Hiswa formations in the Sirhan Basin. The estimated transformation ratio, along with the timing of maturation relative to trap formation and migration pathways, provided a critical discriminator for determining which source intervals are likely to have contributed to the commercial charge. The integrated geochemical and basin‑modeling results emphasized that successful charge prediction requires coupling kerogen quality with robust maturity and burial histories. This cross-basin comparison study highlights the importance of integrating petrophysical analysis with source rock evaluation in informing exploration plans, identifying sweet spots, and managing reservoirs in geologically complex regions.

Harnessing salt tolerance and C4-like traits from the halophyte wild species Oryza coarctata into cultivated rice (Oryza sativa L.)

Effect of deep marginal elevation with different intermediate materials on the fracture resistance of direct and indirect final composite restorations: an in vitro study

Abstract To compare an injectable hybrid composite for deep marginal elevation with a resin-modified glass ionomer and a flowable composite, and to assess their effects on the fracture resistance of direct restorations and indirect resin composite inlays at different depths. Ninety non-carious maxillary premolar teeth were selected to receive standardized mesio-occluso-distal (MOD) cavity preparations. The specimens were systematically classified into subgroups and classes as follows: Groups (Restorative Material): Specimens were divided into two main groups ( N  = 45) based on the nano-ceramic resin composite used (Spectra ST): direct restorations or indirect inlay restorations. Subgroups (Intermediate Material): Each group was further divided into three equal subgroups ( N  = 15) based on the intermediate material (B) used: resin-modified glass ionomer (B1:RMGI), flowable resin composite (B2), and injectable hybrid composite (B3).Classes (Gingival Box Depth): Finally, each subgroup was subdivided into three classes ( N  = 5) based on the depth of the gingival box (D) relative to the cementoenamel junction (CEJ): at the CEJ (D1), 2 mm above the CEJ (D2), and 2 mm below the CEJ (D3). Fracture resistance was tested using a universal testing machine. Fracture types were assessed under a stereomicroscope. Selected samples underwent scanning electron microscopy analysis. Data were statistically analyzed using Fisher’s exact test, z-tests, Shapiro-Wilk tests, Levene’s tests, and a three-way ANOVA. direct restorations and indirect inlay restorations did not significantly affect fracture resistance when considered as a standalone factor ( p  = 0.686). For the injectable hybrid composite, the indirect composite inlay exhibited higher fracture resistance at and above the CEJ ( p  < 0.001*). The flowable composite showed a significant difference across depths when used as a direct composite restoration ( p  = 0.025*). For RMGI, the indirect composite inlay was significantly superior only below the CEJ ( p  < 0.001*). The box depth below the CEJ generally reduced fracture resistance. While a comparison between direct restorations vs. indirect composite inlay restorations alone does not dictate fracture resistance, material choice and cavity depth are critical. Indirect composite inlay restorations using injectable hybrid composites provide superior strength at/above the Cementum-Enamel Junction (CEJ), while RMGI excels only below it; greater depth reduces overall resistance.

HCA-Mamba: a hierarchical cross-attention framework combining Vision Mamba and CNN for night-time image quality assessment

The impact of attention mechanism technology for intelligent text generation in traditional CAD

Simultaneous spectrofluorimetric determination of nasal cyclic nucleotides as biochemical markers in post-COVID-19 olfactory dysfunction using supramolecular-enhanced derivative spectroscopy

Augmented reality adoption and the digital transformation of eco-smart tourism ventures: exploring the mediating role of tourist engagement

Abstract The rapid growth of digital technologies has reshaped the tourism industry, with augmented reality (AR) emerging as a powerful tool for enhancing tourist experiences and driving business innovation. This study examines the role of AR adoption in fostering the digital transformation of eco-smart tourism ventures, with a particular focus on the mediating role of tourist engagement. Guided by the Stimulus–Organism–Response (S-O-R) theory, the research explores how AR adoption stimulates immersive experience, environmental awareness support, cultural enrichment, trust, perceived sustainability value, and innovation orientation, ultimately influencing tourist engagement and digital transformation. The study employed a quantitative research design, collecting data through structured questionnaires distributed to small and medium eco-smart tourism ventures across Malaysia. A total of 367 valid responses were analysed using AMOS - Structural Equation Modelling (AMOS-SEM) to test hypothesized relationships and mediation effects. The findings reveal that AR adoption significantly enhances immersive experience, environmental awareness, cultural enrichment, perceived sustainability value, and innovation orientation, while trust was found to be insignificant. AR adoption positively influences both tourist engagement and digital transformation, with tourist engagement serving as a complementary mediator between AR adoption and digital transformation. These results highlight the strategic role of AR not only as an immersive tool but also as a catalyst for sustainable digital transformation in tourism. The study contributes theoretically by extending the S-O-R framework into AR adoption research and practically by offering insights for tourism entrepreneurs and policymakers to leverage AR for innovation, sustainability, and competitiveness.

Ambivalent self-concept and feared self mediates the relationship between negative early life experiences and obsessive-compulsive symptoms

Association between body mass index and mortality in intensive care unit patients with sepsis: A retrospective cohort study

The prevalence of neck pain and its association with studying device usage and posture among students at the University of Jordan: A cross-sectional study

Background Neck pain is one of the leading causes of discomfort and disability, especially among students who use digital devices and books for studying. This study investigates the prevalence of neck pain among students of the University of Jordan and the association of potential risk factors, including posture and study devices. Methods This cross-sectional survey was conducted at the University of Jordan. Data were collected through structured, face-to-face interviews from eligible students stratified across the University’s 20 faculties to ensure proportional representation. Known musculoskeletal disease, congenital anomalies, or prior neck/shoulder trauma or surgery cases were excluded. The questionnaire covered demographics, study behaviors, primary and secondary study utilities, posture, study duration, physical activity, and pain intensity reported based on NRS-11 scoring. Analysis was performed using JASP. We used univariate and multivariate logistic regression. Results The data analyzed showed that out of the 507 students, (52.4%) complained of neck pain within the last week, with the most complained site of pain being the neck (56%), and the least being the left shoulder (5.2%). Sitting slouched was associated with approximately two-fold higher odds of reporting neck pain compared to sitting upright with full back support (OR = 2.18, CI [1.18, 4.04], p = 0.013). Using laptops (OR = 2.65, CI [1.45, 4.87], p = 0.002) or tablets (OR = 2.68, CI [1.34, 5.36], p = 0.005) for studying was associated with a nearly 2.7-fold increase in the odds of neck pain compared with using phones. In contrast, studying with books was not significantly associated with neck pain. Conclusion Slouched sitting and use of laptops or tablets were associated with higher odds of reporting neck pain, while studying primarily on phones was associated with lower odds of reporting neck pain compared with laptops and tablets. This study identified the urgent need for ergonomic education and interventions to promote healthier study habits and reduce musculoskeletal strain in students.

The limits of debiased clinical language models for cross-hospital generalization

The acceptability of minimally invasive tissue sampling for cause of death determination in rural South Africa: A qualitative analysis

Background Minimally invasive tissue sampling (MITS) has been used to determine cause of death in various low- and middle-income countries. However, information on the acceptability of this procedure in community-based deaths is limited; most studies have focused on facility-based deaths among children. This qualitative study describes factors affecting the prospective acceptability of MITS for community deaths across all ages in a rural South African community and reviews the utility of the Theoretical Framework for Acceptability (TFA). Methods This qualitative study was conducted in the rural Agincourt Health and socio-Demographic Surveillance System (HDSS) site in Mpumalanga, South Africa. Thematic analysis was conducted on 20 in-depth interviews with residents who had experienced a death within the previous 2 years and 6 focus group discussions (FGDs) with key stakeholders. FGD groups included community members, healthcare workers, traditional healers, religious leaders and mortuary workers. Results MITS was considered acceptable by interviewees, who posited that bereaved families had a strong desire to know cause of death, which would drive participation. Limited manipulation of the body and minimal disruption of burial practices were conditions that would make MITS more readily acceptable. Facilitators of participation included engaging with local traditional leaders, rigorous community education, transparency and openness regarding MITS activities, and the provision of emotional and psychological support to the bereaved. Whilst local beliefs did not forbid participation in MITS, acceptability was limited for deaths in traditional healers and infants as it would disrupt burial in these groups. Rumours of organ-trafficking during autopsies made some participants wary of the MITS and a lack of trust in the research team could discourage participation. Conclusion In Agincourt, MITS is an acceptable procedure among community members that are interested in knowing cause of death. Thorough community engagement, open communication and an empathetic approach to bereaved families are crucial for building community support for the implementation of MITS. The TFA provides a valuable outline for the assessment of acceptability but failed to account for trust dynamics between providers and participants. We propose the modification of the TFA to include the domain “trust in providers”.

Ferroptosis-related proteins orchestrate aortic dissection: unveiling novel molecular signatures and therapeutic avenues

TransitNet: A lightweight semantic segmentation network for urban traffic scene understanding

Existing semantic segmentation networks often suffer from large parameter sizes and high computational complexity, making it difficult to deploy them on resource-constrained in-vehicle systems or edge devices. Additionally, these networks lack sufficient cross-domain adaptability, limiting their performance across diverse scenarios such as traffic and remote sensing. To address these issues, this paper proposes a lightweight semantic segmentation network, TransitNet, with enhanced Cross-Domain Adaptation capability. Based on the PSPNet architecture, TransitNet introduces a Rectangular Context Calibration Attention (RCCA) module to adaptively model the long-range spatial dependencies of road structures and traffic flow distributions. It also incorporates a Bidirectional Fusion Attention (BFA) module to enhance hierarchical feature interaction, thereby preserving fine-grained details. Furthermore, StarNet is adopted as the novel backbone network, leveraging star-shaped operations and depthwise separable convolutions to reduce model parameters. By improving the forward propagation mechanism, it outputs low-level spatial features and upsampled high-level features to support multi-scale feature fusion. Experiments demonstrate that TransitNet achieves an mIoU of 86.98% on the VOC2012 dataset, outperforming PSPNet by 1.58%. On the LoveAD remote sensing dataset, it achieves an mIoU of 61.04%, surpassing CM-UNet by 8.87%. Additionally, TransitNet achieves an mIoU of 76.11% on the OST300 dataset, further verifying its strong generalization and Cross-Domain Adaptation capabilities. By balancing efficiency and accuracy, TransitNet provides a high-performance semantic segmentation solution for real-time environmental perception in autonomous driving systems. It also holds significant potential for applications in fine classification of roads and buildings in remote sensing imagery. The code is available at https://github.com/Eric-863/TransitNet .