Manual interpretation of gastrointestinal (GI) endoscopy images is a subjective and time-intensive process, often leading to variability between clinicians and the risk of missed abnormalities. To address these challenges, we investigate the use of StyleGAN2-ADA as a generative model for unsupervised anomaly detection in gastrointestinal endoscopy images. StyleGAN2-ADA combines adaptive discriminator augmentation with style-based generative modelling, enabling the model to capture both global anatomical structures and fine-grained mucosal textures. In this work, the performance of StyleGAN2-ADA is empirically compared with the widely used f-AnoGAN framework to assess its suitability for anomaly detection in gastrointestinal endoscopy images. The model was trained and evaluated on the HyperKvasir dataset, a large and diverse collection of endoscopic images representing both normal and pathological findings. Our method achieved a ROC-AUC of 0.945 and an F1 score of 0.847, outperforming the established f-AnoGAN baseline in anomaly discrimination. These results highlight the ability of StyleGAN2-ADA to model complex multimodal distributions of healthy tissue and to emphasize abnormal regions as reconstruction defects. Clinically, this framework has the potential to support early detection of GI diseases and to integrate into real-time endoscopy workflows, reducing cognitive load for physicians, assisting less experienced clinicians, and accelerating biopsy decision-making.

Tooth axis estimation plays a key role in various digital dental workflows, including orthodontic diagnosis and prosthetic design. In clinical practice, tooth axes are typically determined by jointly considering crown morphology and radiographic information; however, this process is subjective and time-consuming. In particular, anatomical tooth axes are defined based on root information, and cone-beam computed tomography is commonly used for this purpose. However, its routine use is limited by radiation exposure, cost, and workflow complexity. As an alternative, crown-based references such as the facial axis of the clinical crown have been used as auxiliary indicators of tooth orientation in clinical settings. In this study, we propose a deep learning-based framework for automatically estimating crown-based tooth axes using only geometric information from 3D tooth crown point clouds. The proposed method first translates each tooth to its center and applies absolute orientation alignment to normalize all samples into a common coordinate system. Rotation-based data augmentation is then applied to incorporate diverse pose variations. During inference, medial-distal orientation alignment is performed to ensure consistent directional alignment of each tooth. A quaternion-based rotation regression model is used to estimate the tooth axis by rotating a predefined initial axis, and a composite loss function is employed to jointly enforce numerical, geometric, and directional consistency. Experimental results on a controlled patient-level split show that the proposed method achieves an average angular error of 3.23[Formula: see text] (standard deviation: 2.06[Formula: see text]), along with relatively consistent Chamfer distance measures. In addition, qualitative analysis indicates that the predicted tooth axes exhibit relatively consistent directional patterns at the full-arch level. These findings suggest that crown surface geometry, combined with alignment-based preprocessing, may be useful for approximating crown-based tooth axes. Rather than replacing anatomical tooth axes, the proposed framework aims to explore the feasibility of estimating clinically meaningful reference axes under limited information settings, and may serve as a supportive geometric reference for crown-based digital dentistry workflows.

This study examines the factors influencing supply chain performance in the processed fruit and vegetable sector in Jammu and Kashmir (J&K), a region still underrepresented in empirical supply chain research due to its limited geographic area and high perishability. Based on the Resource-Based View and Dynamic Capabilities Theory, the study evaluates how supply chain performance is affected by logistics, procurement, responsiveness, integration, flexibility, information technology, and communication. Sustainable supply chain practices are tested as a mediating mechanism. A structured questionnaire collected primary data from 146 fruit and vegetable processing enterprises. Data were gathered from fruit and vegetable processing units, including respondents such as managers, middle-level employees, and other staff members. Covariance-based structural equation modeling (SEM) and confirmatory factor analysis (CFA) analysed the proposed relationships. The results show that all examined enablers have a significant and positive impact on supply chain performance, with sustainable practices partially mediating these relationships, emphasising their crucial role in improving operational outcomes. The novelty of this research lies in its context-specific empirical modeling of supply chain performance in a fragile, horticulture-dominated region, addressing a notable gap in the agri-food supply chain literature. The findings provide practical insights for policymakers and industry stakeholders by highlighting strategic levers to enhance resilience, sustainability, and competitiveness in perishable agri-food supply chains facing geographic and infrastructural challenges.

Leaf hyperspectral reflectance (HSR) data have gained increasing attention due to their usage in predicting a range of leaf physiological, biochemical, structural, and photosynthetic traits using machine learning (ML) models. The PROSPECT family of models offers a complementary, mechanistic means to estimate leaf traits from HSR data using model inversion. However, a comprehensive evaluation of the accuracy and transferability of the PROSPECT model across a large set of species is hindered by the limited availability of ground truth data sets. Here, we employed a combination of inversion and forward simulation of the PROSPECT-D model across a broad range of species and identified four narrow wavebands linked to environmental effects. We also introduced a novel framework using partial least squares regression to enable the analysis of the transferability of the machine learning models trained base on the PROSPECT-D across species. This analysis revealed trait-specific patterns of transferability for the machine learning surrogate based on the PROSPECT-D forward model. We then extended this analysis to PROSPECT-D inversion using neural networks and developed a fast, accurate deep-learning-based surrogate inversion approach to estimate leaf traits from measured HSR data. Our data-driven framework paves the way for improving the accuracy of PROSPECT and similar mechanistic models.

The common leopard (Panthera pardus) is critically endangered in Pakistan, yet quantitative habitat assessments remain absent from the Himalayan foothills. This study provides the first detection-corrected occupancy assessment of leopards in Margalla Hills National Park (MHNP) using four years of systematic camera-trap data (2021-2024) across 20 sites and 2,400 sampling occasions per season. We modelled seasonal occupancy and detection probability during summer (May-August) and winter (November-February) using environmental and anthropogenic covariates. In summer, the null detection model was top-ranked (weight = 0.59), while prey relative abundance index (RAI) was the strongest occupancy predictor (β = 3.18 ± 1.31 SE, P = 0.015; ψ = 0.39 ± 0.19 SE). In winter, prey RAI strongly predicted detection probability (β = 0.437 ± 0.071 SE, P < 0.001), and occupancy was highest at sites distant from human settlements (β = 4.72 ± 2.19 SE, P = 0.031; ψ = 0.89 ± 0.13 SE), indicating active human-area avoidance during a season of heightened spatial overlap. These findings demonstrate the primacy of prey availability in driving leopard space use and identify settlement proximity as a key winter occupancy determinant. Conservation requires prey management, seasonal regulation of human activity in core leopard zones, and landscape-level planning across the MHNP-Murree Hills-Ayubia corridor.

Psychological responses to the COVID-19 pandemic showed heterogeneous trajectories rather than uniform recovery. We analyzed seven waves of nationwide web-based surveys conducted in Japan between June 2020 and December 2022 (approximately 1,000 participants per wave) to examine long-term trends in depression, anxiety, and stress and their associations with self-evaluative traits. DASS-21 scores were significantly elevated during the initial emergency phase (June 2020) and declined substantially thereafter (Kruskal-Wallis tests, all p < .001). However, distributional analyses indicated that a subgroup of individuals remained persistently distressed despite overall population-level improvement. Decision-tree-based variable-importance analyses identified the impostor phenomenon as the strongest contributor to depression, anxiety, and stress, followed by self-esteem. For anxiety, survey timing showed the strongest linear association, whereas non-linear models consistently highlighted self-evaluative variables as dominant predictors. These findings indicate that prolonged societal crises are associated with divergent mental-health adaptation patterns. Psychological self-evaluation factors, particularly impostor feelings and low self-esteem, may play a central role in persistent distress beyond demographic and social-network factors.

Scalable, low-burden tools are needed to identify individuals at risk of depression before progressing to a clinically significant level of depressive disorder. We evaluated a smartphone framework for subclinical-depression risk using self-supervised speech representations and a field-ready data-collection protocol. Participants (N = 119) were stratified into high-risk (PHQ-9 ≥ 10; n = 64) and low-risk (n = 55) groups. A mobile app elicited two 1-minute narrative recordings using negative and positive mood-induction tasks, separated by a 10-minute neutral interval. We compared four models: extreme gradient boosting on handcrafted features; a convolutional neural network-recurrent neural network (CNN-RNN) on mel-spectrograms; head-only WavLM-MLP; and fully fine-tuned WavLM (WavLM-FT). Performance was estimated using 5-fold cross-validation (CV) and out-of-fold (OOF) aggregation. In a 5-fold CV, WavLM-FT achieved the highest area under the receiver operating characteristic curve (ROC-AUC) of 0.90, area under the precision-recall curve of 0.90, F1 of 0.73, accuracy of 0.68, recall of 0.89, and precision of 0.65. In OOF subject-level predictions, WavLM-FT led (ROC-AUC 0.86; accuracy 0.79) and outperformed CNN-RNN and WavLM-MLP. Results suggest full-model adaptation captures informative paralinguistic cues within standardized smartphone recordings. A brief, ecologically valid protocol with self-supervised learning may enable scalable, non-invasive depression-risk screening.

This study investigates the in-situ alloying of ER70S-6 and 316L stainless steel using the WAAM-HW process through the fabrication of two walls with different wire proportions, employing wire feed speeds of 4m/min for the primary wire and 2m/min for the hot wire. When ER70S-6 served as the primary wire, its higher thermal and electrical conductivity promoted stable arc behavior and efficient melting of both feed materials. In contrast, the wall produced with 316L as the primary wire exhibited porosity, lack of fusion, and elemental segregation, associated with its lower thermal and electrical conductivity as well as its lower melting temperature compared to ER70S-6. Metallographic analysis revealed distinct microstructural regimes: a predominantly martensitic structure in the ER70S-6-rich wall and a heterogeneous mixture of martensite, austenite, and ferrite in the 316L-rich wall. These differences were also reflected in the microhardness results, with the martensitic wall exhibiting higher and more homogeneous hardness values (378.79 ± 17.15 HV), whereas the heterogeneous wall showed pronounced local variations (264.21 ± 42.36 HV). Both alloyed walls demonstrated significantly enhanced corrosion resistance compared with ER70S-6 (Icorr 3.69 × 10⁻⁵ A·cm⁻²), reaching values comparable to 316L (Icorr on the order of 10⁻⁷ A·cm⁻²). However, the passive behaviour characteristic of austenitic stainless steels was not fully retained, as evidenced by a reduced passive range, decreasing from 0.529V in the 316L wall to below 0.042V in the hybrid configurations. Corrosion was observed to initiate preferentially in Cr- and Ni-depleted regions, with increased chemical segregation and microstructural heterogeneity in the 316L-rich wall contributing to reduced electrochemical performance. Overall, the results highlight the potential of WAAM-HW in-situ alloying for the development of novel alloy compositions, while emphasizing the strong influence of process-dependent thermal conditions on melting efficiency, chemical homogeneity, and microstructural uniformity.

Proper monitoring of tumor progression and evaluation of treatment responses highly depend on longitudinal brain tumor segmentation from MRI data. Current deep learning methodologies have mainly concentrated on analyzing single-time-point images, which restricts the ability to incorporate temporal dynamics during the segmentation process. The study proposes a new approach called Temporal-Spatial Transformer Network (TST-Net), which can be used for longitudinal brain tumor segmentation. The proposed methodology involves a temporal attention module to integrate the progression-aware information across temporally successive MRI images and a spatial attention module to improve tumor detection. Preprocessing of BraTS longitudinal MRI data and subsequent training of TST-Net on pre-aligned datasets were done in an end-to-end fashion. Evaluation was done by Dice similarity coefficients and compared with current methods. TST-Net proved to be more effective than previous solutions, obtaining 86.0%, 88.0%, and 91.0% Dice scores for improving tumor segmentation, tumor core, and overall tumor segmentation, respectively. This study reveals that incorporating the temporal dimension along with spatial attention leads to increased accuracy of brain tumor segmentation. Incorporating temporal information and applying spatial attention can effectively decrease inconsistencies between follow-up images and contribute to improved tumor area detection. TST-Net can serve as an important tool for clinical applications. TST-Net presents an efficient solution to brain tumor segmentation due to its ability to combine both spatial and temporal attention.

The long-term stability of water-isolated coal pillars in water-accumulated goafs within ecologically fragile mining areas is severely threatened by the coupled effects of mine water immersion and mining disturbance. In this study, triaxial loading and triaxial loading with unilateral unloading experiments were conducted to simulate the complex stress paths experienced by water-isolated coal pillars throughout their service period. Based on the bearing deformation behavior and failure mechanisms of water-immersed coal samples, the main factors influencing the stability of water-isolated coal pillars during their service period were analyzed, and a comprehensive stability evaluation method was established. Acoustic emission (AE) monitoring results indicate that during triaxial loading, elevated moisture content and high confining pressure promote a transition in the failure mode of coal samples towards a tensile-shear composite pattern, whereas during triaxial loading with unilateral unloading, these factors facilitate a shift towards a predominantly tensile failure pattern. Under both stress paths, the stability of coal samples is affected by moisture content, confining pressure, and their inherent structure. Accordingly, the stability of water-isolated coal pillars is determined to be primarily influenced by three factors: the geomechanical environment of the surrounding rock, the structural characteristics of the water-isolated coal pillar, and anthropogenic mining activities. A multi-level fuzzy comprehensive evaluation system based on the analytic hierarchy process (AHP) was established, comprising these three primary indicators and thirteen secondary indicators. Indicator weights were determined using the scaling method and validated through consistency tests, and a membership function system integrating linear mapping and discrete assignment was constructed to quantify the fuzziness of influencing factors. Taking the 22614 working face of the Daliuta Coal Mine as the engineering background, the stability evaluation value of the water-isolated coal pillar was calculated to be 0.79, indicating an overall stable condition; however, the structural strength degradation under prolonged water immersion poses a potential leakage risk. Based on the evaluation results, a structural model incorporating an elastic zone and an inelastic zone was further established, from which the rational width of the water-isolated coal pillar was derived as no less than 34.12m, with a critical hydraulic head of 40.22m. The comprehensive qualitative and quantitative results suggest that it is necessary to enhance the monitoring of stress, seepage pressure, and deformation of the water-isolated coal pillar, so that effective prevention and control measures can be promptly implemented at potential local seepage points to ensure the safety and stability of the water-isolated coal pillar.