Search Results (2,625)

Background/Objectives: Conventional workflows, peripheral blood smears, and bone marrow assessment supplemented by LDI-PCR, molecular cytogenetics, and array-CGH, are expert-driven in the face of biological and imaging variability. Methods: We propose an AI pipeline that integrates convolutional neural networks (CNNs) and transfer learning-based models with two explainable AI (XAI) approaches, LIME and Grad-Cam, to deliver both high diagnostic accuracy and transparent rationale. Seven public sources were curated into a unified benchmark (66,550 images) covering ALL, AML, CLL, CML, and healthy controls; images were standardized, ROI-cropped, and split with stratification (80/10/10). We fine-tuned multiple backbones (DenseNet-121, MobileNetV2, VGG16, InceptionV3, ResNet50, Xception, and a custom CNN) and evaluated the accuracy and F1-score, benchmarking against the recent literature. Results: On the five-class task (ALL/AML/CLL/CML/Healthy), MobileNetV2 achieved 97.9% accuracy/F1, with DenseNet-121 reaching 97.66% F1. On ALL subtypes (Benign, Early, Pre, Pro) and across tasks, DenseNet121 and MobileNetV2 were the most reliable, achieving state-of-the-art accuracy with the strongest, nucleus-centric explanations. Conclusions: XAI analyses (LIME, Grad-CAM) consistently localized leukemic nuclei and other cell-intrinsic morphology, aligning saliency with clinical cues and model performance. Compared with baselines, our approach matched or exceeded accuracy while providing stronger, corroborated interpretability on a substantially larger and more diverse dataset. Full article

Background/Objectives: Virtual reality (VR) is increasingly used to reduce pain during pediatric needle procedures, but its effectiveness may vary by developmental stage and gender. This study evaluated whether immersive VR reduces venipuncture pain in children and adolescents and examined parent–patient agreement and gender-specific response patterns. Methods: A prospective nonrandomized clinical study was conducted within a hospital-based pediatric venipuncture service using an alternating 1:1 allocation sequence. Participants aged 4–18 years underwent venipuncture with either VR (n = 49) or standard care (n = 29). Procedural pain was measured using the Faces Pain Scale–Revised (FPS-R) with independent parent ratings. Analysis of covariance (ANCOVA) compared post-procedural FPS-R scores while adjusting for baseline pain. Exploratory age and gender-specific analyses were also performed. Results: VR led to a clear reduction in pain for children, even after adjusting for baseline scores (3.55 vs. 4.73; p = 0.003). Adolescents, however, reported similarly low pain in both groups (2.81 vs. 2.79; p = 0.60), and several mentioned that the PEGI 3 content felt too young for them, which likely limited how engaged they were. Among children, girls showed the most noticeable drop in pain, which matches the subgroup’s adjusted significance (p = 0.011). Parent–patient agreement was stronger in children (r ≈ 0.7–0.8) than in adolescents (r ≈ 0.4–0.5), and VR did not change this pattern. Most participants said they would choose VR again for future procedures. Conclusions: Immersive VR helped reduce venipuncture pain in children but had little effect in adolescents, underscoring the need for age-appropriate or more interactive VR content for older patients. Overall, these findings support using VR selectively as a distraction tool that fits the developmental needs of pediatric groups. Full article

Background/Objectives: Healthcare workers (HCW) in primary care settings play a significant role in recommending vaccines to patients. We aimed to describe COVID-19 vaccination knowledge, attitudes, perception, and practices (KAPP) of HCWs in Tunisia and identify associated factors. Methods: We conducted a national cross-sectional survey (29 January to 3 February 2024) among HCWs in primary public healthcare centers using purposive sampling. Factors associated with good knowledge, positive attitude, and good practice, measured through Likert scales using face-to-face questionnaires, were identified using binary logistic regression. Results: We included 906 HCWs (mean age = 41.87 ± 8.89 years). In total, 37.75% (342/906) of HCWs had good knowledge and perception, 4.30% (39/906) had a positive attitude, and 24.9% (226/906) had good practices related to COVID-19 vaccination. Working in urban compared to rural areas was associated with good knowledge (aOR = 1.57, 95%CI = 1.12–2.21) and positive attitude (aOR = 4.94, 95%CI = 1.19–20.44) to COVID-19 vaccination. Physicians had better KAPP scores than other medical professionals. HCWs working in departments with high-risk patients were more likely to have good knowledge (aOR = 1.28, 95%CI = 1.00–1.72). Positive attitude was also associated with being male (aOR = 2.97, 95%CI = 1.75–5.07) and having at least one non-communicable disease (aOR = 1.92, 95%CI = 1.14–3.23). Being male (aOR = 1.97, 95%CI = 1.35–2.88) and having more years of professional experience (aOR = 1.81, 95%CI = 1.29–2.52) were associated with good practice. Conclusions: Just over a third of HCWs in primary healthcare clinics had good knowledge of COVID-19 vaccination, while positive attitudes and good practices were low. Targeted interventions, particularly for HCWs with less professional experience working in rural settings, are needed to increase good practices and improve COVID-19 vaccination coverage in Tunisia. Full article

Aquaculture digitalization is increasingly regarded as a crucial pathway to improving productivity, sustainability, and resilience in the fisheries sector. Policy instruments intended to foster this digital transformation—such as substantial subsidies and stringent regulatory mandates—often face constraints stemming from fiscal limitations, administrative burdens, and implementation inefficiencies. Behavioral interventions (nudges) represent a potentially effective and less resource-intensive alternative, yet their capacity—individually or in conjunction with moderate subsidies and regulatory measures—to foster aquaculture digitalization remains empirically underexplored. Drawing on survey data from 254 fish farmers in the lower Yangtze River region and employing a combination of principal component analysis (PCA), ordinary least squares (OLS) regression, Propensity Score Matching (PSM), and Gradient Boosted Trees (GBT) techniques, this study finds that: (1) Social nudging has a robust and consistent positive effect on digital transformation; (2) The effects of subsidies and regulations are heterogeneous and context-dependent; (3) The negative interactions between nudging and constraints, as well as between nudging and subsidies, are context-dependent and tend to inhibit digital transformation; (4) Policy effects display marked heterogeneity across different contexts, particularly with respect to sales channels, external pressures, producers’ transformation capabilities, and the scale of aquaculture operations. These findings deepen the understanding of how behavioral and structural policies interact in agricultural digitalization, emphasizing that effective policy should combine financial and regulatory measures with efforts to strengthen farmers’ digital awareness and behavioral adaptability. Full article

Current approaches to dynamic sign language recognition commonly rely on dense landmark representations, which impose high computational cost and hinder real-time deployment on resource-constrained devices. To address this limitation, this work proposes a computationally efficient framework for real-time dynamic Mexican Sign Language (MSL) recognition based on a multimodal minimal angular-geometry representation. Instead of processing complete landmark sets (e.g., MediaPipe Holistic with up to 468 keypoints), the proposed method encodes the relational geometry of the hands, face, and upper body into a compact set of 28 invariant internal angular descriptors. This representation substantially reduces feature dimensionality and computational complexity while preserving linguistically relevant manual and non-manual information required for grammatical and semantic discrimination in MSL. A real-time end-to-end pipeline is developed, comprising multimodal landmark extraction, angular feature computation, and temporal modeling using a Bidirectional Long Short-Term Memory (BiLSTM) network. The system is evaluated on a custom dataset of dynamic MSL gestures acquired under controlled real-time conditions. Experimental results demonstrate that the proposed approach achieves 99% accuracy and 99% macro F1-score, matching state-of-the-art performance while using fewer features dramatically. The compactness, interpretability, and efficiency of the minimal angular descriptor make the proposed system suitable for real-time deployment on low-cost devices, contributing toward more accessible and inclusive sign language recognition technologies. Full article

Background/Objectives: Athletes face both daily and sport-related stressors while being expected to perform at an optimal level. Effective recovery, particularly adequate sleep, plays a key role in psychophysiological restoration and performance, whereas sleep deprivation may impair functioning and increase perceived stress. This study examined the associations between coping strategies, sleep quality, athletic experience, competitive level, and perceived stress during the pre-competition period among female volleyball players. Methods: Ninety-one athletes (aged 18–35, M = 23.03, SD = 4.37) from three Polish professional leagues—Tauron (n = 31), First League (n = 30), and Second League (n = 30)—completed an online battery including the Stress Coping Strategies in Sport Questionnaire (SR3S), the Perceived Stress Scale (PSS-4), the Pittsburgh Sleep Quality Index (PSQI), and a demographic survey. Results: Based on PSQI scores, approximately 60% of the athletes were classified as poor sleepers. No significant differences in sleep quality or perceived stress were found across leagues. However, athletes competing in higher leagues reported more frequent use of mental coping strategies. Athletic experience, sleep quality, and the coping strategy of seeking social support were significantly associated with perceived stress. Players with less experience, poorer sleep, and a greater tendency to seek social support reported higher stress levels. The positive association between support-seeking and stress likely reflects reactive coping among more stressed athletes rather than a maladaptive effect of social support. Conclusions: These findings underscore the importance of promoting adaptive coping and sleep hygiene in competitive sport, particularly among less experienced female athletes during the pre-competition period. Full article

A key component of creating robust and sustainable businesses is the digital transformation of business operations. This study examines the impact of distributed digital technology, namely cloud computing and blockchain technology, on an auditor’s behavior, an essential component of the framework for corporate responsibility. This study also highlights the impact of digital transformation on sustainable auditing, urging auditors to improve their technological skills to build trust in evolving entities. We used a unique dataset of Chinese A-share listed companies from 2013 to 2021 to show that this time period is important because it shows the beginning and growth of these technologies in the Chinese business world. This gives us a good starting point for looking at their early-stage audit effects. Our key findings are threefold. First, we found that firms using distributed digital technologies (cloud computing and blockchain) experienced (a) higher audit fees and (b) standard audit opinions, indicating the growing complexity and the requirement that auditors acquire specialized skills in order to evaluate cyber-resilience and technological structures. Second, firms facing substantial profit fluctuations (higher risk level) following digital engagement were subject to higher audit fees and a decreased probability of standard audit outcomes, emphasizing the nuanced risks of digital transformation. Third, the main results were more pronounced in (a) non-state-owned enterprises and (b) high-tech enterprises. Our study is robust to multiple sensitivity analyses, endogeneity tests, and propensity score matching (PSM). The results show that regulators need to create and support specialized auditing regulations regarding distributed technologies. These regulations would assist auditors in evaluating cloud and blockchain engagement and make it clear to businesses what is important to be compliant. Full article

As a non-invasive diagnostic technique, evolved gas analysis (EGA) holds significant value in assessing the insulation conditions of critical equipment such as power cables. Current analytical methods face two major challenges: insulation materials may undergo multiple aging mechanisms simultaneously, leading to interfering characteristic gases; and traditional approaches lack the multi-label recognition capability to address concurrent fault patterns when processing mixed-gas data. These limitations hinder the accuracy and comprehensiveness of insulation condition assessment, underscoring the urgent need for intelligent analytical methods. This study proposes a deep convolutional neural network (DCNN)-based multi-label classification framework to accurately identify the gas generation characteristics of five typical power cable insulation materials—ethylene propylene diene monomer (EPDM), ethylene-vinyl acetate copolymer (EVA), silicone rubber (SR), polyamide (PA), and cross-linked polyethylene (XLPE)—under fault conditions. The method leverages concentration data of six characteristic gases (CO₂, C₂H₄, C₂H₆, CH₄, CO, and H₂), integrating modern data analysis and deep learning techniques, including logarithmic transformation, Z-score normalization, multi-scale convolution, residual connections, channel attention mechanisms, and weighted binary cross-entropy loss functions, to enable simultaneous prediction of multiple degradation states or concurrent fault pattern combinations. By constructing a gas dataset covering diverse materials and operating conditions and conducting comparative experiments to validate the proposed DCNN model’s performance, the results demonstrate that the model can effectively learn material-specific gas generation patterns and accurately identify complex label co-occurrence scenarios. This approach provides technical support for improving the accuracy of insulation condition assessment in power cable equipment. Full article

Background/Objectives: Caregivers of infants with congenital muscular torticollis (CMT) frequently seek information online, although the accuracy, clarity, and safety of web-based content remain variable. As large language models (LLMs) are increasingly used as health information tools, their reliability for caregiver education requires systematic evaluation. This study aimed to assess the reproducibility and quality of ChatGPT-5.1 responses to caregiver-centered questions regarding CMT. Methods: A set of 17 questions was developed through a Delphi process involving clinicians and caregivers to ensure relevance and comprehensiveness. ChatGPT generated responses in two independent sessions. Reproducibility was assessed using TF–IDF cosine similarity and embedding-based semantic similarity. Ten clinical experts evaluated each response for accuracy, readability, safety, and overall quality using a 4-point Likert scale. Results: ChatGPT demonstrated moderate lexical consistency (mean TF–IDF similarity 0.75) and high semantic stability (mean embedding similarity 0.92). Expert ratings indicated moderate to good performance across domains, with mean scores of 3.0 for accuracy, 3.6 for readability, 3.1 for safety, and 3.1 for overall quality. However, several responses exhibited deficiencies, particularly due to omission of key cautions, oversimplification, or insufficient clinical detail. Conclusions: While ChatGPT provides fluent and generally accurate information about CMT, the observed variability across topics underscores the importance of human oversight and content refinement prior to integration into caregiver-facing educational materials. Full article

The proliferation of plastic waste across natural ecosystems has created a global environmental and public health crisis. Monitoring plastic litter using remote sensing remains challenging due to the significant variability in terrain, lighting, and weather conditions. Although earlier approaches, including classical supervised machine learning techniques such as Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM), applied to hyperspectral and multispectral data have shown promise in controlled settings, they often may face challenges in generalizing across diverse environmental conditions encountered in real-world scenarios. In this work, we present a deep learning framework for pixel-wise segmentation of plastic waste in short-wave infrared (900–1700 nm) hyperspectral imagery acquired from an Unmanned Aerial Vehicle (UAV). Our architecture integrates attention gates and residual connections within a U-Net backbone to enhance contextual modeling and spatial-spectral consistency. We introduce a multi-flight dataset spanning over 9 UAV missions across varied environmental settings, consisting of hyperspectral cubes with centimeter-level resolution. Using a leave-one-out cross-validation protocol, our model achieves test accuracy of up to 96.8% (average 90.5%) and a 91.1% F1 score, demonstrating robust generalization to unseen data collected in different environments. Compared to classical models, the deep network captures richer semantic representations, particularly under challenging conditions. This work offers a scalable and deployable tool for automated plastic waste monitoring and represents a significant advancement in remote environmental sensing. Full article

Traditional cardiopulmonary resuscitation (CPR) training faces limitations such as instructor dependency, low efficiency, and subjective assessment. To address these issues, this study proposes a novel computer vision-based method for the automation and objective evaluation of artificial respiration, shifting focus to the long-overlooked ventilation component. We developed an evaluation framework integrating human pose estimation and spatio-temporal graph convolution network (ST-GCN): first, OpenPose is utilized to extract skeletal keypoints of the rescuer, followed by action classification and recognition-including chest compressions, airway opening, and artificial breathing via a ST-GCN. Based on the American Heart Association (AHA) guidelines, this research defines and implements five quantitative metrics for ventilation quality, including CPR operation procedure, chin-frontal angle, interruption time, ventilation time, and ventilation frequency. An automated scoring model was established accordingly. Validated on a self-constructed dataset containing multi-source videos, the model achieved an accuracy of 87.64% in recognizing artificial respiration actions and 84.47% in evaluating action standardization. Experimental results demonstrate that the system can effectively and objectively evaluate the quality of artificial respiration. Compared with traditional instructor-dependent approaches, this study provides a low-cost, scalable technical solution, offering a new pathway for promoting high-quality CPR training. Full article

Polygenic autoimmune diseases (ADs) have several common features that are caused by a complex interplay of genetic and environmental factors. Common pathophysiological mechanisms include dysregulation of the immune system, chronic inflammation, and epigenetic changes influenced by external factors. For the prediction of the genetic predisposition of AD manifestation, polygenic risk scale (PRS), or polygenic scores (PGSs), are used. Use of PRSs faces several challenges such as applicability on a specific population, performance comparison, and estimation of biological relevance based on SNP number. We compared PRS with different numbers of SNPs and tried to find the common genetic core of ADs. Our analysis revealed a list of the most common altered genes, which we annotated and interpreted. Clustering of PRS based on used genes showed that clusters of ADs remained consistent across all chosen PRS sizes. We concluded that PRS size does not have an impact on biological relevance. Full article

Objectives: Intensive care units (ICUs) are characterised by high care complexity and nursing workload, which directly affects patient safety and staff sustainability. Latvia faces a chronic shortage of nurses, particularly in intensive care, yet systematic national data on nursing workload have been lacking. This study aimed to quantitatively assess nursing workload in Latvian ICUs using the Nursing Activities Score (NAS) and to evaluate its relationship with staffing adequacy. Methods: A prospective, multicentre observational study was conducted over six months (May–November 2025) in 14 Latvian ICUs representing all three levels of intensive care. Nursing workload was measured using the NAS during each 12 h shift. A total of 28,079 complete NAS observations were analysed using descriptive statistics, inferential tests (t-tests, ANOVA), mixed-effects modelling, regression analysis, and time-series forecasting. Results: The mean NAS was 65.45 (SD = 25.76), equivalent to an average of 15.71 nursing care hours per patient per day. Workload remained similarly high during day and night shifts. Significant differences were observed between ICUs and care levels, with level 2 units showing the highest workload. The average nursing shortage rate was 42.6% and was strongly predicted by NAS values (R² = 0.115), whereas shift type and unit level had minimal explanatory power. Conclusions: ICU nursing workload in Latvia is persistently high and unevenly distributed across units. Staffing levels are not adequately adjusted to actual care demands. Integrating NAS-based workload monitoring into staffing models is essential for evidence-based workforce planning, improving patient safety, and reducing nurse overburdening. Full article

The coronavirus is highly contagious, making rapid early diagnosis essential. Although deep learning-based diagnostic methods using CT or X-ray images have advanced significantly, they still face limitations in cost, processing time, and radiation exposure. In addition, for the possibility of real-time COVID-19 diagnosis, model lightweighting is required. This study proposes a lightweight deep learning model for COVID-19 diagnosis based on fluorescence images and demonstrates its applicability in embedded environments. To prevent data imbalance caused by noise and experimental variations, images were preprocessed using Gray Scale conversion, CLAHE, and Z-Score normalization to equalize brightness values. Among the tested architectures—VGG, ResNet, DenseNet, and EfficientNet—ResNet152 and VGG13 achieved the highest accuracies of 97.25% and 93.58%, respectively, and were selected for lightweighting. Layer-wise importance was calculated using an imprinting-based method, and less important layers were pruned. The pruned VGG13 maintained its accuracy while reducing model size by 18.9 MB and parameters by 4.2 M. ResNet152 (Prune 39) improved accuracy by 1% while reducing size by 161.5 MB and parameters by 40.22 M. The optimized model achieved 129.97 ms, corresponding to 7.69 frames per second (FPS) on an NPU(Furiosa AI Warboy), proving real-time COVID-19 diagnosis is feasible even on low-power edge devices. Full article

Background/Objective: This study aims to examine changes in hedonic hunger and problematic eating behaviors over a 24-week follow-up period after sleeve gastrectomy and Roux-en-Y gastric bypass and to compare differences between the two surgical procedures. Methods: This prospective observational study included 144 adults who underwent sleeve gastrectomy (n = 74) or Roux-en-Y gastric bypass (n = 70). Hedonic hunger was assessed using the Power of Food Scale, and problematic eating behaviors were evaluated with the Eating Disorder Examination Questionnaire. Data was collected one week before surgery and 24 weeks postoperatively through face-to-face interviews. Results: At the 24-week follow-up, participants who underwent gastric bypass had higher total Power of Food Scale scores than those who underwent sleeve gastrectomy (2.42 vs. 2.15), although reductions from baseline were not significantly different (−1.31 vs. −1.16, p = 0.136). Both procedures resulted in significant decreases in total Eating Disorder Examination Questionnaire scores (sleeve gastrectomy: 2.37 to 1.00, p < 0.01; gastric bypass: 2.41 to 1.36, p < 0.01), as well as in Eating Concern, Shape Concern, and Weight Concern subgroups. Reductions in Eating Disorder Examination Questionnaire total score and in Shape Concern and Weight Concern subgroups score were greater after sleeve gastrectomy (−1.37 vs. −1.05, p = 0.030). Total weight loss percentage was positively correlated with decreases in Eating Disorder Examination Questionnaire scores in both groups (p = 0.010) and was significantly associated with Power of Food Scale reductions only in sleeve gastrectomy (r = 0.163, p = 0.014). Conclusions: Both procedures reduce hedonic hunger and problematic eating behaviors, but the magnitude of change and its effect on weight loss may vary by surgical method. Full article