Search Results (7,258)

This paper presents the results of research on the technical–economic optimization of FDM parameters (L_h—layer height and I_d—infill density percentage) for the manufacture of tensile and compression samples from recycled materials (r) of PETG (polyethylene terephthalate glycol) and ASA (acrylonitrile styrene acrylate) in the context of the transition to a circular economy. To carry out our technical–economic study, the fundamental principle of value analysis was used, which consists of maximizing the ratio between V_i and C_p, where V_i represents the mechanical characteristic (tensile strength or compressive strength) and C_p represents the production cost. The results of this study showed that, in the case of tensile samples manufactured by recycled PETG (rPETG), the parameter that significantly influences the results of the V_i/C_p ratios is L_h (the height of the layer), while for the samples manufactured additively from recycled ASA (rASA), the parameter that decisively influences the tensile strength is I_d (the infill density percentage). In the case of compression samples manufactured by FDM from recycled PETG (rPETG) and recycled ASA (rASA), the parameter that signified influences the results of the V_i/C_p ratios is I_d (the infill density percentage). Following the optimization of the FDM parameters, using multiple-response optimization, we identified the optimal parameters for the manufacture of parts by FDM from rPETG and rASA: L_h = 0.20 mm and I_d = 100%. The results of this study demonstrated that the use of recycled plastics from PETG and ASA lends itself to a production and consumption model based on a circular economy. Full article

The increasing demand for hazelnut kernels is favoring an upsurge in hazelnut cultivation worldwide, but ongoing climate change threatens this crop, affecting yield decreases and subject to uncontrolled pathogen and parasite attacks. Technical advances in precision agriculture are expected to support farmers to more efficiently control the physio-pathological status of crops. Here, we report a straightforward approach to monitoring hazelnut trees in an open field, using aerial multispectral pictures taken by drones. A dataset of 4112 images, each having 2Mpixel resolution per tree and covering RGB, Red Edge, and near-infrared frequencies, was obtained from 185 hazelnut trees located in two different orchards of the Piedmont region (northern Italy). To increase accuracy, and especially to reduce false negatives, the image of each tree was divided into nine quadrants. For each quadrant, nine different vegetation indices (VIs) were computed, and in parallel, each tree quadrant was tagged as “healthy/unhealthy” by visual inspection. Three supervised binary classification algorithms were used to build models capable of predicting the status of the tree quadrant, using the VIs as predictors. Out of the nine VIs considered, only five (GNDVI, GCI, NDREI, NRI, and GI) were good predictors, while NDVI SAVI, RECI, and TCARI were not. Using them, a model accuracy of about 65%, with 13% false negatives was reached in a way that was rather independent of the algorithms, demonstrating that some VIs allow inferring the physio-pathological condition of these trees. These achievements support the use of drone-captured images for performing a rapid, non-destructive physiological characterization of hazelnut trees. This approach offers a sustainable strategy for supporting farmers in their decision-making process during agricultural practices. Full article

Background: The risk of neonatal SARS-CoV-2 infection from the mother’s own milk (MoM) in neonates who are exposed to maternal SARS-CoV-2 during the perinatal period remains unclear. We conducted a systematic review to assess the association between MoM feeding and neonatal SARS-CoV-2 infection in neonates who were born to SARS-CoV-2-positive pregnant persons. Methods: PubMed Central and Google Scholar were searched for studies published by 14 March 2024 that reported neonatal SARS-CoV-2 infection by feeding type. This search, including Scopus, was updated on 17 December 2024. The primary outcome was neonatal SARS-CoV-2 infection. The meta-analysis was conducted using a random effects model with two planned subgroup analyses: time of maternal PCR testing (at admission vs. previous 2 weeks) and dyad handling (isolation vs. some precautions vs. variable/NA). Results: The primary outcome was available in both arms of nine studies, including 5572 neonates who received MoM and 2215 who received no MoM. The GRADE rating was low quality, because the studies were observational (cohorts). The frequency of SARS-CoV-2 infection was similar in both arms (2.7% MoM vs. 2.2% no MoM), with a common risk ratio of 0.82 (95% confidence interval 0.44, 1.53, p = 0.54). No significant differences were observed in the subgroup analyses. Limitations include observational and incomplete data, other possible infection sources, small sample sizes for subgroup analyses, and neonates with more than one feeding type. Conclusions: Feeding MoM was not associated with an increased risk of neonatal SARS-CoV-2 infection among neonates who were born to mothers with perinatal infection. These data, along with reports showing a lack of active replicating SARS-CoV-2 virus in MoM, further support women with perinatal SARS-CoV-2 infection feeding MoM. Registration: PROSPERO ID CRD42021268576. Full article

Person identification is a critical task in applications such as security and surveillance, requiring reliable systems that perform robustly under diverse conditions. This study evaluates the Vision Transformer (ViT) and ResNet34 models across three modalities—RGB, thermal, and depth—using datasets collected with infrared array sensors and LiDAR sensors in controlled scenarios and varying resolutions (16 × 12 to 640 × 480) to explore their effectiveness in person identification. Preprocessing techniques, including YOLO-based cropping, were employed to improve subject isolation. Results show a similar identification performance between the three modalities, in particular in high resolution (i.e., 640 × 480), with RGB image classification reaching 100.0%, depth images reaching 99.54% and thermal images reaching 97.93%. However, upon deeper investigation, thermal images show more robustness and generalizability by maintaining focus on subject-specific features even at low resolutions. In contrast, RGB data performs well at high resolutions but exhibits reliance on background features as resolution decreases. Depth data shows significant degradation at lower resolutions, suffering from scattered attention and artifacts. These findings highlight the importance of modality selection, with thermal imaging emerging as the most reliable. Future work will explore multi-modal integration, advanced preprocessing, and hybrid architectures to enhance model adaptability and address current limitations. This study highlights the potential of thermal imaging and the need for modality-specific strategies in designing robust person identification systems. Full article

This study proposes an advanced plant disease classification framework leveraging the Attention Score-Based Multi-Vision Transformer (Multi-ViT) model. The framework introduces a novel attention mechanism to dynamically prioritize relevant features from multiple leaf images, overcoming the limitations of single-leaf-based diagnoses. Building on the Vision Transformer (ViT) architecture, the Multi-ViT model aggregates diverse feature representations by combining outputs from multiple ViTs, each capturing unique visual patterns. This approach allows for a holistic analysis of spatially distributed symptoms, crucial for accurately diagnosing diseases in trees. Extensive experiments conducted on apple, grape, and tomato leaf disease datasets demonstrate the model’s superior performance, achieving over 99% accuracy and significantly improving F1 scores compared to traditional methods such as ResNet, VGG, and MobileNet. These findings underscore the effectiveness of the proposed model for precise and reliable plant disease classification. Full article

With the increasing demand for accurate multimodal data analysis in complex scenarios, existing models often struggle to effectively capture and fuse information across diverse modalities, especially when data include varying scales and levels of detail. To address these challenges, this study presents an enhanced Swin Transformer V2-based model designed for robust multimodal data processing. The method analyzes urban economic activities and spatial layout using satellite and street view images, with applications in traffic flow and business activity intensity, highlighting its practical significance. The model incorporates a multi-scale feature extraction module into the window attention mechanism, combining local and global window attention with adaptive pooling to achieve comprehensive multi-scale feature fusion and representation. This approach enables the model to effectively capture information at different scales, enhancing its expressiveness in complex scenes. Additionally, a cross-attention-based multimodal feature fusion mechanism integrates spatial structure information from scene graphs with Swin Transformer’s image classification outputs. By calculating similarities and correlations between scene graph embeddings and image classifications, this mechanism dynamically adjusts each modality’s contribution to the fused representation, leveraging complementary information for a more coherent multimodal understanding. Compared with the baseline method, the proposed bimodal model performs superiorly and the accuracy is improved by 3%, reaching 91.5%, which proves its effectiveness in processing and fusing multimodal information. These results highlight the advantages of combining multi-scale feature extraction and cross-modal alignment to improve performance on complex multimodal tasks. Full article

For decades, Alzheimer’s Disease (AD) research has focused on the amyloid cascade hypothesis, which identifies amyloid-beta (Aβ) as the primary driver of the disease. However, the consistent failure of Aβ-targeted therapies to demonstrate efficacy, coupled with significant safety concerns, underscores the need to rethink our approach to AD treatment. Emerging evidence points to microbial infections as environmental factors in AD pathoetiology. Although a definitive causal link remains unestablished, the collective evidence is compelling. This review explores unconventional perspectives and emerging paradigms regarding microbial involvement in AD pathogenesis, emphasizing the gut–brain axis, brain biofilms, the oral microbiome, and viral infections. Transgenic mouse models show that gut microbiota dysregulation precedes brain Aβ accumulation, emphasizing gut–brain signaling pathways. Viral infections like Herpes Simplex Virus Type 1 (HSV-1) and Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) may lead to AD by modulating host processes like the immune system. Aβ peptide’s antimicrobial function as a response to microbial infection might inadvertently promote AD. We discuss potential microbiome-based therapies as promising strategies for managing and potentially preventing AD progression. Fecal microbiota transplantation (FMT) restores gut microbial balance, reduces Aβ accumulation, and improves cognition in preclinical models. Probiotics and prebiotics reduce neuroinflammation and Aβ plaques, while antiviral therapies targeting HSV-1 and vaccines like the shingles vaccine show potential to mitigate AD pathology. Developing effective treatments requires standardized methods to identify and measure microbial infections in AD patients, enabling personalized therapies that address individual microbial contributions to AD pathogenesis. Further research is needed to clarify the interactions between microbes and Aβ, explore bacterial and viral interplay, and understand their broader effects on host processes to translate these insights into clinical interventions. Full article

Addressing groundwater contamination, this study applies machine learning (ML) algorithms to explore the spatiotemporal dynamics of hexavalent chromium (Cr[VI]) at the Hanford 100-Area. The research uses an extensive long-term monitoring dataset focused on groundwater wells and aquifers to enhance the understanding and management strategies of this complex environmental issue and predict the impact on aquifers due to the contamination in groundwater wells. The challenging nature of the task is due to various factors, such as the geological nature of the soil, pipeline leaks, and mobility of the particles that impact the speed of contamination. The findings demonstrate a random forest (ML)-based approach to predict the contaminant distributions accurately, thus significantly reducing uncertainties in contamination assessments and refining conceptual site models. This approach advances groundwater quality management and sets a precedent for future AI-driven environmental studies. Full article

Fires are classified into five types: A, B, C, D, and F/K, according to the components involved in combustion. Recognizing fire classes is critical, since each kind demands a unique suppression approach. Proper fire classification helps to decrease the risk to both life and property. The fuel type is used to determine the fire class, so that the appropriate extinguishing agent can be selected. This study takes advantage of recent advances in deep learning, employing YOLOv11 variants (YOLO11n, YOLO11s, YOLO11m, YOLO11l, and YOLO11x) to classify fires according to their class, assisting in the selection of the correct fire extinguishers for effective fire control. Moreover, a hybrid model that combines YOLO11n and MobileNetV2 is developed for multi-class classification. The dataset used in this study is a combination of five existing public datasets with additional manually annotated images, to create a new dataset covering the five fire classes, which was then validated by a firefighting specialist. The hybrid model exhibits good performance across all classes, achieving particularly high precision, recall, and F1 scores. Its superior performance is especially reflected in the macro average, where it surpasses both YOLO11n and YOLO11m, making it an effective model for datasets with imbalanced classes, such as fire classes. The YOLO11 variants achieved high performance across all classes. YOLO11s exhibited high precision and recall for Class A and Class F, achieving an F1 score of 0.98 for Class A. YOLO11m also performed well, demonstrating strong results in Class A and No Fire with an F1 score of 0.98%. YOLO11n achieved 97% accuracy and excelled in No Fire, while also delivering good recall for Class A. YOLO11l showed excellent recall in challenging classes like Class F, attaining an F1 score of 0.97. YOLO11x, although slightly lower in overall accuracy of 96%, still maintained strong performance in Class A and No Fire, with F1 scores of 0.97 and 0.98, respectively. A similar study employing MobileNetV2 is compared to the hybrid model, and the results show that the hybrid model achieves higher accuracy. Overall, the results demonstrate the high accuracy of the hybrid model, highlighting the potential of the hybrid models and YOLO11n, YOLO11m, YOLO11s, and YOLO11l models for better classification of fire classes. We also discussed the potential of deep learning models, along with their limitations and challenges, particularly with limited datasets in the context of the classification of fire classes. Full article

In this research, a numerical model for simulating dam break floods was developed utilizing ArcGIS 10.8, 3ds Max 2021, and Flow-3D v11.2 software, with the aim of accurately representing the dam break disaster at Tianchi Lake in Changbai Mountain. The study involved the construction of a Triangulated Irregular Network (TIN) terrain surface and the application of 3ds Max 2021 to enhance the precision of the three-dimensional terrain data, thereby optimizing the depiction of the region’s topography. The finite volume method, along with multi-block grid technology, was employed to model the dam break scenario at Tianchi Lake. To evaluate the severity of the dam break disaster, the research integrated land use classifications within the study area with the simulated flood depths resulting from the dam break, applying the natural breaks method for hazard level classification. The findings indicated that the computational fluid dynamics (CFD) numerical model developed in this study significantly enhanced both the efficiency and accuracy of the simulations. Furthermore, the disaster assessment methodology that incorporated land use types facilitated the generation of inundation maps and disaster zoning maps across two scenarios, thereby effectively assessing the impacts of the disaster under varying conditions. Full article

Objectives: Accurate kidney and tumor segmentation of computed tomography (CT) scans is vital for diagnosis and treatment, but manual methods are time-consuming and inconsistent, highlighting the value of AI automation. This study develops a fully automated AI model using vision transformers (ViTs) and convolutional neural networks (CNNs) to detect and segment kidneys and kidney tumors in Contrast-Enhanced (CECT) scans, with a focus on improving sensitivity for small, indistinct tumors. Methods: The segmentation framework employs a ViT-based model for the kidney organ, followed by a 3D UNet model with enhanced connections and attention mechanisms for tumor detection and segmentation. Two CECT datasets were used: a public dataset (KiTS23: 489 scans) and a private institutional dataset (Private: 592 scans). The AI model was trained on 389 public scans, with validation performed on the remaining 100 scans and external validation performed on all 592 private scans. Tumors were categorized by TNM staging as small (≤4 cm) (KiTS23: 54%, Private: 41%), medium (>4 cm to ≤7 cm) (KiTS23: 24%, Private: 35%), and large (>7 cm) (KiTS23: 22%, Private: 24%) for detailed evaluation. Results: Kidney and kidney tumor segmentations were evaluated against manual annotations as the reference standard. The model achieved a Dice score of 0.97 ± 0.02 for kidney organ segmentation. For tumor detection and segmentation on the KiTS23 dataset, the sensitivities and average false-positive rates per patient were as follows: 0.90 and 0.23 for small tumors, 1.0 and 0.08 for medium tumors, and 0.96 and 0.04 for large tumors. The corresponding Dice scores were 0.84 ± 0.11, 0.89 ± 0.07, and 0.91 ± 0.06, respectively. External validation on the private data confirmed the model’s effectiveness, achieving the following sensitivities and average false-positive rates per patient: 0.89 and 0.15 for small tumors, 0.99 and 0.03 for medium tumors, and 1.0 and 0.01 for large tumors. The corresponding Dice scores were 0.84 ± 0.08, 0.89 ± 0.08, and 0.92 ± 0.06. Conclusions: The proposed model demonstrates consistent and robust performance in segmenting kidneys and kidney tumors of various sizes, with effective generalization to unseen data. This underscores the model’s significant potential for clinical integration, offering enhanced diagnostic precision and reliability in radiological assessments. Full article

Additive manufacturing (AM) processes, such as fused deposition modelling (FDM), have emerged as transformative technologies in pharmaceutical manufacturing, enabling the production of drug delivery systems with complex and customised geometries. These advancements provide precise control over drug release profiles and facilitate the development of patient-specific medicines. This study investigates the dissolution behaviour of AM-fabricated tablets made from polyvinyl alcohol (PVA), a hydrophilic and biocompatible polymer widely used in drug delivery systems. The influence of the initial mass, surface area, and surface-area-to-volume ratio (S/V) on dissolution kinetics is evaluated for tablets with intricate geometries. Our findings demonstrate that these parameters, while critical for conventional tablet shapes, are insufficient to fully predict the dissolution behaviour of complex geometries. Furthermore, this study highlights how geometric modifications can enable the administration of the same drug dosage through sustained or immediate release profiles, offering enhanced versatility in drug delivery. By leveraging the geometric design freedom provided by AM technologies, this research underscores the potential for optimising drug delivery systems to improve therapeutic outcomes and patient compliance. Full article

When picking strawberries outdoors, due to factors such as light changes, obstacle occlusion, and small target detection objects, the phenomena of poor strawberry recognition accuracy and low recognition rate are caused. An improved YOLOv5n strawberry high-precision recognition algorithm is proposed. The algorithm uses FasterNet to replace the original YOLOv5n backbone network and improves the detection rate. The MobileViT attention mechanism module is added to improve the feature extraction ability of small target objects so that the model has higher detection accuracy and smaller module sizes. The CBAM hybrid attention module and C2f module are introduced to improve the feature expression ability of the neural network, enrich the gradient flow information, and improve the performance and accuracy of the model. The SPPELAN module is added as well to improve the model’s detection efficiency for small objects. The experimental results show that the detection accuracy of the improved model is 98.94%, the recall rate is 99.12%, the model volume is 53.22 MB, and the mAP value is 99.43%. Compared with the original YOLOv5n, the detection accuracy increased by 14.68%, and the recall rate increased by 11.37%. This technology has effectively accomplished the accurate detection and identification of strawberries under complex outdoor conditions and provided a theoretical basis for accurate outdoor identification and precise picking technology. Full article

DC micro-grids are emerging as a promising solution for efficiently integrating renewable energy into power systems. These systems offer increased flexibility and enhanced energy management, making them ideal for applications such as heat pump (HP) systems. However, the integration of intermittent renewable energy sources with optimal energy management in these micro-grids poses significant challenges. This paper proposes a novel control strategy designed specifically to improve the performance of DC micro-grids. The strategy enhances energy management by leveraging an environmental mission profile that includes real-time measurements for energy generation and heat pump performance evaluation. This micro-grid application for heat pumps integrates photovoltaic (PV) systems, wind generators (WGs), DC-DC converters, and battery energy storage (BS) systems. The proposed control strategy employs an intelligent maximum power point tracking (MPPT) approach that uses optimization algorithms to finely adjust interactions among the subsystems, including renewable energy sources, storage batteries, and the load (heat pump). The main objective of this strategy is to maximize energy production, improve system stability, and reduce operating costs. To achieve this, it considers factors such as heating and cooling demand, power fluctuations from renewable sources, and the MPPT requirements of the PV system. Simulations over one year, based on real meteorological data (average irradiance of 500 W/m², average annual wind speed of 5 m/s, temperatures between 2 and 27 °C), and carried out with Matlab/Simulink R2022a, have shown that the proposed model predictive control (MPC) strategy significantly improves the performance of DC micro-grids, particularly for heat pump applications. This strategy ensures a stable DC bus voltage (±1% around 500 V) and maintains the state of charge (SoC) of batteries between 40% and 78%, extending their service life by 20%. Compared with conventional methods, it improves energy efficiency by 15%, reduces operating costs by 30%, and cuts CO₂; emissions by 25%. By incorporating this control strategy, DC micro-grids offer a sustainable and reliable solution for heat pump applications, contributing to the transition towards a cleaner and more resilient energy system. This approach also opens new possibilities for renewable energy integration into power grids, providing intelligent and efficient energy management at the local level. Full article

(1) Background: Volumetric modulated arc therapy (VMAT) can deliver more accurate dose distribution and reduce radiotherapy-induced toxicities for postoperative cervical and endometrial cancer. This study aims to retrospectively analyze the relationship between dosimetric parameters of organs at risk (OARs) and acute toxicities and provide suggestions for the dose constraints. (2) Methods: A total of 164 postoperative cervical and endometrial cancer patients were retrospectively analyzed, and the endpoints were grade ≥ 2 acute urinary toxicity (AUT) and acute lower gastrointestinal toxicity (ALGIT). The normal tissue complication probability (NTCP) model was established using the logistic regression model. Restricted cubic spline (RCS) curves were used to explore the association between dosimetric parameters and toxicities. The receiver operating characteristic (ROC) curve, calibration curve, Akaike’s corrected information criterion (AICc), decision curve analysis (DCA), and clinical impact curve (CIC) were analyzed to evaluate the performance of NTCP models. (3) Results: Bladder V_40Gy was identified to develop the NTCP model of AUT, and the mean AUC was 0.69 (CI: 0.58–0.80). Three candidate predictors, namely the small intestine V_30Gy, colon D_45%, and rectum D_55%, were identified to develop the NTCP model of ALGIT, and the mean AUC was 0.71 (CI: 0.61–0.80). Both models were considered to have relatively good discriminative accuracy and could provide a high net benefit in clinical applications. (4) Conclusions: We developed NTCP models to predict the probability for grade ≥ 2 AUT and ALGIT. We recommend that bladder V_40Gy, the small intestine V_30Gy, colon D_45%, and rectum D_55% be controlled below 42%, 20.4%, 16.9 Gy, and 32.0 Gy, respectively. Full article