Search Results (12)

In object detection, boundary blurring caused by occlusion and background interference often hinders effective feature extraction. To address this challenge, we propose Edge Aware-YOLO, a novel framework designed to enhance edge awareness and efficient feature fusion. Our method integrates three key contributions. First, the Variable Sobel Compact Inverted Block (VSCIB) employs convolution kernels with adjustable orientation and size, enabling robust multi-scale edge adaptation. Second, the Spatial Pyramid Shared Convolution (SPSC) replaces standard pooling with shared dilated convolutions, minimizing detail loss during feature reconstruction. Finally, the Efficient Downsampling Convolution (EDC) utilizes a dual-branch architecture to balance channel compression with semantic preservation. Extensive evaluations on public datasets demonstrate that Edge Aware-YOLO significantly outperforms state-of-the-art models. On MS COCO, it achieves 56.3% mAP₅₀ and 40.5% mAP_50–95 (gains of 1.5% and 1.0%) with only 2.4M parameters and 5.8 GFLOPs, surpassing advanced models like YOLOv11. Full article

Background: Pregnancy induces substantial cardiovascular remodeling, yet whether maternal cardiac adaptation differs across ethnic groups remains unclear. Body size, ventricular geometry, and thoracoabdominal configuration may modulate key functional indices such as left ventricular ejection fraction (LVEF) and global longitudinal strain (LV-GLS). This systematic review compared echocardiographic characteristics between Asian and Western healthy pregnant women in late gestation and explored physiological mechanisms underlying observed differences. Methods: A comprehensive search of PubMed, Scopus, and EMBASE identified studies reporting transthoracic echocardiography in healthy singleton third-trimester pregnancies across Asian and Western populations. Extracted variables included anthropometry, ventricular dimensions and volumes, LVEF, and LV-GLS. Pooled estimates were calculated using inverse-variance weighting, with heterogeneity quantified using the I² statistic. Study quality was assessed with the NIH Case–Control Quality Assessment Tool. Comparative forest plots visualized population differences. Results: Twenty studies involving 1431 participants (578 Asian and 853 Western women) met inclusion criteria. Asian women consistently exhibited smaller ventricular chambers, higher LVEF, and more favorable LV-GLS. Importantly, these differences persisted after indexing LV-GLS to BSA, indicating that body-size normalization attenuates—but does not eliminate—population differences in myocardial deformation. Western women demonstrated slightly attenuated GLS despite preserved LVEF, plausibly attributable to larger cardiac size, higher wall stress, greater diaphragmatic elevation, and increased extrinsic thoracic compression. Between-study heterogeneity was substantial (I² > 95%) due to variation in imaging platforms, strain software, and population characteristics. Methodological quality was fair, with frequent lack of sample-size justification and incomplete confounder adjustment. Conclusions: Healthy Asian pregnant women display a hyperdynamic systolic phenotype, whereas Western women show a physiologically appropriate, load-related attenuation of LV-GLS with preserved LVEF. These findings highlight the need for ethnicity-associated and anatomy-aware echocardiographic reference values and support incorporating thoracic geometric indices, such as the modified Haller Index, into strain interpretation during pregnancy. Full article

Chili pepper, a widely cultivated and consumed crop, faces challenges in accurately determining maturity due to issues such as occlusion, small target size, and similarity between fruit color and background. This study presents an enhanced YOLOv8n-based object detection model, YOLO-ALW, designed to address these challenges. The model introduces the AKConv (Alterable Kernel Convolution) module in the head section, which adaptively adjusts the convolution kernel shape and size based on the target and scene, improving detection performance under occlusion and dense environments. In the backbone, the SPPF_LSKA (Spatial Pyramid Pooling Fast-Large Separable Kernel Attention) module enhances the integration of multi-scale features, facilitating accurate differentiation of peppers at various maturity stages while maintaining low computational complexity. Additionally, the Wise-IoU (Wise Intersection over Union) loss function optimizes bounding box learning, further improving the detection of peppers in occluded or background-similar scenarios. Experimental results demonstrate that YOLO-ALW achieves a mean average precision (mAP_0.5) of 99.1%, with precision and recall rates of 98.3% and 97.8%, respectively, outperforming the original YOLOv8n by 3.4%, 5.1%, and 9.0%, respectively. Grad-CAM feature visualization highlights the model’s improved focus on key fruit features. YOLO-ALW shows significant promise for high-precision chili pepper detection and maturity recognition, offering valuable support for automated harvesting applications. Full article

Aiming at many challenges in the recognition task of peach leaf shrink disease, such as the diversity of object size of diseased leaf disease, complex background interference, and inflexible adjustment of model training learning rate, we propose a peach leaf shrink disease recognition algorithm based on an attention generalized efficient layer aggregation network. Firstly, the rectified linear unit activation function is used to effectively improve the stability and performance of the model in low-precision computing environments and solve the problem of partial gradient disappearance. Secondly, the integrated squeeze-and-excitation network attention mechanism can adaptively focus on the key areas of pests and diseases in the image, which significantly enhances the recognition ability of the model to the characteristics of pests and diseases. Finally, combined with fast pyramid pooling enhanced with Local Attention Networks, the deep fusion of cross-layer features is realized to improve the ability of the model to identify complex features and optimize the operation efficiency. The experimental results on the peach leaf shrink disease recognition dataset show that the proposed algorithm achieves a significant improvement in performance compared with the original YOLOv8 algorithm. Specifically, mF1, mPrecision, mRecall, and mAP increased by 0.1075, 0.0723, 0.1224, and 0.1184, respectively, which provided strong technical support for intelligent and automatic monitoring of peach pests and diseases. Full article

Bovine respiratory syncytial virus (BRSV) is an economically important pathogen of cattle and contributes to the bovine respiratory disease complex (BRDC). Despite individual studies investigating BRSV prevalence, risk factors, and detection methodologies, a systematic review and meta-analysis have been lacking. The aim of the current study was to conduct a systematic review and meta-analysis to determine the prevalence and detection rate of BRSV and identify associated risk factors. Additionally, the study aimed to explore the variability in BRSV prevalence based on different detection methods and associated risk factors. Adhering to PRISMA guidelines, data from three databases—Web of Science, PubMed, and Scopus—were systematically retrieved, screened and extracted. Out of 2790 initial studies, 110 met the inclusion criteria. The study found that prevalence and detection rates varied based on the detection methods used (antibody, antigen, and nucleic acid), study populations, production systems, and geographic locations. Findings were reported as a pooled proportion. The pooled proportion, hereafter referred to as prevalence or detection rate, was determined by calculating the ratio of cattle that tested positive for BRSV to the total number of cattle tested. Key findings include a pooled prevalence of 0.62 for antibody-based methods, 0.05 for antigen-based methods, and 0.09 (adjusted to 0.03) for nucleic acid-based methods. Detection rates in BRDC cases also varied, with antibody methods showing a rate of 0.34, antigen methods 0.16, and nucleic acid methods 0.13. The certainty of evidence of the meta-analysis results, assessed using GRADE, was moderate for antibody detection methods and low for antigen and nucleic acid methods. The study identified significant risk factors and trends affecting BRSV prevalence, such as geographical location, herd size, age, and co-infections. The results of the current study showed the complexity of understanding BRSV prevalence in different settings. The variability in BRSV prevalence based on detection methods and associated risk factors, such as geographic location and herd size, highlights the need for tailored approaches to detect and manage BRSV accurately. Full article

In metal additive manufacturing (AM), precise temperature field prediction is crucial for process monitoring, automation, control, and optimization. Traditional methods, primarily offline and data-driven, struggle with adapting to real-time changes and new process scenarios, which limits their applicability for effective AM process control. To address these challenges, this paper introduces the first physics-informed (PI) online learning framework specifically designed for temperature prediction in metal AM. Utilizing a physics-informed neural network (PINN), this framework integrates a neural network architecture with physics-informed inputs and loss functions. Pretrained on a known process to establish a baseline, the PINN transitions to an online learning phase, dynamically updating its weights in response to new, unseen data. This adaptation allows the model to continuously refine its predictions in real-time. By integrating physics-informed components, the PINN leverages prior knowledge about the manufacturing processes, enabling rapid adjustments to process parameters, geometries, deposition patterns, and materials. Empirical results confirm the robust performance of this PI online learning framework in accurately predicting temperature fields for unseen processes across various conditions. It notably surpasses traditional data-driven models, especially in critical areas like the Heat Affected Zone (HAZ) and melt pool. The PINN’s use of physical laws and prior knowledge not only provides a significant advantage over conventional models but also ensures more accurate predictions under diverse conditions. Furthermore, our analysis of key hyperparameters—the learning rate and batch size of the online learning phase—highlights their roles in optimizing the learning process and enhancing the framework’s overall effectiveness. This approach demonstrates significant potential to improve the online control and optimization of metal AM processes. Full article

The occurrence of maize diseases is frequent but challenging to manage. Traditional identification methods have low accuracy and complex model structures with numerous parameters, making them difficult to implement on mobile devices. To address these challenges, this paper proposes a corn leaf disease recognition model SNMPF based on convolutional neural network ShuffleNetV2. In the down-sampling module of the ShuffleNet model, the max pooling layer replaces the deep convolutional layer to perform down-sampling. This improvement helps to extract key features from images, reduce the overfitting of the model, and improve the model’s generalization ability. In addition, to enhance the model’s ability to express features in complex backgrounds, the Sim AM attention mechanism was introduced. This mechanism enables the model to adaptively adjust focus and pay more attention to local discriminative features. The results on a maize disease image dataset demonstrate that the SNMPF model achieves a recognition accuracy of 98.40%, representing a 4.1 percentage point improvement over the original model, while its size is only 1.56 MB. Compared with existing convolutional neural network models such as EfficientNet, MobileViT, EfficientNetV2, RegNet, and DenseNet, this model offers higher accuracy and a more compact size. As a result, it can automatically detect and classify maize leaf diseases under natural field conditions, boasting high-precision recognition capabilities. Its accurate identification results provide scientific guidance for preventing corn leaf disease and promote the development of precision agriculture. Full article

Dynamic monitoring of cropland using high spatial resolution remote sensing images is a powerful means to protect cropland resources. However, when a change detection method based on a convolutional neural network employs a large number of convolution and pooling operations to mine the deep features of cropland, the accumulation of irrelevant features and the loss of key features will lead to poor detection results. To effectively solve this problem, a novel cropland change detection network (CroplandCDNet) is proposed in this paper; this network combines an adaptive receptive field and multiscale feature transmission fusion to achieve accurate detection of cropland change information. CroplandCDNet first effectively extracts the multiscale features of cropland from bitemporal remote sensing images through the feature extraction module and subsequently embeds the receptive field adaptive SK attention (SKA) module to emphasize cropland change. Moreover, the SKA module effectively uses spatial context information for the dynamic adjustment of the convolution kernel size of cropland features at different scales. Finally, multiscale features and difference features are transmitted and fused layer by layer to obtain the content of cropland change. In the experiments, the proposed method is compared with six advanced change detection methods using the cropland change detection dataset (CLCD). The experimental results show that CroplandCDNet achieves the best F1 and OA at 76.04% and 94.47%, respectively. Its precision and recall are second best of all models at 76.46% and 75.63%, respectively. Moreover, a generalization experiment was carried out using the Jilin-1 dataset, which effectively verified the reliability of CroplandCDNet in cropland change detection. Full article

Quantum key distribution (QKD) is an encrypted communication technique based on the principles of quantum mechanics that ensures communication security by exploiting the properties of quantum states. Currently, the transmission efficiency of the QKD system is low. Trusted relay technology is used to solve this problem and achieve long-distance transmission. However, trusted relaying alone cannot decrypt the issues of poor link stability and the low utilization of key resources. To further optimize the system performance, we propose a dynamic routing algorithm. One of the improvement schemes includes the following: firstly, an adjustable-size quantum key pool (QKP) is designed, which can dynamically adjust the size of the refreshing pool according to the actual demand. Secondly, the utilization of key resources is improved by using the residual quantum key model to dynamically obtain the remaining key amount in the QKP and set the key amount threshold. We calculate the link-blocking probability and track the blocking intensity and blocking entry by combining the Poisson process, thus realizing the evaluation of the link stability. Finally, the number of remaining keys in the QKP and the link-blocking probability combine with the random wandering model as the basis of the route selection for the QKD dynamic routing algorithm to achieve efficient key path selection. We validated the algorithm by comparing it with other algorithms on the Mininet simulation platform, and the algorithm proved to have a better performance in terms of congestion avoidance, delay reduction, and improved QKD efficiency. This scheme provides a novel and efficient way to solve the problems in existing QKD systems. It effectively improves the transmission efficiency and strengthens the system’s security by dynamically obtaining the critical volume, accurately evaluating the link state, and selecting the optimal critical path. Full article

Background: Hepatitis C virus (HCV) is still common in patients with chronic kidney disease. It has been recently discovered that chronic HCV is a risk factor for increased incidence of CKD in the adult general population. According to a systematic review with a meta-analysis of clinical studies, pooling results of longitudinal studies (n = 2,299,134 unique patients) demonstrated an association between positive anti-HCV serologic status and increased incidence of CKD; the summary estimate for adjusted HR across the surveys was 1.54 (95% CI, 1.26; 1.87), (p < 0.0001). The introduction of direct-acting antiviral drugs (DAAs) has caused a paradigm shift in the management of HCV infection; recent guidelines recommend pan-genotypic drugs (i.e., drugs effective on all HCV genotypes) as the first-choice therapy for HCV, and these promise to be effective and safe even in the context of chronic kidney disease. Aim: The purpose of this narrative review is to show the most important data on pan-genotypic DAAs in advanced CKD (CKD stage 4/5). Methods: We recruited studies by electronic databases and grey literature. Numerous key-words (‘Hepatitis C’ AND ‘Chronic kidney disease’ AND ‘Pan-genotypic agents’, among others) were adopted. Results: The most important pan-genotypic combinations for HCV in advanced CKD are glecaprevir/pibrentasvir (GLE/PIB) and sofosbuvir/velpatasvir (SOF/VEL). Two clinical trials (EXPEDITION-4 and EXPEDITION-5) and some ‘real-world’ studies (n = 6) reported that GLE/PIB combinations in CKD stage 4/5 gave SVR12 rates ranging between 86 and 99%. We retrieved clinical trials (n = 1) and ‘real life’ studies (n = 6) showing the performance of SOF/VEL; according to our pooled analysis, the summary estimate of SVR rate was 100% in studies adopting SOF/VEL antiviral combinations. The drop-out rate (due to AEs) in patients on SOF/VEL ranged between 0 and 4.8%. Conclusions: Pan-genotypic combinations, such as GLE/PIB and SOF/VEL, appear effective and safe for HCV in advanced CKD, even if a limited number of studies with small sample sizes currently exist on this issue. Studies are under way to assess whether successful antiviral therapy with DAAs will translate into better survival in patients with advanced CKD. Full article

With the application of Selective Laser Melting (SLM) technology becoming more and more widespread, it is important to note the process parameters that have a very important effect on the forming quality. Key process parameters such as laser power (P), scan speed (s), and scanning strategy (μ) were investigated by determining the correlation between the microstructure and residual stress in this paper. A total of 10 group 316L specimens were fabricated using SLM for comprehensive analysis. The results show that the key process parameters directly affect the morphology and size of the molten pool in the SLM deposition, and the big molten pool width has a direct effect on the larger grain size and crystal orientation distribution. In addition, the larger grain size and misorientation angle also affect the size of the residual stress. Therefore, better additive manufacturing grain crystallization can be obtained by reasonably adjusting the process parameter combinations. The transfer energy density can synthesize the influence of four key process parameters (P, v, the hatching distance (δ), and the layer thickness (h)). In this study, it is proposed that the accepted energy density will reflect the influence of five key process parameters, including the scanning trajectory (μ), which can reflect the comprehensive effect of process parameters more accurately. Full article

Cells polarize for growth, motion, or mating through regulation of membrane-bound small GTPases between active GTP-bound and inactive GDP-bound forms. Activators (GEFs, GTP exchange factors) and inhibitors (GAPs, GTPase activating proteins) provide positive and negative feedbacks. We show that a reaction–diffusion model on a curved surface accounts for key features of polarization of model organism fission yeast. The model implements Cdc42 membrane diffusion using measured values for diffusion coefficients and dissociation rates and assumes a limiting GEF pool (proteins Gef1 and Scd1), as in prior models for budding yeast. The model includes two types of GAPs, one representing tip-localized GAPs, such as Rga3; and one representing side-localized GAPs, such as Rga4 and Rga6, that we assume switch between fast and slow diffusing states. After adjustment of unknown rate constants, the model reproduces active Cdc42 zones at cell tips and the pattern of GEF and GAP localization at cell tips and sides. The model reproduces observed tip-to-tip oscillations with periods of the order of several minutes, as well as asymmetric to symmetric oscillations transitions (corresponding to NETO “new end take off”), assuming the limiting GEF amount increases with cell size. Full article