课程文献中心 更多>>

关键词： GSRCA   Interpretability analysis   Raman spectroscopy classification   Spectral region contribution quantification  

摘要： To address the fragmented point-wise gradients and limited physical interpretability of standard Grad-CAM in spectral analysis, this study proposes GSRCA (Grad-CAM-based Spectral Region and Contribution Analysis). Its core innovation lies in an adaptive region segmentation strategy that aggregates fragmented gradient points into continuous intervals corresponding to molecular vibrational modes. Based on a CNN framework, GSRCA first computes class-specific gradients via Grad-CAM, then employs a weight-distribution-based adaptive region segmentation strategy to aggregate discrete gradient responses into physically meaningful continuous spectral regions. Finally, the relative contribution of each region to the classification result is quantified through gradient integration and expressed as a percentage.On a Raman spectroscopy dataset containing six categories of illicit drugs, the key regions identified by GSRCA are generally consistent with characteristic vibrational bands reported in the literature, and region deletion/insertion tests further validate the effectiveness of the proposed contribution quantification scheme. Compared with conventional Grad-CAM, LIME, and SHAP-based methods such as GRASS, GSRCA shows better performance in regional integrity, physical interpretability, and computational efficiency. Experiments also show that the method requires no architecture-specific parameter tuning and can be transferred to dilated CNNs, Transformers, and remote sensing hyperspectral data, demonstrating model compatibility and generalization capability. GSRCA provides a physically interpretable and versatile explanation method for spectral classification models.

关键词： 3D printing   Biomechanics   CAD   Finite element analysis (FEA)   Parametric model   Simulation  

摘要： Purpose

关键词： Cancer survivorship   Complementary and alternative medicine   Chinese immigrants   Information practices   Information encountering   Trust   Credibility   Cognitive authority   Fractured information landscapes  

摘要： Purpose

关键词： Coronary artery disease   FV   Monocyte   Residual inflammation   Tissue factor  

摘要： The underlying proinflammatory mechanism of coagulation factor VⅡ in coronary artery disease (CAD) remains unclear. According to the inclusion criteria, 24 non-CAD subjects and 24 patients with CAD were enrolled. The expression levels of pro-inflammatory (CD14⁺CD86⁺, CD14⁺CD163^-), anti-inflammatory monocytes (CD14⁺CD163⁺), CD14⁺CD142⁺ and CD14⁺PAR2⁺ monocytes in peripheral blood were detected by flow cytometry. The levels of TNF-α, IL-1β, IL-6, IL-10, and FVⅡ in plasma were determined by ELISA. In vitro experiments were conducted to induce inflammatory THP-1 by FVⅡa. The FVⅡa/TF complex inhibitor PCI-27,483 and PAR2 antagonist AZ3451 were used to clarify the mechanisms. Compared with the control group, the percentage of CD14⁺CD86⁺ pro-inflammatory monocytes were significantly increased in the CAD group, and further elevated after PCI. The results of the levels of CD14⁺CD142⁺ and CD14⁺PAR2⁺ monocytes showed the same trends with CD14⁺CD86⁺ pro-inflammatory monocytes. The levels of CD14⁺CD163⁺ anti-inflammatory monocytes and IL-10 were decreased in CAD patients, but increased after PCI. The pro-inflammatory factors and FVⅡ were significantly elevated in the CAD group and significantly decreased after PCI. Correlation analysis revealed that FVⅡ, TF, and PAR2 were significantly associated with CD14⁺CD86⁺ pro-inflammatory monocytes, and inflammatory factors. In vitro studies further confirmed that the FVⅡa/TF coagulation complex and PAR2 could activate NF-κB in monocytes and aggravate inflammation. Our findings suggest that crosstalk between monocytes and the coagulation complex might contribute to residual inflammation in CAD, potentially involving the FVIIa/TF/PAR2-NF-κB signaling pathway. Based on these observations, we hypothesize

关键词： complementary and alternative medicine   herd health management   dairy cows   qualitative research   document analysis  

摘要： The use of complementary and alternative medicines (CAM) in UK dairy farming has sparked considerable debate. The Royal College of Veterinary Surgeons (RCVS) released a position statement in 2017, emphasizing that treatments should be underpinned by evidence. Since then, there are ongoing debates around CAM use in farming media. We conducted a document analysis to examine how these debates may influence how CAM is used on UK dairy farms.

关键词： Additive manufacturing   Accuracy   Lithium disilicate   CAD CAM  

摘要： This study compared subtractive and additive fabrication methods regarding their impact on the marginal/internal fit, trueness, and precision of lithium disilicate ceramic veneers.

关键词： Reverse engineering  

摘要： Reverse engineering of degraded service components becomes difficult if there is uncertainty in the nominal geometry due to corrosion, fracture or missing material, and original drawings are not available. This study proposes a two-stage dimensional validation procedure for a worn cast-iron foot valve, which can be broken down into reconstruction-stage deviation and fabrication-stage deviation. The valve was scanned with structured light and converted to a watertight mesh of around 3.1 million facets. The damaged geometry was recreated in the CAD environment using a combination of NURBS surface fitting, symmetry guided repair and controlled fitting of missing or corroded geometry. A PLA prototype was then fabricated by fused deposition modelling using 0.15 mm layer height, 100% infill, and 45 mm/s print speed. Dimensional accuracy was evaluated using signed deviation mapping, rigid alignment, and a ± 2 mm tolerance band. In Stage 1, the CAD-scan comparison produced an RMS deviation of 2.67 mm, with 36.5% of the surface lying within the tolerance band. In Stage 2, the CAD-printed-part comparison showed an RMS deviation of 2.9 mm and 35% in-tolerance surface. The small RMS increase of + 0.23 mm indicates that, under the selected printing conditions, FDM reproduction introduced only a limited additional global deviation relative to the CAD-scan deviation. The main contribution of this work is the proposed two-stage error-separation strategy, which provides traceable evidence of dimensional error propagation across scanning, CAD repair, and physical prototyping. The workflow is suitable for preliminary geometric prototyping of degraded cast components, while functional reuse requires further datum-based and feature-level validation.

关键词： Endoscopy  

摘要： Artificial Intelligence (AI), particularly in the domains of computer vision and deep learning (DL), is increasingly being integrated into medical diagnostics. Although it has not yet entirely transformed clinical practice, its role in medical image analysis is rapidly expanding, offering the potential for more accurate, efficient, and accessible diagnostic capabilities that contribute to improved patient outcomes. Currently, DL techniques are employed in a variety of medical imaging tasks, such as anomaly detection, pathology classification, and anatomical structure segmentation across modalities, including endoscopy, MRI, and CT scans. Despite these advancements, the widespread clinical adoption of DL models remains constrained by several limitations, including the opacity of model decision-making, data imbalance, and lack of interpretability. To address these challenges, this study proposes an explainable knowledge distillation (KD) framework that integrates Gradient-weighted Class Activation Mapping (Grad-CAM) with a ResNet-based teacher-student architecture. The proposed methodology was evaluated on a class-imbalanced dataset comprising ten categories of gastrointestinal (GI) abnormalities derived from the Capsule Vision 2024 Challenge. The resulting student model achieved a classification accuracy of 96.52% and an F1-score of 96.37%, thereby surpassing previously reported methods.

关键词： Inspection  

摘要： Deep learning-based autonomous inspection of power grid insulators is challenged by data imbalance and model opacity. This paper presents an end-to-end solution integrating advanced data synthesis, detection, classification, and explainability. First, a conditional diffusion model generates realistic synthetic fault images to balance the dataset. A two-stage architecture based on You Only Look Once version 26 (YOLO26) extra-large and Shifted windows (Swin)-V2-B, called YOLO26-Swin, fine-tuned with Bayesian optimization, performs robust insulator detection and then fault classification. Finally, a novel SHapley Additive exPlanations with Class Activation Mapping (SHAP-CAM) method provides intuitive visual explanations for model predictions. Extensive experiments validate our framework’s superiority: it achieves an F1-score of 0.98149 and a mean Average Precision (mAP)@[0.5] of 0.98951, exceeding leading detection and classification models. This work highlights the efficacy of diffusion models for data augmentation in critical infrastructure and advances the interpretability of vision-based inspection systems.

关键词： Artificial intelligence  

摘要： Gender recognition is a significant subfield in artificial intelligence and can be used in many applications, for example: security systems, suggestions, and biometric identification. This work evaluates the performance of MobileNet and ViT architectures for gender classification using real world and synthetic datasets. From the experimental results, we find that the overall performance of MobileNet is outstanding in real-life applications with the highest accuracy, recall, and F1 score due to feature extraction with its lightweight architecture. On the other hand, ViT is powerful, but it needs a lot of training data to fine-tune its performance. Various Explainable AI (XAI) techniques, including Grad CAM and LIME, were utilized to improve model interpretability by shedding light on the rationale behind the decision-making process and highlighting the most salient features that contribute to the classification process. These results indicate that MobileNet is suitable for real-time applications, while ViT shows potential for optimization through fine tuning and additional data. Future studies should also investigate hybrid models, strategies for mitigating bias, and improved explainability for fairness and reliability in gender classification systems.