课程文献中心 更多>>

关键词： Computer-Aided Design (CAD)   Human-Computer Interaction (HCI)   Interactive design   Large Language Models (LLMs)   Parametric modeling   Programming-based CAD  

摘要： Recent advances in generative Artificial Intelligence (AI)—particularly Large Language Models (LLMs)—offer a new paradigm for CAD interaction by enabling natural and intuitive input through texts, images, and context-aware selections. In this study, we present CADialogue, a multimodal LLM-powered conversational assistant to enable intuitive parametric CAD modeling through natural language, speech, image, and selection-based geometry interactions. Built on a general-purpose large language model, CADialogue translates user prompts into executable code to support geometry creation and context-aware editing. The system features a modular architecture that decouples prompt handling, refinement logic, and execution—allowing seamless model replacement as LLMs develop—and includes caching for rapid reuse of validated designs. We evaluate the system on 70 modeling and 10 editing tasks across varying difficulty levels, assessing performance in terms of accuracy, refinement behavior, and execution time. Results show an overall success rate of 95.71%, combining a 91.43% baseline under Text-Only input with additional recoveries enabled by Text + Image input, with robust recovery from failure via self-correction and human-in-the-loop refinement. Comparative analysis reveals that image input improves success in semantically complex prompts but introduces additional processing time. Furthermore, caching confirmed macros yields over 85.71% speedup in repeated executions. These findings highlight the potential of general-purpose LLMs for enabling accessible, iterative, and accurate CAD modeling workflows without domain-specific fine-tuning. The source code and dataset for CADialogue are available at https://***/Hiram31/CADialogue

来源 ScienceDirectJournal 详细信息

关键词： 3D printing   CAD/CAM post   Relined fibre post   Customised post   Internal adaptation   Micro-CT   Push-out bond strength  

摘要： This study assessed the internal adaptation and push-out bond strength (PBS) of customised CAD/CAM 3D-printed (Varseosmile Crown Plus, Bego, Germany - 3DP) and milled (Enamic, Vita, Germany - MP) resin-based posts vs prefabricated fiber posts relined with resin composite (RelyX fiber post and Filtek Bulkfil Composite, 3M Health Care, USA - RFP).

关键词： Complementary and alternative medicine   CAM   depoliticisation   political participation   voting   survey  

摘要： The increasing cultural prevalence and appeal of complementary and alternative medicine (CAM) have raised concerns about its potential depoliticising effects. Due to its focus on the self, CAM is believed to cultivate individualism and orient transformative energies towards self-improvement rather than the collective struggle for social justice. However, despite these concerns, few quantitative studies have examined CAM from the perspective of the depoliticisation thesis. This article aims to address this gap by studying the association between political participation and CAM with data from the European Social Survey (ESS). These data enable us to capture a diverse range of CAM practices and to scrutinise both institutionalised forms of political participation, such as voting, and non-institutionalised participation taking place in civil society. Contrary to the depoliticisation thesis, our results show that CAM users either engage more actively in voting or do not significantly differ in their voting behaviour compared to non-users, depending on the specific CAM modalities considered. Moreover, CAM users participate more actively in non-institutionalised activities than non-users across all types of CAM modalities. This article offers valuable insights into the relationship between CAM and political engagement and challenges prevailing assumptions about the depoliticising effects of CAM.

关键词： breast lesions   classification   deep learning   dynamic tissue stimulation   grading   RFTSDP Method   ultrasound RF time series  

摘要： Objectives Accurate detection of breast lesion type is crucial for optimizing treatment;however, due to the limited precision of current diagnostic methods, biopsies are often required. To address this limitation, we proposed radio frequency time series dynamic processing (RFTSDP) in 2020, which analyzes the dynamic response of tissue and the impact of scatterer displacement on RF echoes during controlled stimulations to enhance diagnostic information. Methods We developed a vibration-generating device and collected ultrafast ultrasound data from 11 ex vivo breast tissue samples under different stimulations. Deep learning (DL) was used for automated feature extraction and lesion classification into 2, 3, and 5 categories. The performance of the convolutional neural network (CNN)-based RFTSDP method was compared with traditional machine learning techniques, which involved spectral and nonlinear feature extraction from RF time series, followed by a support vector machine (SVM). Results With 65 Hz vibration, the DL-based RFTSDP method achieved 99.53 +/- 0.47% accuracy in classifying and grading breast lesions. CNN consistently outperformed SVM, particularly under vibratory stimulation. In 5-class classification, CNN reached 98.01% versus 95.64% for SVM, with the difference being statistically significant (P < .05). Furthermore, the CNN-based RFTSDP method showed a 28.67% improvement in classification accuracy compared to the non-stimulation condition and the analysis of focused raw data. Conclusions We developed a DL-based CAD system capable of classifying and grading breast lesions. This study demonstrates that the proposed system not only enhances classification but also ensures increased stability and robustness compared to traditional methods.

摘要： Serving as the primary barrier, skin protects the human body against oxidative stress caused by ultraviolet radiation, pollutants, and free radicals. Antioxidants are essential for preventing oxidative damage. Polyphenols such as ferulic acid (FA) and kaempferol (KMP) demonstrate potent free radical scavenging properties. However, poor solubility, stability, and bioavailability often limit their clinical application. This study aimed to investigate the effect of independent variables (PVA and PLGA concentrations, and sonication time on %EE of payloads, and the particle size (PS) using Box-Behnken design (BBD). PS of the nanoparticles (NPs) ranged from 171.40 ± 2.09 to 212.60 ± 2.37 nm with polydispersity index (PDI) of 0.14–0.287. PS, PDI, and zeta (ζ) potential of the optimized NPs was 189.8 nm, PDI 0.217, and − 11.1 mV, respectively. FE-SEM analysis demonstrated that the NPs were well-dispersed and spherical. Pure FA and KMP showed rapid burst release with 83.15% and 60.52% within 2 h (complete by 4 h), whereas PLGA NPs exhibited sustained biphasic release (FA: 48.00% in 1 h, 81.71% in 6 h, 98.99% in 24 h; KMP: 34.00% in 1 h, 73.83% in 6 h, 98.98% in 24 h). Free FA and KMP showed moderate antioxidant activity with IC50 values of 22.92 ± 3.80 µg/mL and 26.53 ± 2.01 µg/mL, respectively, while the optimized NPs exhibited improved activity with a lower IC50 of 18.69 ± 2.08 µg/mL, compared to ascorbic acid (AA) (4.05 ± 2.98 µg/mL). In the HET-CAM assay, pure FA and KMP showed mild irritation (IS 2.56 and 2.89, respectively), while the optimized NPs were nearly non-irritant (IS 0.52), indicating that encapsulation reduced irritation and improved biocompatibility of payload. The results showed that the NPs prolonged drug release profiles, making them promising delivery carrier. This nanotechnology-driven approach enhances the protective and therapeutic potential of antioxidants in skin care and oxidati

关键词： Centroid radial distribution   Convolution neural network   Grad-CAM   Regional maritime collision risk assessment   Relative traffic representation   Risk influence distribution map  

摘要： Maritime transport handles over 80 % of global trade but rising traffic density and congestion have increased collision risk. Traditional assessment methods, including the ship domain and closest point-of-approach approaches, are limited to pairwise interactions and cannot capture regional traffic dynamics. To overcome these shortcomings, we propose a regional collision risk assessment framework integrating AIS-derived relative traffic representations, convolutional neural networks, and explainable artificial intelligence. A gradient-weighted class activation mapping framework generated risk influence distribution maps and a centroid radial distribution, while a novel indicator, the risk influence radius (RIR), quantified radial influence patterns. Experiments with 3,660 low-risk and 366 high-risk cases from South Korean coastal waters revealed that the Visual Geometry Group (VGG) network backbone achieved the highest performance, with an average F1-score of 0.9324, outperforming ResNeXt, EfficientNet, ResNet, and MobileNetV3 models. Statistical tests revealed a significant difference in RIR (p < 0.05) between high- and low-risk collision areas at 10 and 20 km, but not at 30 km. These results confirm RIR as an interpretable indicator of collision risk at regional scales, while highlighting its limitations at broader ranges. The framework combines accuracy with interpretability and supports practical applications such as maritime monitoring, route planning, and early warning systems.

关键词： Batch normalization   Bayesian optimized SVM   CNN   KNN   Layer fusion   RF   Spectral features  

摘要： Coronary Artery Disease (CAD) and Congestive Heart Failure (CHF) are significant contributors to global mortality. Early diagnosis and intervention in cardiac subjects are crucial for reducing mortality rates and minimizing the associated treatment costs linked to these conditions. This highlights the critical need for efficient healthcare strategies aimed at the early detection and classification of cardiac diseases to improve patient outcomes. This study proposed a hybrid model combination of layer fusion Convolutional Neural Network (LF-CNN) and Machine Learning (ML) techniques, including Bayesian Optimized Support Vector Machine (BO-SVM), k-nearest neighbor (KNN), and Random Forest (RF) for detecting Normal (healthy), CAD, and CHF arrhythmia conditions. For training and validation, time-series and spectral features are extracted from HRV database of cardiac diseases to feed input of proposed model. Additionaly, to speedup the training, validation and to overcome overfitting, batch normalization was employed to adjust the distribution of features. This study also employed t-distributed Stochastic Neighbor Embedding (t-SNE) within deep CNN to visualize data and ensuring that fusion of different features. The simulation results demonstrated that the proposed method achieved an accuracy (Ac) of 99.88 %, a precision (Pr) of 99.89 %, a recall (sensitivity) of 99.68 %, and an F1-score of 99.79 % by integrating BO-SVM, KNN, and RF ML techniques with LF-CNN. The findings indicate that classifier performance can be significantly enhanced by fusion of multiple features instead of relying on a single feature and emphasizing the benefits of integrating ML with deep learning (DL) techniques.

关键词： Computer-Aided Design   Multimedia Technology   Fashion Design   Sustainability   Cultural Heritage  

摘要： Computer-aided design (CAD) and multimedia technology, as important tools of digital design, are widely used in the field of fashion design. In this article, the author discusses the synergy between computer-aided design and multimedia technology in improving the efficiency of fashion design resources and cultural expression. Mathematical modeling, quantitative evaluation, and case analysis indicated that the integration of the two technologies not only optimizes the design process and improves the utilization rate of materials but also realizes the digital extraction, redesign, and modern expression of cultural elements. The author puts forward a set of comprehensive frameworks that integrate concepts of green design and technical reality and provides theoretical support and practical reference for the sustainable development and cultural innovation of fashion design.

摘要： The transportation system depends on existing road networks that connect destinations, forming the backbone of urban mobility. This study focuses on the creation and analysis of a road network dataset (ND) designed for spatial network analysis, with an emphasis on route optimization within urban areas and support for sustainable urban planning and smart infrastructure. The methodology includes data acquisition and digitization using geographic information system (GIS) technologies, specifically OpenStreetMap and Google MyMaps, followed by CAD-to-GIS format conversion to ensure precise spatial alignment. This process yields a comprehensive ND that enables shortest path analysis and optimized route planning. Validation showed distance discrepancies under ± 0.52% between ArcMap and QGIS tools, with GPS error within ± 2.09%, confirming high spatial accuracy. Well-connected nodes contribute to a robust geospatial network foundation. The final dataset supports transportation planning in academic environments and demonstrates potential for broader urban deployment. It enhances traffic flow, connectivity, and environmental outcomes through improved route strategies. The study offers a transferable framework applicable across cities, addressing infrastructure challenges and advancing smart, resilient, and sustainable urban growth. These findings contribute to spatial information science through practical methods that improve data reliability and decision-making.