Search Results (13,914)

Urban roof segmentation plays a pivotal role in applications such as urban planning, infrastructure management, and renewable energy deployment. This study explores the evolution of deep learning techniques from traditional Convolutional Neural Networks (CNNs) to cutting-edge transformer-based models in the context of roof segmentation from satellite imagery. We highlight the limitations of conventional methods when applied to urban environments, including resolution constraints and the complexity of roof structures. To address these challenges, we evaluate two advanced deep learning models, Mask R-CNN and MaskFormer, which have shown significant promise in accurately segmenting roofs, even in dense urban settings with diverse roof geometries. These models, especially the one based on transformers, offer improved segmentation accuracy by capturing both global and local image features, enhancing their performance in tasks where fine detail and contextual awareness are critical. A case study on Ben Guerir City in Morocco, an urban area experiencing rapid development, serves as the foundation for testing these models. Using high-resolution satellite imagery, the segmentation results offer a deeper understanding of the accuracy and effectiveness of these models, particularly in optimizing urban planning and renewable energy assessments. Quantitative metrics such as Intersection over Union (IoU), precision, recall, and F1-score are used to benchmark model performance. Mask R-CNN achieved a mean IoU of 74.6%, precision of 81.3%, recall of 78.9%, and F1-score of 80.1%, while MaskFormer reached a mean IoU of 79.8%, precision of 85.6%, recall of 82.7%, and F1-score of 84.1% (pixel-level, micro-averaged at IoU = 0.50 on the held-out test set), highlighting the transformative potential of transformer-based architectures for scalable and precise urban imaging. The study also outlines future work in 3D modeling and height estimation, positioning these advancements as critical tools for sustainable urban development. Full article

Foreign object debris (FOD) in civil aviation environments poses severe risks to flight safety. Conventional detection primarily relies on manual visual inspection, which is inefficient, susceptible to fatigue-related errors, and carries a high risk of missed detections. Therefore, there is an urgent need to develop an efficient and convenient intelligent method for detecting aircraft FOD. This study proposes a detection model based on a Siamese network architecture integrated with a spatial transformation module. The proposed model identifies FOD by comparing the registered features of evidence-retention images with their corresponding normally distributed features. A dedicated aircraft FOD dataset was constructed for evaluation, and extensive experiments were conducted. The results indicate that the proposed model achieves an average improvement of 0.1365 in image-level AUC (Area Under the Curve) and 0.0834 in pixel-level AUC compared to the Patch Distribution Modeling (PaDiM) method. Additionally, the effects of the spatial transformation module and training dataset on detection performance were systematically investigated, confirming the robustness of the model and providing guidance for parameter selection in practical deployment. Overall, this research introduces a novel and effective approach for intelligent aircraft FOD detection, offering both methodological innovation and practical applicability. Full article

Background: Neuroimaging-based diagnostic approaches are of critical importance for the accurate diagnosis and treatment of major depressive disorder (MDD). However, multisite neuroimaging data often exhibit substantial heterogeneity in terms of scanner protocols and population characteristics. Moreover, concerns over data ownership, security, and privacy make raw MRI datasets from multiple sites inaccessible, posing significant challenges to the development of robust diagnostic models. Federated learning (FL) offers a privacy-preserving solution to facilitate collaborative model training across sites without sharing raw data. Methods: In this study, we propose the personalized Federated Gradient Matching and Contrastive Optimization (pF-GMCO) algorithm to address domain shift and support scalable MDD classification using multimodal MRI. Our method incorporates gradient matching based on cosine similarity to weight contributions from different sites adaptively, contrastive learning to promote client-specific model optimization, and multimodal compact bilinear (MCB) pooling to effectively integrate structural MRI (sMRI) and functional MRI (fMRI) features. Results and Conclusions: Evaluated on the Rest-Meta-MDD dataset with 2293 subjects from 23 sites, pF-GMCO achieved accuracy of 79.07%, demonstrating superior performance and interpretability. This work provides an effective and privacy-aware framework for multisite MDD diagnosis using federated learning. Full article

In response to the inability to quickly assess wind noise performance during the early stages of automotive styling design, this paper proposes a method for predicting interior wind noise by integrating automotive point cloud models with the Gray Wolf Optimization Residual Network model (GWO-ResNet). Based on wind tunnel test data under typical operating conditions, the point cloud model of the test vehicle is compressed using an auto-encoder and used as input features to construct a nonlinear mapping model between the whole vehicle point cloud and the wind noise level at the driver’s left ear. Through adaptive optimization of key hyperparameters of the ResNet model using the gray wolf optimization algorithm, the accuracy and generalization of the prediction model are improved. The prediction results on the test set indicate that the proposed GWO-ResNet model achieves prediction results that are consistent with the actual measured values for the test samples, thereby validating the effectiveness of the proposed method. A comparative analysis with traditional ResNet models, GWO-LSTM models, and LSTM models revealed that the GWO-ResNet model achieved Mean Absolute Percentage Error (MAPE) and mean squared error (MSE) of 9.72% and 20.96, and 9.88% and 19.69, respectively, on the sedan and SUV test sets, significantly outperforming the other comparison models. The prediction results on the independent validation set also demonstrate good generalization ability and stability (MAPE of 10.14% and 10.15%, MSE of 23.97 and 29.15), further proving the reliability of this model in practical applications. The research results provide an efficient and feasible technical approach for the rapid evaluation of wind noise performance in vehicles and provide a reference for wind noise control in the early design stage of vehicles. At the same time, due to the limitations of the current test data, it is impossible to predict the wind noise during the actual driving of the vehicle. Subsequently, the wind noise during actual driving can be predicted by the test data of multiple working conditions. Full article

Outdoor ice rinks have long been a staple for inexpensive exercise and entertainment in cold environments. However, the possible deterioration of or impact on outdoor ice rinks from a changing climate is poorly understood due to no or little monitoring of data of such facilities. To investigate long-term changes in ice rinks over recent decades, an energy-balance-based ice rink model (with three versions considering precipitation and melt) was applied to a simulated ice rink for two representative area—Dorset of south-central Ontario and the Experimental Lakes Area (ELA) of northwestern Ontario, Canada. The model was calibrated and tested using four-year ice rink data (since limited data are available) and applied to a 40-year period starting in 1978 to reproduce the dates of rink-on and rink-off, rink duration in a season, and ice water equivalent under daily climate inputs, and to illustrate any changing trend in these variables, i.e., the ice rink responses to changed climate. Results showed no clear trend in any ice rink features over four decades, attributed to winter temperature that did not increase substantially (a weak driver), no change in events of rain-on-ice and snowfall-on-rink, and reduced wind speed (possibly slowing ice melting). This is the first trial of a physically based rink model to evaluate outdoor ice rinks. More in situ monitoring and in-depth modelling are necessary, and this model can help guide the monitoring. Full article

Composite pressure vessels have attracted significant attention in recent years owing to their lightweight characteristics and superior mechanical performance. However, analyzing composite layers remains challenging due to complex filament-winding (FW) pattern structures and the associated high computational costs. This study introduces a homogenization method to achieve cross-scale modeling of carbon fiber-reinforced plastic (CFRP) layers, accounting for both lay-up sequence and in-plane FW diamond-shaped form. The stacking sequence in an FW Type IV composite pressure vessel is numerically investigated through ply modeling and cross-scale homogenization. The composite tank structure, featuring a polyamide PA66 liner, is designed for a working pressure of 70 MPa and comprises 12 helical winding layers and 17 hoop winding layers. An FW cross-undulation representative volume element (RVE) is developed based on actual in-plane mesostructures, suggesting an equivalent laminate RVE effective elastic modulus. Furthermore, six different lay-up sequences are numerically compared using ply models and fully and partially homogenized models. The structural displacements in both radial and axial directions are validated across all modeling approaches. The partial homogenization method successfully captures the detailed fiber-direction stress distribution in the innermost two hoop or helical layers. By applying the Hashin tensile failure criterion, the burst pressure of the composite tank is evaluated, revealing 7.56% deviation between the partial homogenization model and the ply model. Fatigue life analysis of the Type IV composite pressure vessel is conducted using ABAQUS® coupled with FE-SAFE, incorporating an S-N curve for polyamide PA66. The results indicate that the fatigue cycles of the liner exhibit only 0.28% variation across different stacking sequences, demonstrating that homogenization has a negligible impact on liner lifecycle predictions. The proposed cross-scale modeling framework offers an effective approach for multiscale simulation of FW composite pressure vessels, balancing computational efficiency with accuracy. Full article

Currently, the customs inspection of wood-boring pests in timber still primarily relies on manual visual inspection, which involves observing insect holes on the timber surface and splitting the timber for confirmation. However, this method has significant drawbacks such as long detection time, high labor cost, and accuracy relying on human experience, making it difficult to meet the practical needs of efficient and intelligent customs quarantine. To address this issue, this paper develops a rapid identification system based on the peristaltic signals of wood-boring pests through the PyQt framework. The system employs a deep learning model with multi-attention mechanisms, namely the Residual Denoising Vision Network (RDVNet). Firstly, a LabVIEW-based hardware–software system is used to collect pest peristaltic signals in an environment free of vibration interference. Subsequently, the original signals are clipped, converted to audio format, and mixed with external noise. Then signal features are extracted through three cepstral feature extraction methods Mel-Frequency Cepstral Coefficients (MFCC), Power-Normalized Cepstral Coefficients (PNCC), and RelAtive SpecTrAl-Perceptual Linear Prediction (RASTA-PLP) and input into the model. In the experimental stage, this paper compares the denoising module of RDVNet (de-RDVNet) with four classic denoising models under five noise intensity conditions. Finally, it evaluates the performance of RDVNet and four other noise reduction classification models in classification tasks. The results show that PNCC has the most comprehensive feature extraction capability. When PNCC is used as the model input, de-RDVNet achieves an average peak signal-to-noise ratio (PSNR) of 29.8 and a Structural Similarity Index Measure (SSIM) of 0.820 in denoising experiments, both being the best among the comparative models. In classification experiments, RDVNet has an average F1 score of 0.878 and an accuracy of 92.8%, demonstrating the most excellent performance. Overall, the application of this system in customs timber quarantine can effectively improve detection efficiency and reduce labor costs and has significant practical value and promotion prospects. Full article

In our study, we examine how the temporal window shift—the step between consecutive analysis windows—affects EEG-based cognitive fatigue detection while keeping the window length fixed. Using a reference workload dataset and a pipeline that includes preprocessing and feature extraction, we vary the shift to control segment overlap and, consequently, the number and independence of training samples. We evaluate six machine-learning models (decision tree, random forest, SVM, kNN, MLP, and a transformer). Across the models, smaller shifts generally increase accuracy and F1 score, consistent with the larger sample count; however, they also reduce sample independence and can inflate performance if evaluation splits are not sufficiently stringent. Class-wise analyses reveal persistent confusion for the moderate-fatigue class, the severity of which depends on the chosen shift. We discuss the methodological trade-offs, provide practical recommendations for choosing and reporting shift parameters, and argue that temporal segmentation decisions should be treated as first-class design choices in EEG classification. Our findings highlight the need for transparent reporting of window length, shift/overlap, and subject-wise evaluation protocols to ensure reliable and reproducible results in cognitive fatigue detection. Our conclusions pertain to subject-wise generalization on the STEW dataset; cross-dataset validation is an important next step. Full article

Short-term building cooling load prediction is crucial for optimizing building energy management and promoting sustainability. While data-driven models excel in this task, their performance heavily depends on the input feature set. Feature selection must balance predictive accuracy (relevance) and model simplicity (minimal redundancy), a challenge that existing methods often address incompletely. This study proposes a novel feature optimization framework that integrates the Maximum Information Coefficient (MIC) to measure non-linear relevance and the Maximum Relevance Minimum Redundancy (MRMR) principle to control redundancy. The proposed MRMR-MIC method was evaluated against four benchmark feature selection methods using three predictive models in a simulated office building case study. The results demonstrate that MRMR-MIC significantly outperforms other methods: it reduces the feature dimensionality from over 170 to merely 40 variables while maintaining a prediction error below 5%. This represents a substantial reduction in model complexity without sacrificing accuracy. Furthermore, the selected features cover a more comprehensive and physically meaningful set of attributes compared to other redundancy-control methods. The study concludes that the MRMR-MIC framework provides a robust, systematic methodology for identifying essential feature variables, which can not only enhance the performance of prediction models, but also offer practical guidance for designing cost-effective data acquisition systems in real-building applications. Full article

Early and differential diagnosis of dementia is essential for timely and targeted care. This study investigated the feasibility of using an artificial intelligence (AI)-based system to discriminate between different stages and etiologies of dementia by analyzing facial emotions. We collected video recordings of 64 participants exposed to standardized audio-visual stimuli. Facial emotion features in terms of valence and arousal were extracted and used to train machine learning models on multiple classification tasks, including distinguishing individuals with mild cognitive impairment (MCI) and overt dementia from healthy controls (HCs) and differentiating Alzheimer’s disease (AD) from other types of cognitive impairment. Nested cross-validation was adopted to evaluate the performance of different tested models (K-Nearest Neighbors, Logistic Regression, and Support Vector Machine models) and optimize their hyperparameters. The system achieved a cross-validation accuracy of 76.0% for MCI vs. HCs, 73.6% for dementia vs. HCs, and 64.1% in the three-class classification (MCI vs. dementia vs. HCs). Among cognitively impaired individuals, a 75.4% accuracy was reached in distinguishing AD from other etiologies. These results demonstrated the potential of AI-driven facial emotion analysis as a non-invasive tool for early detection of cognitive impairment and for supporting differential diagnosis of AD in clinical settings. Full article

Lung cancer remains one of the most lethal cancers globally. Its early detection is vital to improving survival rates. In this work, we propose a hybrid computer-aided diagnosis (CAD) pipeline for lung cancer classification using Computed Tomography (CT) scan images. The proposed CAD pipeline integrates ten image preprocessing techniques and ten pretrained deep learning models for feature extraction including convolutional neural networks and transformer-based architectures, and four classical machine learning classifiers. Unlike traditional end-to-end deep learning systems, our approach decouples feature extraction from classification, enhancing interpretability and reducing the risk of overfitting. A total of 400 model configurations were evaluated to identify the optimal combination. The proposed approach was evaluated on the publicly available Lung Image Database Consortium and Image Database Resource Initiative dataset, which comprises 1018 thoracic CT scans annotated by four thoracic radiologists. For the classification task, the dataset included a total of 6568 images labeled as malignant and 4849 images labeled as benign. Experimental results show that the best performing pipeline, combining Contrast Limited Adaptive Histogram Equalization, Swin Transformer feature extraction, and eXtreme Gradient Boosting, achieved an accuracy of 95.8%. Full article

This study reconstructs seven decades (1949–2019) of morphodynamic changes and sediment dynamics in the Diable River (Québec, Canada) using nine series of aerial photographs, a high-resolution LiDAR Digital Elevation Model (2021), and grain-size analysis. The objectives were to document long-term river evolution, quantify erosion and deposition, and evaluate sediment connectivity between eroding sandy bluffs and depositional zones. Planform analysis and sediment budgets derived from DEMs of Difference (DoD) reveal an oscillatory trajectory characterized by alternating phases of sediment export and temporary stabilization, rather than a simple trend of degradation or aggradation. The most dynamic interval (1980–2001) was marked by widespread meander migration and the largest net export (−142.5 m³/km/year), whereas the 2001–2007 interval showed net storage (+70.8 m³/km/year) and short-term geomorphic recovery. More recent floods (2017, 2019; 20–50-year return periods) induced localized but persistent sediment loss, underlining the structuring role of extreme events. Grain-size results indicate partial connectivity: coarse fractions tend to remain in local depositional features, while finer sediments are preferentially exported downstream. These findings emphasize the geomorphic value of temporary sediment sinks (bars, beaches) and highlight the need for adaptive river management strategies that integrate sediment budgets and local knowledge into floodplain governance. Full article

In the face of ever-evolving cyber threats, modern intrusion detection systems (IDS) must achieve long-term adaptability without sacrificing performance on previously encountered attacks. Traditional IDS approaches often rely on static training assumptions, making them prone to forgetting old patterns, underperforming in label-scarce conditions, and struggling with imbalanced class distributions as new attacks emerge. To overcome these limitations, we present a continual learning framework tailored for adaptive intrusion detection. Unlike prior methods, our approach is designed to operate under real-world network conditions characterized by high-dimensional, sparse traffic data and task-agnostic learning sequences. The framework combines three core components: a clustering-based memory strategy that selectively retains informative historical samples using DP-Means; multi-level knowledge distillation that aligns current and previous model states at output and intermediate feature levels; and a meta-learning-driven class reweighting mechanism that dynamically adjusts to shifting attack distributions. Empirical evaluations on benchmark intrusion detection datasets demonstrate the framework’s ability to maintain high detection accuracy while effectively mitigating forgetting. Notably, it delivers reliable performance in continually changing environments where the availability of labeled data is limited, making it well-suited for real-world cybersecurity systems. Full article

The escalating prevalence of malware poses a significant threat to digital infrastructure, demanding robust yet efficient detection methods. In this study, we evaluate multiple Convolutional Neural Network (CNN) architectures, including basic CNN, LeNet, AlexNet, GoogLeNet, and DenseNet, on a dataset of 11,000 malware images spanning 452 families. Our experiments demonstrate that CNN models can achieve reliable classification performance across both multiclass and binary tasks. However, we also uncover a critical weakness in that even minimal image perturbations, such as pixel modification lower than 1% of the total image pixels, drastically degrade accuracy and reveal CNNs’ fragility in adversarial settings. A key contribution of this work is spatial analysis of malware images, revealing that discriminative features concentrate disproportionately in the bottom-left quadrant. This spatial bias likely reflects semantic structure, as malware payload information often resides near the end of binary files when rasterized. Notably, models trained in this region outperform those trained in other sections, underscoring the importance of spatial awareness in malware classification. Taken together, our results reveal that CNN-based malware classifiers are simultaneously effective and vulnerable to learning strong representations but sensitive to both subtle perturbations and positional bias. These findings highlight the need for future detection systems that integrate robustness to noise with resilience against spatial distortions to ensure reliability in real-world adversarial environments. Full article

Accurate prediction of reinforced concrete shear strength is essential for structural safety, yet datasets often contain a mix of raw geometric and material properties alongside physics-informed engineered features, making optimal feature selection challenging. This study introduces a statistically principled framework that advances feature reduction for neural networks in three novel ways. First, it extends the artificial neural network-based signal-to-noise ratio (ANN-SNR) method, previously limited to classification, into regression tasks for the first time. Second, it couples ANN-SNR with a confidence-interval (CI)-based stopping rule, using the lower bound of the baseline ANN’s R² confidence interval as a rigorous statistical threshold for determining when feature elimination should cease. Third, it systematically evaluates both raw experimental variables and physics-informed engineered features, showing how their combination enhances both robustness and interpretability. Applied to experimental concrete shear strength data, the framework revealed that many low-SNR features in conventional formulations contribute little to predictive performance and can be safely removed. In contrast, hybrid models that combined key raw and engineered features consistently yielded the strongest performance. Overall, the proposed method reduced the input feature set by approximately 45% while maintaining results statistically indistinguishable from baseline and fully optimized models (R² ≈ 0.85). These findings demonstrate that ANN-SNR with CI-based stopping provides a defensible and interpretable pathway for reducing model complexity in reinforced concrete shear strength prediction, offering practical benefits for design efficiency without compromising reliability. Full article