Search Results (948)

In recent years, convolutional neural network-based object detectors have achieved extensive applications in remote sensing (RS) image interpretation. While multi-scale feature modeling optimization remains a persistent research focus, existing methods frequently overlook the symmetrical balance between feature granularity and morphological diversity, particularly when handling high-aspect-ratio RS targets with anisotropic geometries. This oversight leads to suboptimal feature representations characterized by spatial sparsity and directional bias. To address this challenge, we propose the Parallel Interleaved Convolutional Kernel Network (PICK-Net), a rotation-aware detection framework that embodies symmetry principles through dual-path feature modulation and geometrically balanced operator design. The core innovation lies in the synergistic integration of cascaded dynamic sparse sampling and symmetrically decoupled feature modulation, enabling adaptive morphological modeling of RS targets. Specifically, the Parallel Interleaved Convolution (PIC) module establishes symmetric computation patterns through mirrored kernel arrangements, effectively reducing computational redundancy while preserving directional completeness through rotational symmetry-enhanced receptive field optimization. Complementing this, the Global Complementary Attention Mechanism (GCAM) introduces bidirectional symmetry in feature recalibration, decoupling channel-wise and spatial-wise adaptations through orthogonal attention pathways that maintain equilibrium in gradient propagation. Extensive experiments on RSOD and NWPU-VHR-10 datasets demonstrate our superior performance, achieving 92.2% and 84.90% mAP, respectively, outperforming state-of-the-art methods including EfficientNet and YOLOv8. With only 12.5 M parameters, the framework achieves symmetrical optimization of accuracy-efficiency trade-offs. Ablation studies confirm that the symmetric interaction between PIC and GCAM enhances detection performance by 2.75%, particularly excelling in scenarios requiring geometric symmetry preservation, such as dense target clusters and extreme scale variations. Cross-domain validation on agricultural pest datasets further verifies its rotational symmetry generalization capability, demonstrating 84.90% accuracy in fine-grained orientation-sensitive detection tasks. Full article

Accurate and robust power quality disturbance (PQD) classification is critical for modern electrical grids, particularly in noisy environments. This study presents a comprehensive comparative evaluation of machine learning (ML) and deep learning (DL) models for automatic PQD identification. The models evaluated include Support Vector Machines (SVM), Decision Trees (DT), Random Forest (RF), k-Nearest Neighbors (kNN), Gradient Boosting (GB), and Dense Neural Networks (DNN). For experimentation, a hybrid dataset, comprising both synthetic and real signals, was used to assess model performance. The robustness of the models was evaluated by systematically introducing Gaussian noise across a wide range of Signal-to-Noise Ratios (SNRs). A central objective was to directly benchmark the practical implementation and performance of these models across two widely used platforms: MATLAB R2024a and Python 3.11. Results show that ML models achieve high accuracies, exceeding $95\%$ at an SNR of 10 dB. DL models exhibited remarkable stability, maintaining $97\%$ accuracy for SNRs above 10 dB. However, their performance degraded significantly at lower SNRs, revealing specific confusion patterns. The analysis underscores the importance of multi-domain feature extraction and adaptive preprocessing for achieving resilient PQD classification. This research provides valuable insights and a practical guide for implementing and optimizing robust PQD classification systems in real-world, noisy scenarios. Full article

The burgeoning Internet of Medical Things (IoMT) offers unprecedented opportunities for real-time patient monitoring and predictive diagnostics, yet the current systems struggle with scalability, data confidentiality against quantum threats, and real-time privacy-preserving intelligence. This paper introduces Med-Q Ledger, a novel, multi-layered framework designed to overcome these critical limitations in the Medical IoT domain. Med-Q Ledger integrates a permissioned Hyperledger Fabric for transactional integrity with a scalable Holochain Distributed Hash Table for high-volume telemetry, achieving horizontal scalability and sub-second commit times. To fortify long-term data security, the framework incorporates post-quantum cryptography (PQC), specifically CRYSTALS-Di lithium signatures and Kyber Key Encapsulation Mechanisms. Real-time, privacy-preserving intelligence is delivered through an edge-based federated learning (FL) model, utilizing lightweight autoencoders for anomaly detection on encrypted gradients. We validate Med-Q Ledger’s efficacy through a critical application: the prediction of intestinal complications like necrotizing enterocolitis (NEC) in preterm infants, a condition frequently necessitating emergency colostomy. By processing physiological data from maternal wearable sensors and infant intestinal images, our integrated Random Forest model demonstrates superior performance in predicting colostomy necessity. Experimental evaluations reveal a throughput of approximately 3400 transactions per second (TPS) with ~180 ms end-to-end latency, a >95% anomaly detection rate with <2% false positives, and an 11% computational overhead for PQC on resource-constrained devices. Furthermore, our results show a 0.90 F1-score for colostomy prediction, a 25% reduction in emergency surgeries, and 31% lower energy consumption compared to MQTT baselines. Med-Q Ledger sets a new benchmark for secure, high-performance, and privacy-preserving IoMT analytics, offering a robust blueprint for next-generation healthcare deployments. Full article

Considering the growing demand for modern facility agriculture, it is essential to develop non-destructive technologies for assessing lettuce nutritional status. To overcome the limitations of traditional methods, which are destructive and time-consuming, this study proposes a multimodal non-destructive nitrogen detection method for lettuce based on multi-source imaging. The approach integrates terahertz time-domain spectroscopy (THz-TDS) and near-infrared hyperspectral imaging (NIR-HSI) to achieve rapid and non-invasive nitrogen detection. Spectral imaging data of lettuce samples under different nitrogen gradients (20–150%) were simultaneously acquired using a THz-TDS system (0.2–1.2 THz) and a NIR-HSI system (1000–1600 nm), with image segmentation applied to remove background interference. During data processing, Savitzky–Golay smoothing, MSC (for THz data), and SNV (for NIR data) were employed for combined preprocessing, and sample partitioning was performed using the SPXY algorithm. Subsequently, SCARS/iPLS/IRIV algorithms were applied for THz feature selection, while RF/SPA/ICO methods were used for NIR feature screening, followed by nitrogen content prediction modeling with LS-SVM and KELM. Furthermore, small-sample learning was utilized to fuse crop feature information from the two modalities, providing a more comprehensive and effective detection strategy. The results demonstrated that the THz-based model with SCARS-selected power spectrum features and an RBF-kernel LS-SVM achieved the best predictive performance (R² = 0.96, RMSE = 0.20), while the NIR-based model with ICO features and an RBF-kernel LS-SVM achieved the highest accuracy (R² = 0.967, RMSE = 0.193). The fusion model, combining SCARS and ICO features, exhibited the best overall performance, with training accuracy of 96.25% and prediction accuracy of 95.94%. This dual-spectral technique leverages the complementary responses of nitrogen in molecular vibrations (THz) and organic chemical bonds (NIR), significantly enhancing model performance. To the best of our knowledge, this is the first study to realize the synergistic application of THz and NIR spectroscopy in nitrogen detection of facility-grown lettuce, providing a high-precision, non-destructive solution for rapid crop nutrition diagnosis. Full article

This article presents an optimization of the explicit Euler method for a heat conduction model. The starting point of the paper was the analysis of the limitations of the explicit Euler scheme and the classical CFL condition in the transient domain, which pointed to the oscillation occurring in the intermediate states. To eliminate this phenomenon, we introduced the No-Sway Threshold given for the Fourier number (K), stricter than the CFL, which guarantees the monotonic approximation of the temperature–time evolution. Thereafter, by means of the identical inequalities derived based on the Method of Equating Coefficients, we determined the optimal values of $\Delta t$ and $\Delta x$. Finally, for the construction of the variable grid spacing (M2), we applied the equation expressing the R of the identical inequality system and accordingly specified the thickness of the material elements ($\Delta \xi$). As a proof-of-concept, we demonstrate the procedure on an application case with major simplifications: during an emergency shutdown of the Flexblue® SMR, the temperature of the air inside the tank instantly becomes 200 °C, while the initial temperatures of the water and the steel are 24 °C. For a 50.003 mm × 50.003 mm surface patch of the tank, we keep the leftmost and rightmost material elements of the uniform-grid (M1) and variable-grid (M2) single-line models at constant temperature; we scale the results up to the total external surface (6714.39 m²). In the M2 case, a larger portion of the heat power taken up from the air is expended on heating the metal, while the rise in the heat power delivered to the seawater is more moderate. At the 3000th min, the steel-wall temperature in M1 falls between 26.229 °C and 25.835 °C, whereas in M2 the temperature gradient varies between 34.648 °C and 30.041 °C, which confirms the advantage of the combination of variable grid spacing and the No-Sway Threshold. Full article

The experimental study has been carried out using advanced computer vision methods in order to visualize the moment of excitation and further propagation of a non stationary isotropic domain in a hybrid aligned nematic (HAN) microsized volume under the effect of a laser beam focused on a bounding liquid crystal surface. It has been shown that, when the laser power exceeds a certain threshold value, in bulk of the HAN microvolume, an isotropic circular domain is formed. We also observed a structure of alternating concentric rings around the isotropic circular region, which increases with distance from the center of the isotropic domain. The formation of a sequence of rings in a polarizing microscopic image indicates the formation of a complex topology of the director field in the HAN cell under study. The following evolution of the texture can be represented by two modes. Firstly, the “fast” heating mode, which is responsible for the formation and explosive expansion of an isotropic zone in bulk of the HAN microvolume with characteristic time ${\tau }_1$ due to a laser spot heating on the upper indium tin oxide (ITO) layer. Secondly, the “slow” heating mode, when an isotropic zone and concentric rings slowly expand with characteristic time ${\tau }_2$ mainly due to the finite thermoconductivity of ITO layer. When the laser power significantly exceeds the threshold value, damped oscillations of the isotropic domain are observed. We also introduced the metrics that allows quantitatively estimate the behavior of texture observed. The results obtained form an experimental basis for further investigation of thermomechanical force appearing in the LC system with coupled gradients of temperature and director fields. Full article

In order to enhance the tracking accuracy and disturbance rejection capability in the speed loop of an optoelectronic tracking servo control system, a parameter self-adjusting disturbance rejection generalized predictive control method (STGPC) based on a continuous-time model is proposed in this paper. First, a dynamic model of the servo turntable system is established, and a linear extended state observer (LESO) is designed to perform real-time estimation of internal and external disturbances in the system. Second, a generalized predictive control law incorporating the predictive model, performance metrics, and rolling optimization is systematically derived, where the reference trajectory is generated by a tracking differentiator and the system state is provided in real time by the LESO. Furthermore, a gradient descent method is innovatively introduced to achieve adaptive adjustment in the predictive time domain, and the stability of the closed-loop system is rigorously proven based on Lyapunov theory. Finally, simulation experiments were conducted to verify the tracking performance, disturbance rejection performance, and time-domain parameter self-adjustment effects of the control method. Simulation results show that compared with PID control and traditional linear generalized predictive control (LGPC), the proposed STGPC method reduces speed tracking residuals by 73.79% and 51.04%, respectively, enhances disturbance suppression capability for speed vibration disturbances by 50.55% and 47.55%, respectively, and enhances compensation capability for friction torque disturbances by 68.03% and 59.33%, respectively. The system demonstrates outstanding velocity tracking accuracy and disturbance rejection while exhibiting good robustness against system parameter perturbations. Full article

The automotive and agricultural industries face increasingly stringent demands with technological advancements and rising living standards, resulting in substantially heightened engineering complexity. In this background, optimization methods become indispensable tools for solving diverse engineering challenges. This narrative review paper provides a comprehensive overview of the application and challenges of five optimization algorithms, including gradient-based optimization algorithms, heuristic algorithms, surrogate model-based optimization algorithms, Bayesian optimization algorithms, and hybrid cellular automata algorithms in two fields. To accomplish this objective, the research literature published from 2000 to the present is analyzed, focusing on automotive structural optimization, material optimization, crashworthiness, and lightweight design, as well as agricultural product inspection, mechanical parameter optimization, and ecological system optimization. A classification framework for optimization methods is established based on problem characteristics, elucidating the core strengths and limitations of each method. Cross-domain comparative studies are conducted to provide reference guidance for researchers in related fields. Full article

With the rapid expansion of digital knowledge platforms and intelligent information systems, organizations and communities are producing a vast number of unstructured knowledge data, including annotated corpora, technical diagrams, collaborative whiteboard content, and domain-specific multimedia archives. However, knowledge sharing across institutions is hindered by privacy risks, high communication overhead, and fragmented ownership of data. Federated learning promises to overcome these barriers by enabling collaborative model training without exchanging raw knowledge artifacts, but its success depends on motivating data holders to undertake the additional computational and communication costs. Most existing incentive schemes, which are based on non-cooperative game formulations, neglect unstructured interactions and communication efficiency, thereby limiting their applicability in knowledge-driven scenarios. To address these challenges, we introduce SC-NBTI, a smart contract and Nash bargaining-based incentive framework for federated learning in knowledge collaboration environments. We cast the reward allocation problem as a cooperative game, devise a heuristic algorithm to approximate the NP-hard Nash bargaining solution, and integrate a probabilistic gradient sparsification method to trim communication costs while safeguarding privacy. Experiments on the FMNIST image classification task show that SC-NBTI requires fewer training rounds while achieving 5.89% higher accuracy than the DRL-Incentive baseline. Full article

Traditional fault detection methods, often designed for centralized or cloud-based systems, are ill-suited for the edge. The deployment of predictive maintenance solutions on ultra-low-cost embedded platforms remains a significant challenge due to strict limitations in memory, processing capacity, and energy availability. To address these challenges, vibration and motor current signals were analyzed using an ultra-low-cost RP2040 microcontroller. For fault detection, this study uses statistical time-domain features and principal component analysis (PCA), followed by classification with eXtreme Gradient Boosting (XGBoost) models distilled for resource-constrained deployment. Experimental evaluation demonstrated that vibration-based features achieved a diagnostic accuracy of 94.1%, while current-based representations obtained 95.5% accuracy when using principal components, compared to 83.2% with handcrafted statistical features. Model distillation reduced memory footprint by up to 2.5× (from 0.42 MB to 0.15 MB) without compromising diagnostic fidelity, enabling deployment within the 264 KB RAM and 2 MB Flash constraints of the RP2040 microcontroller. This study proposes a modular framework that systematically evaluates statistical features, dimensionality reduction, sensor synchronization, and model distillation, thereby identifying the most cost-efficient combination of techniques that balances diagnostic accuracy with strict memory and processing constraints. The findings establish that accurate fault detection can be realized directly on severely resource-limited hardware, thereby extending the practical applicability of condition monitoring to cost-sensitive industrial environments. Full article

The widespread presence of security patterns in modern anti-forgery systems has given rise to an urgent need for reliable smartphone authentication. However, persistent recognition inaccuracies occur because of the inherent degradation of patterns during smartphone capture. These acquisition-related artifacts are manifested as both spectral distortions in high-frequency components and structural corruption in the spatial domain, which essentially limit current verification systems. This paper addresses these two challenges through four key innovative aspects: (1) It introduces a chromatic-adaptive coupled oscillation mechanism to reduce noise. (2) It develops a DFT-domain processing pipeline. This pipeline includes micro-feature degradation modeling to detect high-frequency pattern elements and directional energy concentration for characterizing motion blur. (3) It utilizes complementary spatial-domain constraints. These involve brightness variation for local consistency and edge gradients for local sharpness, which are jointly optimized by combining maximum a posteriori estimation and maximum likelihood estimation. (4) It proposes an adaptive graph-based partitioning strategy. This strategy enables spatially variant kernel estimation, while maintaining computational efficiency. Experimental results showed that our method achieved excellent performance in terms of deblurring effectiveness, runtime, and recognition accuracy. This achievement enables near real-time processing on smartphones, without sacrificing restoration quality, even under difficult blurring conditions. Full article

Gold flotation performance is influenced by multiple interacting variables, yet most predictive studies in this area emphasize accuracy while neglecting interpretability, limiting their practical value for process engineers. This study applies explainable machine learning techniques to identify and interpret key variables, controlling cumulative gold recovery and grade using a small, experimentally derived dataset (n = 11) from Ballarat gold ore flotation. A Gradient Boosting Regressor, combined with SHAP (Shapley Additive Explanations), permutation importance, and feature importance analyses, was employed to uncover both linear and non-linear relationships. Power, head grade, and processing time consistently emerged as dominant predictors, while interaction effects (e.g., head grade × collector, size × head grade) provided additional explanatory insights. The findings reveal actionable process implications, including trade-offs between energy input and flotation efficiency, and highlight operational conditions for improved recovery and grade. This study demonstrates that interpretable machine learning can bridge the gap between statistical modeling and process optimization, delivering transparent, domain-specific insights even in data-constrained environments. Full article

The resistivity method and time-domain induced polarization (TDIP) method are two branches of electrical geophysical prospecting. In recent years, researchers have implemented time-lapse resistivity inversion and time-lapse TDIP inversion based on time-lapse constraint theory. Although time-lapse inversion ensures temporal continuity between inversion results obtained at distinct epochs, it may not only cause the results to deviate from the true subsurface conditions, but also result in significant structural discrepancies resistivity and TDIP inversion results, thereby reducing inversion accuracy. To address these issues, the joint inversion of time-lapse resistivity and TDIP data was implemented based on cross-gradient constraint theory and time-lapse constraint theory. Using synthetic data from the theoretical model, we conducted separate inversion, time-lapse inversion, and time-lapse joint inversion. Comparative analysis of the results from these inversion schemes reveals that, compared with separate inversion and time-lapse inversion, time-lapse joint inversion not only maintains the temporal continuity of inverted models across consecutive monitoring epochs but also enforces structural similarity among distinct physical property models. This approach significantly increases the accuracy of the inversion results and exhibits superior noise robustness. These findings confirm the stability, reliability, and superiority of the algorithm developed in this study, providing a novel approach for addressing geological monitoring challenges. Full article

To improve the performance of image super-resolution reconstruction, this paper optimizes the classical SRGAN model architecture. The original SRResNet is replaced with the EDSR network as the generator, which effectively enhances the ability to restore image details. To address the issue of insufficient multi-scale feature extraction in SRGAN during image reconstruction, an LSK attention mechanism is introduced into the generator. By fusing features from different receptive fields through parallel multi-scale convolution kernels, the model improves its ability to capture key details. To mitigate the instability and overfitting problems in the discriminator training, the Mish activation function is used instead of LeakyReLU to improve gradient flow, and a Dropout layer is introduced to enhance the discriminator’s generalization ability, preventing overfitting to the generator. Additionally, a staged training strategy is employed during adversarial training. Experimental results show that the improved model effectively enhances image reconstruction quality while maintaining low complexity. The generated results exhibit clearer details and more natural visual effects. On the public datasets Set5, Set14, and BSD100, compared to the original SRGAN, the PSNR and SSIM metrics improved by 13.4% and 5.9%, 9.9% and 6.0%, and 6.8% and 5.8%, respectively, significantly enhancing the reconstruction of super-resolution images, achieving more refined and realistic image quality improvement. The model also demonstrates stronger generalization ability on complex cross-domain data, such as remote sensing images and medical images. The improved model achieves higher-quality image reconstruction and more natural visual effects while maintaining moderate computational overhead, validating the effectiveness of the proposed improvements. Full article

Ovarian cancer remains a significant global health concern, and its diagnosis heavily relies on whole-slide images (WSIs). Due to their gigapixel spatial resolution, WSIs must be split into patches and are usually modeled via multi-instance learning (MIL). Although previous studies have achieved remarkable performance comparable to that of humans, in clinical practice WSIs are distributed across multiple hospitals with strict privacy restrictions, necessitating secure, efficient, and effective federated MIL. Moreover, heterogeneous data distributions across hospitals lead to model heterogeneity, requiring a framework flexible to both data and model variations. This paper introduces HFed-MIL, a heterogeneous federated MIL framework that leverages gradient-based attention distillation to tackle these challenges. Specifically, we extend the intuition of Grad-CAM to the patch level and propose Patch-CAM, which computes gradient-based attention scores for each patch embedding, enabling structural knowledge distillation without explicit attention modules while minimizing privacy leakage. Beyond conventional logit distillation, we designed a dual-level objective that enforces both class-level and structural-level consistency, preventing the vanishing effect of naive averaging and enhancing the discriminative power and interpretability of the global model. Importantly, Patch-CAM scores provide a balanced solution between privacy, efficiency, and heterogeneity: they contain sufficient information for effective distillation (with minimal membership inference risk, MIA AUC ≈ 0.6) while significantly reducing communication cost (0.32 MB per round), making HFed-MIL practical for real-world federated pathology. Extensive experiments on multiple cancer subtypes and cross-domain datasets (Camelyon16, BreakHis) demonstrate that HFed-MIL achieves state-of-the-art performance with enhanced robustness under heterogeneity conditions. Moreover, the global attention visualizations yield sharper and clinically meaningful heatmaps, offering pathologists transparent insights into model decisions. By jointly balancing privacy, efficiency, and interpretability, HFed-MIL improves the practicality and trustworthiness of deep learning for ovarian cancer WSI analysis, thereby increasing its clinical significance. Full article