Search Results (48)

Renewable energy’s growing penetration erodes traditional power systems’ inherent dynamic symmetry—balanced inertia, damping, and frequency response. This paper proposes a self-adaptive virtual synchronous generator (VSG) control strategy for a photovoltaic hybrid energy storage system (PV-HESS) based on a radial basis function (RBF) neural network. The strategy establishes a dynamic adjustment framework for inertia and damping parameters via online learning, demonstrating enhanced system stability and robustness compared to conventional VSG methods. In the structural design, the DC-side energy storage system integrates a passive filter to decouple high- and low-frequency power components, with the supercapacitor attenuating high-frequency power fluctuations and the battery stabilizing low-frequency power variations. A small-signal model of the VSG active power loop is developed, through which the parameter ranges for rotational inertia (J) and damping coefficient (D) are determined by comprehensively considering the active loop cutoff frequency, grid connection standards, stability margin, and frequency regulation time. Building on this analysis, an adaptive parameter control strategy based on an RBF neural network is proposed. Case studies show that under various conditions, the proposed RBF strategy significantly outperforms conventional methods, enhancing key performance metrics in stability and dynamic response by 16.98% to 70.37%. Full article

Change detection plays a crucial role in geological disaster tasks such as landslide identification, post-earthquake building reconstruction assessment, and unstable rock mass monitoring. However, real-world scenarios often pose significant challenges, including complex surface backgrounds, illumination and seasonal variations between temporal phases, and diverse change patterns. To address these issues, this paper proposes a multi-stage model for geological disaster change detection, termed MSMCD, which integrates strategies of global dependency modeling, local difference enhancement, edge constraint, and frequency-domain fusion to achieve precise perception and delineation of change regions. Specifically, the model first employs a DualTimeMamba (DTM) module for two-dimensional selective scanning state-space modeling, explicitly capturing cross-temporal long-range dependencies to learn robust shared representations. Subsequently, a Multi-Scale Perception (MSP) module highlights fine-grained differences to enhance local discrimination. The Edge–Change Interaction (ECI) module then constructs bidirectional coupling between the change and edge branches with edge supervision, improving boundary accuracy and geometric consistency. Finally, the Frequency-domain Change Fusion (FCF) module performs weighted modulation on multi-layer, channel-joint spectra, balancing low-frequency structural consistency with high-frequency detail fidelity. Experiments conducted on the landslide change detection dataset (GVLM-CD), post-earthquake building change detection dataset (WHU-CD), and a self-constructed unstable rock mass change detection dataset (TGRM-CD) demonstrate that MSMCD achieves state-of-the-art performance across all benchmarks. These results confirm its strong cross-scenario generalization ability and effectiveness in multiple geological disaster tasks. Full article

Surface defect detection is vital in industrial quality control. While deep learning has largely automated inspection, accurately locating defects with large-scale variations or those difficult to distinguish from similar backgrounds remains challenging. Furthermore, achieving high-precision and real-time performance under limited computational resources in deployment environments complicates effective solutions. In this work, we propose SD-GASNet, a network based on a self-distillation model compression strategy. To identify subtle defects, we design an Alignment, Enhancement, and Synchronization Feature Pyramid Network (AES-FPN) fusion network incorporating the Frequency Domain Information Gathering-and-Allocation (FIGA) mechanism and the Channel Synchronization (CS) module for industrial images from different sensors. Specifically, FIGA refines features via the Multi-scale Feature Alignment (MFA) module, then the Frequency-Guided Perception Enhancement Module (FGPEM) extracts high- and low-frequency information to enhance spatial representation. The CS module compensates for information loss during feature fusion. Addressing computational constraints, we adopt self-distillation with an Enhanced KL divergence loss function to boost lightweight model performance. Extensive experiments on three public datasets (NEU-DET, PCB, and TILDA) demonstrate that SD-GASNet achieves state-of-the-art performance with excellent generalization, delivering superior accuracy and a competitive inference speed of 180 FPS, offering a robust and generalizable solution for sensor-based industrial imaging applications. Full article

Semiconductor wafer manufacturing is one of the most complex and data-intensive processes in the industry, encompassing the front-end (FEOL), middle-end (MOL), and back-end (BEOL) stages, involving thousands of interdependent processes. Each stage can introduce potential variability, thereby reducing yield, making metrology and inspection crucial for process control. However, due to capacity, cost, and destructive testing constraints, exhaustive metrology for every wafer or die is impractical. Therefore, this study aims to introduce sampling strategies that have evolved to balance the accuracy, risk, and efficiency of measurement allocation. This review presents a literature review of static, adaptive, and dynamic sampling and discusses recent intelligent sampling techniques. The results show that traditional static sampling provides fixed, rule-based inspection schemes that ensure comparability and compliance but lack responsiveness to process variations. Adaptive sampling introduces flexibility, allowing measurement density to be adjusted based on detected drift, anomalies, or statistical control limits. Building on this, dynamic sampling represents a paradigm shift towards predictive, real-time decision-making driven by machine learning, risk analysis, and digital twin integration. The dynamic framework continuously assesses process uncertainties and prioritizes metrology to maximize information gain, thereby significantly reducing metrology workload without impacting yield or quality. Static, adaptive, and dynamic sampling together constitute a continuous evolution from deterministic control to self-optimizing intelligence. As semiconductor nodes move towards sub-3 nm, this intelligent sampling technology is crucial for maintaining yield, cost competitiveness, and process flexibility in autonomous, data-centric wafer fabs. Full article

Self-regulated learning (SRL) is crucial for effective learning, supporting academic achievement and lifelong competencies. Fostering SRL in primary education is important, yet teachers’ understanding and use of strategies are underexplored. This study provides an innovative, multi-method investigation of whether and how primary school teachers’ knowledge and beliefs about SRL align with their classroom practices. Video-based classroom observations were combined with semi-structured interviews to capture both what teachers think and what they do. The study addressed three research questions: (1) how and to what extent teachers implement SRL; (2) their knowledge and beliefs regarding SRL and alignment of these with classroom practice; (3) factors perceived as facilitating or constraining SRL implementation. Eight teachers participated, providing 16 h of observations and 11 h of interview data. Observations were analyzed using the ATES instrument, and interviews were coded thematically. Findings revealed variation in SRL implementation and misalignments between knowledge, beliefs, and practice. Teachers held misconceptions and focused mainly on metacognitive and motivational strategies in classroom practice. Limited self-efficacy and school- and classroom-level factors further constrained SRL implementation. Results indicate a need for professional development addressing knowledge gaps, misconceptions, and teachers’ self-efficacy, while encouraging school-wide reflective practices to support SRL in primary classrooms. Full article

This study investigates gender and disciplinary differences in self-directed study strategies with digital tools among university students. Grounded in Activity Theory (AT), Gender Similarities Hypothesis, and Self-Determination Theory, the research explores how students from STEM and non-STEM fields interact with digital and paper-based materials during individual study sessions. A mixed-methods design was employed, combining eye-tracking data with qualitative interviews. Forty students (mean age: 21.5; equally distributed by gender and disciplinary field) participated in 15 min study sessions using the Pupil Invisible eye-tracker. Fixation durations and heat maps were analyzed through RStudio (Version 2024.04.2+764r), while semi-structured interviews explored students’ motivations, study habits, and perceptions of strategy effectiveness. A theory-driven codebook was developed to analyze interview data, incorporating cognitive, emotional, socio-cultural, and metacognitive dimensions. Results indicate that the disciplinary field plays a more decisive role than gender in shaping study strategies. Female STEM students alternated between digital and paper tools, while non-STEM females used digital tools more continuously. Among males, non-STEM students favored paper, whereas STEM students engaged more with digital materials. Interview data confirmed intra-gender variation and emphasized the influence of context, autonomy, and study planning. The integration of eye-tracking and qualitative inquiry effectively captured both behavioral patterns and students’ perspectives. Findings suggest the need for inclusive, flexible educational practices that respect diverse learning preferences and disciplinary cultures. Full article

Homework is a key learning activity that promotes students’ self-regulation, motivation, and academic achievement. Previous studies highlight the importance of parental and teacher autonomy support in fostering these outcomes, but the mechanisms underlying these relationships require further investigation. This study investigates the effects of parental and teacher autonomy support on students’ self-regulation skills, mathematics homework completion, and academic achievement. Additionally, it examines whether gender moderates these relationships. The research was conducted with 530 middle school students from five public schools in Istanbul, covering 5th, 6th, and 7th grades. Data were collected on teachers’ and parents’ autonomy support in homework, students’ self-regulation strategies, homework behaviors, and academic performance. Analyses were performed using SPSS 25 and AMOS 25 software, employing structural equation modeling (SEM) with mediation paths, multi-group path analysis, and correlation tests. The results indicate that both parental and teacher autonomy support positively influence students’ use of self-regulation strategies, which in turn enhances homework completion and academic success. Self-regulation was found to mediate these relationships, confirming its crucial role in academic outcomes. However, gender did not significantly moderate these associations. This study advances the understanding of how parental and teacher autonomy support influence self-regulation, homework behavior, and academic achievement, contributing to the existing literature. By examining the mediating role of self-regulation and the moderating effect of gender, it provides in-depth insights into variations in homework engagement and academic outcomes. Findings highlight the importance of autonomy-supportive practices by parents and teachers to foster students’ independent study skills. Future studies could extend these findings by examining subject-specific differences and longitudinal effect. Full article

Automated Optical Inspection (AOI) of Printed Circuit Boards (PCBs) suffers from scarce labeled data and frequent domain shifts caused by variations in camera optics, illumination, and product design. These limitations hinder the development of accurate and reliable deep-learning models in manufacturing settings. To address this challenge, this study systematically benchmarks three Parameter-Efficient Fine-Tuning (PEFT) strategies—Linear Probe, Low-Rank Adaptation (LoRA), and Visual Prompt Tuning (VPT)—applied to two representative foundation vision models: the Contrastive Language–Image Pretraining Vision Transformer (CLIP-ViT-B/16) and the Self-Distillation with No Labels Vision Transformer (DINOv2-S/14). The models are evaluated on six-class PCB defect classification tasks under few-shot (k = 5, 10, 20) and full-data regimes, analyzing both performance and reliability. Experiments show that VPT achieves 0.99 ± 0.01 accuracy and 0.998 ± 0.001 macro–Area Under the Precision–Recall Curve (macro-AUPRC), reducing classification error by approximately 65% compared with Linear and LoRA while tuning fewer than 1.5% of backbone parameters. Reliability, assessed by the stability of precision–recall behavior across different decision thresholds, improved as the number of labeled samples increased. Furthermore, class-wise and few-shot analyses revealed that VPT adapts more effectively to rare defect types such as Spur and Spurious Copper while maintaining near-ceiling performance on simpler categories (Short, Pinhole). These findings collectively demonstrate that prompt-based adaptation offers a quantitatively favorable trade-off between accuracy, efficiency, and reliability. Practically, this positions VPT as a scalable strategy for factory-level AOI, enabling the rapid deployment of robust defect inspection models even when labeled data is scarce. Full article

Person Re-Identification (Re-ID), a critical component of intelligent surveillance and security systems, seeks to match individuals across disjoint camera networks under complex real-world conditions. While deep learning has revolutionized Re-ID through enhanced feature representation and domain adaptation, a holistic synthesis of its advancements, unresolved challenges, and ethical implications remains imperative. This survey offers a structured and critical examination of Re-ID in the deep learning era, organized into three pillars: technological innovations, persistent barriers, and future frontiers. We systematically analyze breakthroughs in deep architectures (e.g., transformer-based models, hybrid global-local networks), optimization paradigms (contrastive, adversarial, and self-supervised learning), and robustness strategies for occlusion, pose variation, and cross-domain generalization. Critically, we identify underexplored limitations such as annotation bias, scalability-accuracy trade-offs, and privacy-utility conflicts in real-world deployment. Beyond technical analysis, we propose emerging directions, including causal reasoning for interpretable Re-ID, federated learning for decentralized data governance, open-world lifelong adaptation frameworks, and human-AI collaboration to reduce annotation costs. By integrating technical rigor with societal responsibility, this review aims to bridge the gap between algorithmic advancements and ethical deployment, fostering transparent, sustainable, and human-centric Re-ID systems. Full article

Accurate mapping of wetland vegetation is essential for ecological monitoring and conservation, yet it remains challenging due to the spatial heterogeneity of wetlands, the scarcity of ground-truth data, and the spread of invasive species. Invasive plants alter native vegetation patterns, making their early detection critical for preserving ecosystem integrity. This study proposes a novel framework that integrates self-supervised learning (SSL), supervised segmentation, and multi-sensor data fusion to enhance vegetation classification in the Bayinbuluke Alpine Wetland, China. High-resolution satellite imagery from PlanetScope-3 and Jilin-1 was fused, and SSL methods—including BYOL, DINO, and MoCo v3—were employed to learn transferable feature representations without extensive labeled data. The results show that SSL methods exhibit consistent variations in classification performance, while multi-sensor fusion significantly improves the detection of rare and fragmented vegetation patches and enables the early identification of invasive species. Overall, the proposed SSL–fusion strategy reduces reliance on labor-intensive field data collection and provides a scalable, high-precision solution for wetland monitoring and invasive species management. Full article

To tackle the challenges of complex construction environment interference (e.g., lighting variations, occlusion, and marker contamination) and the demand for high-precision alignment during the hoisting process of bridge-erecting machines, this paper presents a two-stage marker detection–localization network tailored to hoisting alignment. The proposed network adopts a “coarse detection–fine estimation” phased framework; the first stage employs a lightweight detection module, which integrates a dynamic hybrid backbone (DHB) and dynamic switching mechanism to efficiently filter background noise and generate coarse localization boxes of marker regions. Specifically, the DHB dynamically switches between convolutional and Transformer branches to handle features of varying complexity (using depthwise separable convolutions from MobileNetV3 for low-level geometric features and lightweight Transformer blocks for high-level semantic features). The second stage constructs a Transformer-based homography estimation module, which leverages multi-head self-attention to capture long-range dependencies between marker keypoints and the scene context. By integrating enhanced multi-scale feature interaction and position encoding (combining the absolute position and marker geometric priors), this module achieves the end-to-end learning of precise homography matrices between markers and hoisting equipment from the coarse localization boxes. To address data scarcity in construction scenes, a multi-dimensional data augmentation strategy is developed, including random homography transformation (simulating viewpoint changes), photometric augmentation (adjusting brightness, saturation, and contrast), and background blending with bounding box extraction. Experiments on a real bridge-erecting machine dataset demonstrate that the network achieves detection accuracy (mAP) of 97.8%, a homography estimation reprojection error of less than 1.2 mm, and a processing frame rate of 32 FPS. Compared with traditional single-stage CNN-based methods, it significantly improves the alignment precision and robustness in complex environments, offering reliable technical support for the precise control of automated hoisting in bridge-erecting machines. Full article

Precise segmentation of cell nuclei in pathological images is the foundation of cancer diagnosis and quantitative analysis, but blurred boundaries, scale variability, and staining differences have long constrained its reliability. To address this, this paper proposes AL-Net—an adaptive learning network that breaks through these bottlenecks through three innovative mechanisms: First, it integrates dilated convolutions with attention-guided skip connections to dynamically integrate multi-scale contextual information, adapting to variations in cell nucleus morphology and size. Second, it employs self-scheduling loss optimization: during the initial training phase, it focuses on region segmentation (Dice loss) and later switches to a boundary refinement stage, introducing gradient manifold constraints to sharpen edge localization. Finally, it designs an adaptive optimizer strategy, leveraging symbolic exploration (Lion) to accelerate convergence, and switches to gradient fine-tuning after reaching a dynamic threshold to stabilize parameters. On the 2018 Data Science Bowl dataset, AL-Net achieved state-of-the-art performance (Dice coefficient 92.96%, IoU 86.86%), reducing boundary error by 15% compared to U-Net/DeepLab; in cross-domain testing (ETIS/ColonDB polyp segmentation), it demonstrated over 80% improvement in generalization performance. AL-Net establishes a new adaptive learning paradigm for computational pathology, significantly enhancing diagnostic reliability. Full article

Domain shift poses a major challenge in many Machine Learning applications due to variations in data acquisition protocols, particularly in the medical field. Test-time augmentation (TTA) can solve the domain shift issue and improve robustness by aggregating predictions from multiple augmented versions of the same input. However, TTA may inadvertently generate unrealistic or Out-of-Distribution (OOD) samples that negatively affect prediction quality. In this work, we introduce a filtering procedure that removes from the TTA images all the OOD samples whose representations lie far from the training data distribution. Moreover, all the retained TTA images are weighted inversely to their distance from the training data. The final prediction is provided by a Self-Ensemble with Confidence, which is a lightweight ensemble strategy that fuses predictions from the original and retained TTA samples using a weighted soft voting scheme, without requiring multiple models or retraining. This method is model-agnostic and can be integrated with any deep learning architecture, making it broadly applicable across various domains. Experiments on cross-domain leukocyte classification benchmarks demonstrate that our method consistently improves over standard TTA and Baseline inference, particularly when strong domain shifts are present. Ablation studies and statistical tests confirm the effectiveness and significance of each component. Full article

Traditional landslide susceptibility assessment (LSA) methods typically adopt a global modeling strategy, which struggles to account for the pronounced spatial heterogeneity arising from variations in topography, geology, and vegetation conditions within a region. Furthermore, model predictive performance is often undermined by feature redundancy. To address these limitations, this study focuses on the landslide disaster early-warning demonstration area in Honghe Prefecture, Yunnan Province. It proposes an ensemble learning model termed heterogeneity feature optimized stacking (HF-stacking), which integrates spatial heterogeneity partitioning (SHP) with feature selection to improve the scientific rigor of LSA. This method initially establishes an LSA system comprising 15 static landslide conditioning factors (LCFs) and two dynamic factors representing the average annual deformation rates derived from interferometric synthetic aperture radar (InSAR) technology. Based on landslide inventory data, an SHP method combining t-distributed stochastic neighbor embedding (t-SNE) and iterative self-organizing (ISO) clustering was developed to divide the study area into subregions. Within each subregion, a tailored feature selection strategy was applied to determine the optimal feature subset. The final LSA was performed using the stacking ensemble learning approach. The results show that the HF-stacking model achieved the best overall performance, with an average AUC of 95.90% across subregions, 4.23% higher than the traditional stacking model. Other evaluation metrics also demonstrated comprehensive improvements. This study confirms that constructing an SHP framework and implementing feature selection strategies can effectively reduce the impact of spatial heterogeneity and feature redundancy, thereby significantly enhancing the predictive performance of LSA models. The proposed method contributes to improving the reliability of regional landslide risk assessments. Full article

Existing methods for diagnosing depression rely heavily on subjective evaluations, whereas electroencephalography (EEG) emerges as a promising approach for objective diagnosis due to its non-invasiveness, low cost, and high temporal resolution. However, current EEG analysis methods are constrained by volume conduction effect and class imbalance, both of which adversely affect classification performance. To address these issues, this paper proposes a multi-stage deep learning model for EEG-based depression classification, integrating a cortical feature extraction strategy (CFE), a feature attention module (FA), a graph convolutional network (GCN), and a focal adversarial domain adaptation module (FADA). Specifically, the CFE strategy reconstructs brain cortical signals using the standardized low-resolution brain electromagnetic tomography (sLORETA) algorithm and extracts both linear and nonlinear features that capture cortical activity variations. The FA module enhances feature representation through a multi-head self-attention mechanism, effectively capturing spatiotemporal relationships across distinct brain regions. Subsequently, the GCN further extracts spatiotemporal EEG features by modeling functional connectivity between brain regions. The FADA module employs Focal Loss and Gradient Reversal Layer (GRL) mechanisms to suppress domain-specific information, alleviate class imbalance, and enhance intra-class sample aggregation. Experimental validation on the publicly available PRED+CT dataset demonstrates that the proposed model achieves a classification accuracy of 85.33%, outperforming current state-of-the-art methods by 2.16%. These results suggest that the proposed model holds strong potential for improving the accuracy and reliability of EEG-based depression classification. Full article