Search Results (355)

Icing and galloping disasters threaten the safe operation of power transmission lines, and effective response depends on multi-agent collaboration. To address the insufficient attention paid in existing studies to grassroots execution constraints, this paper constructs a tripartite evolutionary game model involving local governments, power grid enterprise O&M management, and grassroots O&M teams. The model integrates collaborative investment, agency costs, benefit sharing, and multi-layer reward–punishment mechanisms into a unified framework. Replicator dynamics and numerical simulations are then used to analyze the evolution of collaborative strategies; the parameters are non-dimensional benchmark values rather than empirically calibrated estimates. The results show that the system exhibits multi-stability and path dependence, with fully non-collaborative and fully collaborative equilibria possibly remaining stable simultaneously. The combination of strong government regulation and incentives with high-level enterprise collaborative management is the key mechanism for overcoming the low-collaboration trap, and strict accountability has higher marginal policy efficiency than equivalent subsidies. Reducing grassroots execution costs and moderately increasing their share of disaster mitigation benefits can accelerate collaborative convergence and expand the attraction basin of the high-collaboration equilibrium. This study provides theoretical support and mechanism-design implications for enhancing the resilience of collaborative emergency response for power transmission lines under extreme weather conditions. Full article

Background/Objectives: The “Land of Fires” (LF) in the Campania Region has attracted considerable attention due to massive environmental contamination deriving from decades of illegal disposal, burial, and burning of urban, industrial, and toxic waste. Cadmium (Cd) has been repeatedly proven to affect male reproductive function by a plethora of endocrine and non-endocrine mechanisms. The scientific literature is almost devoid of large studies addressing semen quality in this area, particularly by directly correlating seminal parameters to objectively measured pollutant burden in biological samples. Therefore, the aim of the current study was to comprehensively evaluate semen quality of males of reproductive age living in the LF, by correlating seminal parameters to cumulative local male reproductive tract Cd burden objectively quantified in whole semen samples. Methods: The current single-center, observational, cross-sectional study evaluated semen quality in 493 males aged 14–50 (29.07 ± 7.17) years living in three LF municipalities. Moreover, the association of semen quality with whole semen Cd (sCd) levels measured by inductively coupled plasma mass spectrometry (ICP-MS) was addressed in a subgroup of participants; semen samples suitable for semen Cd measurements were available from 383/493 (77.7%) participants of the total cohort, and all analyses involving semen Cd were performed within the measured subset. Results: In the total cohort, seminal parameters were as follows: semen pH 8.32 ± 0.3, semen volume 3.13 ± 1.67 mL, sperm concentration 37.58 ± 30.18 × 10⁶/mL, total count 111.2 ± 104 × 10⁶/ejaculate, total motility 56.83 ± 16.09%, progressive motility 50.22 ± 16.63%, in situ motility 6.72 ± 3.43%, immotile spermatozoa 43.07 ± 15.88%, normal morphology 7.97 ± 4.02%, and viability 64.75 ± 15.34%. Prevalence of normozoospermia and pathological seminal parameters was as follows: normozoospermia 66.5% (328/493), pathological seminal parameters 33.5% (165/493), specifically, oligozoospermia 14% (69/493), cryptozoospermia 0.8% (4/493), azoospermia 2.2% (11/493), asthenozoospermia 3% (15/493), teratozoospermia 0.6% (3/493), oligo-astheno-teratozoospermia 6.1% (30/493), necrozoospermia 5.7% (28/493), and different combined seminal parameters alterations 7.1% (35/493). Whole semen Cd was below (undetectable) or above (detectable) the limit of detection (LoD) (0.2 μg/L) in 66.6% (255/383) and 33.4% (128/383) whole semen samples, respectively. In samples with detectable sCd, sCd level was below or above the median value (0.76 μg/L; min–max 0.1–5.95 μg/L) in 23.4% (30/128) and 76.6% (98/128) whole semen samples, respectively. Participants with detectable sCd levels had a significantly reduced sperm total count (93.28 ± 84.88 × 10⁶/ejaculate vs. 113.2 ± 101.5 × 10⁶/ejaculate; p = 0.037), and normal morphology (7.29 ± 3.71% vs. 8.23 ± 3.91%; p = 0.034), and a significantly lower prevalence of normozoospermia (60.2% vs. 72.2%; p = 0.02) and significantly higher prevalence of pathological seminal parameters (39.8% vs. 27.8%; p = 0.02), specifically, a significantly higher prevalence of oligozoospermia (21.1% vs. 12.6%; p = 0.036) than those with undetectable sCd levels. Whole semen Cd levels were significantly higher in participants with pathological seminal parameters (1.08 ± 0.84 μg/L vs. 0.93 ± 0.74 μg/L; p = 0.037) than those with normozoospermia. Participants with sCd levels above the median value (N = 98) had a significantly reduced sperm concentration (29.12 ± 24.84 × 10⁶/mL vs. 43.62 ± 29.55 × 10⁶/mL; p = 0.015) and displayed a trend towards reduced sperm normal morphology (6.92 ± 3.38% vs. 8.55 ± 4.49%; p = 0.057) than those with sCd levels below the median value (N = 30). Moreover, participants with sCd levels above the median value (N = 98) had a significantly reduced sperm concentration (29.12 ± 24.84 × 10⁶/mL vs. 35.3 ± 26.29 × 10⁶/mL; p = 0.03), total count (85.77 ± 80.52 × 10⁶/ejaculate vs. 113.2 ± 101.5 × 10⁶/ejaculate; p = 0.008) and normal morphology (6.92 ± 3.38% vs. 8.23 ± 3.91%; p = 0.006), and a significantly lower prevalence of normozoospermia (57.1% vs. 72.2%; p = 0.008) and significantly higher prevalence of pathological seminal (42.9% vs. 27.8%; p = 0.008), specifically, a significantly higher prevalence of oligozoospermia (23.5% vs. 12.6%; p = 0.014) than those with undetectable sCd levels. Conclusions: The results of the current study demonstrate an association between the environmental Cd exposure and the impairment of seminal parameters, with a significantly poorer semen quality in participants with detectable sCd, and, more markedly, in those with sCd level above the median value, compared to participants with undetectable sCd, although subgroups comparisons highlighted a homogeneous profile in major confounders including age, BMI, and smoking habits among subgroups of participants with different sCd burden. Full article

Millimeter-wave radar enables non-contact monitoring of cardiac activity and therefore has the potential to reconstruct electrocardiogram signals without surface electrodes. However, existing radar-based electrocardiogram reconstruction methods still suffer from incomplete extraction of heartbeat-related information and insufficient modeling of electrocardiogram-related features, which limits reconstruction accuracy. To address these issues, this study proposes a millimeter-wave radar-based electrocardiogram reconstruction method that integrates a respiratory-harmonic-suppressed multi-channel signal-processing frontend with the proposed CA-WTBNet deep reconstruction network. First, based on maximal overlap discrete wavelet transform-based multi-resolution analysis, respiratory harmonics mixed into heartbeat-related components are suppressed by combining respiratory harmonic detection with a heart-rate frequency protection strategy, while cardiac-related information is preserved as much as possible. A multi-channel input representation is then constructed. Meanwhile, the proposed deep reconstruction network is developed to jointly model complementary channel-wise features, local waveform morphology, and temporal dependencies by integrating channel-attention mechanisms, convolutional residual modules, window-based Transformer blocks, and bidirectional long short-term memory. Experiments conducted on the public dataset show that our method achieves an average Pearson correlation coefficient of 0.9641, a mean normalized root mean square error of 0.0458, an average R-peak F1 score of 0.9956, and an average R-peak timing error of 3.13 ms on the test set. In comparison with related studies on the same public Resting dataset, the proposed method achieves the best overall performance among the compared methods, with a 0.53% improvement in Pearson correlation coefficient and a 10.20% reduction in normalized root mean square error over the best-performing compared method. Full article

Traditional acupoint localization methods rely heavily on manual operation, resulting in high subjectivity and limited accuracy. To improve the precision and stability of acupoint detection, this study integrates machine vision technology with in situ projection to achieve automated recognition and real-time visualization of human acupoints. First, an automatic calibration method based on image processing is proposed for back acupoints. Spinal features are extracted from the blue channel, enhanced using adaptive histogram equalization, and processed through region of interest extraction, minimum-threshold binarization, and morphological operations. Key spinal curve points are then fitted using Bézier functions. Canny edge detection is used to extract the human silhouette, locate the acromion, and derive the pixel scale of the “cun” measurement, enabling coordinate computation for 141 back acupoints. In the deep learning component, an improved YOLOv8-Pose model is developed for acupoint localization. Unlike existing methods that use local attention or the original Object Keypoint Similarity (OKS) loss, we introduce two innovations: a non-local attention module for global dependency modeling, and a novel Efficient Object Keypoint Similarity (EOKS) loss function that incorporates geometric constraints—namely, width, height, and center distance—in addition to Euclidean distance. A non-local attention mechanism is incorporated into the backbone to enhance global feature extraction, and the EOKS loss function is designed to improve spatiogeometric regression accuracy. An inference mechanism is further introduced to derive the remaining acupoints from 49 detected keypoints; experiments demonstrate that the improved model achieves 95.0% detection accuracy, outperforming the baseline by 2.62%, with an inference time of 14.5 ms. Finally, an in situ projection platform is constructed, combining camera calibration, four-point proportional scaling, and an OpenCV 4.5.4-based interactive interface. The system supports real-time translation, rotation, and scaling, enabling accurate projection of detected acupoints onto the human body. Full article

Vision Transformer (ViT) has been widely adopted in the computer vision community. However, the standard ViT often contains many parameters, usually performs poorly when trained from scratch on medium-scale datasets, and does not explicitly preserve the local spatial and channel-wise structures within each token. This work proposes a novel model called the Token-Shared Convolutional Projection Vision Transformer (TSCP-ViT). The core idea of TSCP-ViT is to integrate convolutional layers into the multi-head attention mechanism and to apply the same convolutional operation independently to each token, where each token exhibits spatial 2D multi-channel characteristics. In addition, this work introduces a Transformer decoder immediately after each Transformer encoder, enabling the classification tokens to aggregate information from all tokens and be updated using statistical information. Moreover, a trainable Non-Reversing Gate GELU (NRG-GELU) activation is also proposed. Comparative experiments on CIFAR-100, Food-101, and ImageNet100 show that, under comparable parameter counts and without pretraining or knowledge distillation, TSCP-ViT substantially surpasses ViT, outperforms CvT, outperforms ResNet on Food-101, and approaches ResNet on CIFAR-100 and ImageNet100, although with considerably higher FLOPs. Full article

Wheat is a major staple crop, and storage mold growth poses a severe threat to grain safety and quality stability. Natural mold development in stored wheat exhibits subtle, localized, and highly heterogeneous characteristics. Existing unimodal methods and global fusion approaches generally suffer from insufficient local feature sensitivity, hindering fine-grained mold severity grading. To address this limitation, we propose a Mask-Guided Fine-Grained Fusion Network, a weakly supervised framework based on local RGB–HSI fusion. This framework employs a dynamic parallel A/B experimental design to construct time-matched proxy labels via weakly supervised learning. A standardized preprocessing pipeline including single-kernel extraction, foreground segmentation, and cross-modal registration is established to resolve RGB–HSI spatial misalignment, ensuring physical-level spatial consistency of multimodal features. The model incorporates a Foreground-Aware Spectral Recalibration (FASR) module to suppress background noise, a Mask-Guided Dilated Cross-modal Local Attention (MDCLA) mechanism to establish fine-grained local mappings between RGB visual phenotypes and hyperspectral responses, and a sample-level adaptive fusion strategy to dynamically weight features by modal reliability, enhancing representation of complex samples across all mold stages. Experiments show that the Mask-Guided Fine-Grained Fusion Network achieves 0.9689 classification accuracy, 0.9698 Macro-F1 score, and 0.0593 Mean Absolute Error (MAE), significantly outperforming state-of-the-art unimodal deep models and global attention fusion baselines. This work provides a proof-of-principle framework for fine-grained non-destructive mold risk assessment in stored wheat. Full article

With the widespread adoption of intelligent terminals, edge devices, and distributed information systems in the financial domain, financial security sensing data exhibit multisource heterogeneity, dynamic temporal patterns, and high privacy sensitivity. Traditional centralized anomaly detection methods are no longer able to simultaneously satisfy the requirements of cross-institutional or cross-node collaborative modeling, client data privacy protection, and robust monitoring of transaction and system anomalies. To address this challenge, a data-local federated deep anomaly detection framework has been proposed for distributed financial security sensing systems. Initially, a local deep financial security sensing representation module is constructed to perform temporal encoding and attention-based modeling on multisource financial signals, including terminal operation status, network transaction communication, backend server operation, identity authentication, and anomaly alerts, thereby extracting representations relevant to anomalous behaviors. Subsequently, a data-local federated optimization and personalized aggregation mechanism is developed to enable cross-node knowledge sharing without transmitting raw transaction or client data, while local personalized detection heads are employed to adapt to non-independent and identically distributed (non-IID) financial institution data. Furthermore, an adversarially robust security detection and trust-aware aggregation strategy is introduced to enhance model stability under input noise, feature masking, anomaly camouflage, and potential malicious client updates. Experimental results demonstrate that the proposed method achieves an Accuracy of $92.37\%$, a Precision of $89.41\%$, a Recall of $88.26\%$, an F1-score of $88.83\%$, an AUC of $93.06\%$, and a PR-AUC of $89.15\%$ in the primary financial anomaly detection task, significantly outperforming baseline methods such as Isolation Forest, Autoencoder, LSTM, Transformer, FedAvg, FedProx, SCAFFOLD, and MOON. In robustness experiments, the method attains F1-scores of $87.95\%$, $86.42\%$, $86.88\%$, $84.57\%$, $86.73\%$, and $83.91\%$ under Gaussian noise, feature masking, temporal shift, adversarial perturbation, and $20\%$ and $30\%$ malicious client scenarios, respectively. Ablation studies further confirm the effectiveness of local representation learning, personalized federated optimization, adversarial training, and trust-aware aggregation mechanisms. Overall, the proposed approach provides an efficient intelligent anomaly detection solution for financial AI security monitoring scenarios characterized by data localization requirements, node heterogeneity, and attack perturbations. Full article

Photodynamic therapy (PDT) represents a promising non-antibiotic strategy for addressing bacterial and fungal infections, particularly in the context of increasing antimicrobial resistance and biofilm-associated disease. PDT is based on the light-induced activation of photosensitizers, leading to the generation of reactive oxygen species (ROS), including singlet oxygen (¹O₂), which induce oxidative damage to multiple microbial targets. Unlike conventional antimicrobial drugs that often act through specific molecular pathways, antimicrobial PDT produces simultaneous damage to membranes, proteins, nucleic acids, and extracellular biofilm components, thereby reducing the probability of resistance development. This review critically analyzes the cellular, biochemical, and biophysical determinants that govern PDT selectivity toward bacterial and fungal cells in comparison with mammalian host tissues. Particular attention is given to photosensitizer localization, membrane interactions, photobleaching, oxygen dependence, light penetration, and the balance between Type I and Type II photochemical mechanisms. The review provides a comparative overview of major molecular photosensitizer classes, including phenothiazines, porphyrins, chlorins, phthalocyanines, xanthene dyes, natural polyphenols, endogenous compounds, and advanced targeted photosensitizers. In addition, this review distinguishes molecular photosensitizers from nanotechnology-based platforms and delivery systems. Nanoparticles, polymeric carriers, hydrogels, and light-activated coatings are discussed not only as photosensitizer delivery tools, but also as systems that modulate aggregation, improve localization, enhance biofilm penetration, and enable surface-confined ROS generation. ROS are capable of causing phototoxic effects wherever they are located. Unless selectively accumulated by target organisms, there can be systemic phototoxicity. Overall, PDT should be regarded as a modular antimicrobial platform in which photosensitizer chemistry, formulation, light delivery, oxygen availability, and infection biology must be co-optimized. Although further studies are required to address clinical translation, regulatory complexity, material safety, and standardized treatment protocols, PDT offers a scientifically robust and clinically relevant approach that may complement conventional antibacterial and antifungal therapies, especially in localized, biofilm-associated, and device-related infections. Full article

Effective cross-modal feature alignment and interaction are central challenges in multispectral object detection. Although global cross-attention provides strong long-range modeling ability, its quadratic complexity with respect to feature size limits deployment on resource-constrained platforms. We therefore propose Progressive Pixel-Neighborhood Deformable Cross-Attention for multispectral feature fusion, termed PNAFusion. The proposed framework is motivated by two observations: weak misalignment between visible and thermal images is usually concentrated around local neighborhoods, and semantic correspondence across modalities often follows non-linear spatial mappings that fixed receptive fields cannot model well. To address these issues, PNAFusion incorporates local spatial priors into its architectural design to concentrate feature interaction and alignment on the most relevant neighborhoods. Specifically, a Pixel-Neighborhood Cross-Attention (PNCA) module is introduced to avoid redundant global feature matching and suppress background noise. Meanwhile, an Adaptive Deformable Alignment (ADA) module captures non-linear spatial correspondences through learned pixel-wise offsets. These components are further integrated through an iterative feedback mechanism to progressively refine cross-modal feature alignment. Experiments on FLIR, M3FD, and DroneVehicle show that PNAFusion achieves 84.2, 90.5, and 85.5 mAP@0.5, respectively, under the YOLOv5 detector, and further reaches 86.8 mAP@0.5 on FLIR and 90.8 mAP@0.5 on M3FD when transferred to Co-DETR. Efficiency analysis indicates that PNAFusion reduces allocated GPU memory by 33.0% compared with ICAFusion and reduces theoretical FLOPs from 194.8 G to 156.4 G, although the deformable sampling and iterative refinement introduce additional latency. These results demonstrate that PNAFusion provides a practical accuracy–memory trade-off for weakly aligned multispectral object detection. Full article

Orthodontic tooth movement (OTM) is a biologically driven process resulting from the mechanically induced remodeling of the periodontal ligament (PDL) and alveolar bone. A marked inter-individual variability exists in the rate of tooth movement, susceptibility to adverse outcomes such as external apical root resorption (EARR), and overall treatment response. This narrative review synthesizes current evidence on molecular, genetic, and epigenetic biomarkers that underline these differences. We summarize established local biomarkers derived from gingival crevicular fluid and saliva, including inflammatory cytokines, matrix metalloproteinases, and bone remodeling mediators reflecting OTM compression- and tension-side biology. Beyond fluid biomarkers, growing attention is given to genetic and epigenetic determinants of OTM. Specific gene mutations are associated with impaired or absent tooth movement, while multiple single-nucleotide polymorphisms have been linked to increased risk of EARR. Recent studies further demonstrate that orthodontic forces induce epigenetic remodeling in PDL cells, including DNA methylation changes in the gene promoters, histone modifications, and force-responsive non-coding RNAs such as miR-21 and miR-146a, which collectively regulate osteoclastogenesis, inflammation, and tissue adaptation. These findings indicate that OTM is governed by an integrated network combining mechanical stimuli with genetic predisposition and dynamic epigenetic regulation. Understanding these mechanisms provides a foundation for the development of biomarker-guided, patient-specific therapeutic strategies. Full article

Background/Objectives: Early and accurate discrimination of neurological conditions, dementia, stroke and healthy aging, remains a critical clinical challenge. Electroencephalography (EEG) is a non-invasive measure of brain dynamics and entropy-based features obtained from multichannel EEG have shown strong discriminative ability. However, existing deep learning approaches do not sufficiently address the combined challenges of small clinical cohorts and high-dimensional entropy feature spaces. In this study, a novel architecture is proposed for multi-class neurological EEG classification under extreme small-sample conditions. Methods: A novel dual-branch Channel-wise Transformer and Attention-Branch Network (EEG-ChTABNet) are pr to classify 19-channel EEG entropy features into three classes (dementia, stroke, healthy control; N = 45; 15 per class). The architecture suggests four new designs. First, the Channel Importance Attention (CIA) block, which adaptively learns to re-weight the importance of electrodes via squeeze-excitation. Second, the dual-branch encoder, which combines the global multi-head self-attention with the local depthwise-separable convolution. Third, the gated sigmoid fusion mechanism. Fourth, the bottleneck residual classification head, to solve overfitting. Eight entropy feature sets: Amplitude-Aware Permutation Entropy (AAPE), Attention Entropy (AttEn), Dispersion Entropy (DisEn), Distribution Entropy (DistrEn), Fluctuation-based Dispersion Entropy (FDispEn), Fuzzy Entropy (FuzEn), Linear Gaussian Estimation of the Conditional Entropy (LinEn), and Symbolic Dynamics (SyDy) were evaluated individually with stratified 5-fold cross-validation on within-fold SMOTE augmentation. Results: EEG-ChTABNet consistently outperformed the baseline Transformer on all 8 feature sets. DisEn and SyDy features yielded peak classification accuracy of 73.3% (AUC: 0.823 and 0.857, respectively) compared to the corresponding baseline of 57.8% and 55.6%. SyDy achieved the best overall AUC of 0.857 and the dementia detection sensitivity was up to 86.7% over multiple feature sets. Conclusions: EEG-ChTABNet shows the effectiveness of channel-adaptive, dual-branch Transformer Designs for EEG-based neurological classification from Small-Sample Entropy Feature Data, and Identifying SyDy and DisEn as the Most Discriminative Feature Representations for Three-Class Neurological EEG Classification. Full article

With the increasing demand for low-carbon energy, automated defect detection using unmanned aerial vehicle (UAV)-based thermal inspection has become essential for maintaining the reliability of photovoltaic systems. However, existing methods still suffer from low-contrast thermal imagery, large-scale variations of defects, and subtle thermal anomalies. To address these challenges, this study proposes Grouped-Hypergraph-Modulation DEIM (GHM-DEIM), a robust end-to-end detection framework based on an improved DEIM architecture. Specifically, a grouped multi-scale aggregation attention network is introduced to enhance global thermal perception and recover discriminative features from blurred backgrounds. In addition, an enhanced encoder incorporating a hypergraph-based context encoding mechanism is designed to model high-order non-local relationships and improve feature representation across different defect scales. Furthermore, a modulation fusion module is employed to adaptively refine multi-scale feature responses and suppress environmental noise interference. Extensive experiments conducted on the ThermoSolar-PV and PV-HSD-2025 datasets demonstrate that the proposed method consistently outperforms state-of-the-art detectors, achieving mAP@50 values of 88.6% and 74.2%, respectively, with improvements of 4.7% and 2.9% over the baseline. These results demonstrate the effectiveness and robustness of GHM-DEIM for UAV-based PV thermal defect inspection. Full article

Flexible strain sensors have attracted considerable attention in gait recognition owing to their ability to adhere directly to the skin near joints and transduce local deformation. In existing work, however, sensor placement and orientation are largely determined by anatomical experience, while multi-channel classification still relies on back-end digital processors, whose power consumption and latency constrain system practicality in wearable scenarios. This paper presents an integrated design path that proceeds from skin-mechanics theory through sensor-layout optimization to analog-domain front-end inference. On the layout side, the lines-of-non-extension (LoNE) theory is employed to convert the selection of sensor attachment angles from empirical judgment into a calculable mechanics problem; guided by the spatial course of LoNE in the ankle and knee regions, the positions and angles of the nine sensors are determined individually—channels perpendicular to the LoNE capture maximum strain, channels offset by 45 degrees supplement non-sagittal-plane information, and a channel aligned along the LoNE provides a near-zero-strain reference. On the circuit side, the mathematical equivalence between the weighted summation of a linear classifier and Kirchhoff’s current law (KCL) nodal current superposition is exploited to map the classification operation onto current aggregation in an analog circuit, yielding an in-sensor computing (ISC) front end in which the nine-channel weighted summation is completed in a single analog step. The sensors are fabricated by screen-printing a liquid-metal–polymer composite conductive ink onto a TPU film substrate, with a gauge factor RSD of $6.8\%$ and a tensile linearity $R^2>0.99$. Using walking, running, and stair descent as verification targets, the analog classifier reaches 99% accuracy at the circuit-level functional-verification stage. On real multi-subject data, it achieves $87.0\%\pm 8.4\%$ accuracy under intra-subject cross-session validation, with an analog-domain inference response faster than $100\mathsf{\mu }\mathrm{s}$. This design path is not bound to a specific joint or sensor material; when the layout methodology is extended to additional joint regions and the circuit architecture incorporates multiple outputs to cover more classification categories, the same workflow remains applicable, offering a promising low-power, lightweight technical solution for wearable motion monitoring. Full article

Industrial sensor systems are increasingly vulnerable to both physical anomalies and cyberattacks, while their collected time series typically present complex periodic and non-stationary characteristics, along with dynamic spatial dependencies among sensors. To address these issues, this paper proposes a dual-branch graph learning framework with frequency gating for simultaneous industrial sensor anomaly and cyberattack detection. The model first divides the input time series into multiple patches and decomposes each patch into periodic and non-stationary components via frequency analysis. Two graph isomorphism network branches, namely periodic GIN (P-GIN) and non-stationary GIN (NS-GIN), are designed to model the spatial dependencies of the two components separately, where the graph structure is adaptively learned using a Gaussian kernel-based mechanism. Furthermore, a frequency gating module is introduced in the non-stationary branch to enhance the representation of abnormal and attack-related features. Hierarchical temporal encoding is performed via intra-patch attention and inter-patch attention to capture both local and long-range temporal dependencies. Extensive experimental results on real-world industrial sensor datasets demonstrate that the proposed method achieves superior performance compared with state-of-the-art methods in both anomaly detection and cyberattack detection tasks. Full article

Accurate short-term traffic flow prediction is crucial for managing macroscopic Intelligent Transportation Systems (ITS). To overcome limitations in capturing complex spatiotemporal dependencies and the severe challenges of hyperparameter tuning, this paper proposes IHO-CNN-BiLSTM-Attention, a novel hybrid deep learning framework. Specifically, a Convolutional Neural Network (CNN) extracts local spatial features, a Bidirectional Long Short-Term Memory (BiLSTM) network captures temporal dependencies, and an attention mechanism dynamically weights key timesteps. To maximize the architecture’s performance, an Improved Hippopotamus Optimization (IHO) algorithm is proposed for automatic hyperparameter optimization. The IHO algorithm effectively overcomes the premature convergence of traditional optimizers by integrating a Piecewise Linear Chaotic Map (PWLCM) for initialization, tangent-based non-linear adaptive weights, a Tangent Flight defense mechanism, and Lens Opposition-Based Learning (LOBL) for local optimum escape. Evaluated comprehensively across three distinct macroscopic traffic benchmark datasets (a multimodal intersection, METR-LA velocity, and PeMSD4 volume), the IHO algorithm first demonstrated statistically significant superiority on standard CEC benchmark functions. Subsequently, the proposed hybrid model achieved state-of-the-art traffic state classification performance, maintaining peak F1-Scores of 0.9798, 0.8436, and 0.9561 across the highly diverse datasets. It significantly outperformed both classical optimized baselines (e.g., PSO, GWO) and contemporary heavy deep learning architectures (e.g., ASTformer, DiffSTG) under severe class imbalance and varying topological conditions. This work offers a robust, scalable, and highly generalized spatiotemporal forecasting solution with strong theoretical guarantees for intelligent traffic control. Full article