Search Results (5,680)

Printed circuit board (PCB) defect inspection is critical for ensuring product reliability, yet it remains challenging due to the microscopic scale of defects and complex background patterns. To improve the localization of fine anomalies, this paper proposes EAS-DETR, an efficient and highly sensitive real-time end-to-end detector. First, we reconstruct the feature extraction backbone by introducing a novel C2f-EC module, which jointly models local textures and global structural dependencies. Second, an Adaptive Sparse Attention-based Intra-scale Feature Interaction (ASAFI) module is proposed to suppress background noise and focus the network’s attention on sparse defect regions. Finally, an optimized feature pyramid network, SGO-FPN, is designed to mitigate cross-scale feature misalignment and preserve high-resolution spatial details for small object localization. Experiments demonstrate that EAS-DETR achieves an mAP@0.5 of 93.0% and a 91.9% recall on a multi-source PCB dataset. The model outperforms mainstream YOLO variants and baseline RT-DETR models while maintaining a moderate parameter count of 14.6M and achieving a real-time inference speed of over 70 FPS. Furthermore, cross-domain validations on public benchmarks confirm its robust generalization capability for complex tiny object detection tasks. Full article

In recent years, as underwater vehicles continue to improve their noise reduction capabilities, sonar-based detection has faced significant challenges, and non-acoustic detection has become a research focus. Gravity gradient detection, owing to its excellent concealment and anti-interference capability, is regarded as an important non-acoustic means for underwater target detection. Based on the structural characteristics of an underwater vehicle, this paper establishes a homogeneous ellipsoidal hull (HEH) model composed of two similar rotating ellipsoids. This model assumes that the mass of an underwater vehicle is completely uniformly distributed over the outer hull. Analytical formulas for the gravity anomaly and gravity gradient anomaly generated by this model are derived, and their spatial distribution characteristics are analyzed. Furthermore, based on the HEH model, the feasibility underwater vehicle detection using the vertical gravity gradient component is analyzed. Results show that when the accuracy of the gravity gradiometer reaches 10⁻⁴ E, the detection distance for a large underwater vehicle with a displacement of 18,750 t can reach 570 m. Full article

While vision-language models (VLMs) have been widely applied in zero-shot anomaly detection (ZSAD), their performance remains limited by the inability to distinguish fine-grained normal and abnormal textures, coupled with inadequate capabilities in detecting complex morphological anomalies. To address these limitations, this paper proposes BAG-CLIP (Bifurcated Attention Graph-Enhanced CLIP), a dual-path graph-enhanced zero-shot anomaly detection method. This approach employs a Bifurcated Self-Attention (BSA) module to decouple visual features, processing global semantics and spatial details separately to mitigate the inherent conflict between abstract semantic representation and precise spatial localization. A Self-Attention Graph (SAG) module is designed to model the topological structure of complex morphological anomalies. This module dynamically constructs visual features’ topological relationships and utilizes graph convolutions to aggregate neighborhood information, thereby enhancing the model’s representational capacity for diverse and complex morphological anomalies. Extensive experiments are conducted on five diverse industrial datasets, featuring complex transmission line backgrounds alongside general industrial scenarios. The proposed method is comprehensively evaluated against 11 state-of-the-art (SOTA) methods. On the EPED (Electrical Power Equipment Dataset) and MPDD datasets, BAG-CLIP outperforms the second-best methods in image-level AUROC (Area Under the Receiver Operating Characteristic Curve) by 3.7% and 2.8%, respectively. BAG-CLIP achieves superior performance in both zero-shot anomaly detection and segmentation. Full article

Reliable change detection in indoor environments remains a challenge for autonomous robotic systems using 3D LiDAR. Existing methods often require manual annotation, computationally intensive architectures, or focus on outdoor scenes. This paper presents SPICD-Net, a lightweight Siamese PointNet framework for indoor 3D change detection trained exclusively on synthetically generated anomalies, eliminating manual labeling. The framework offers three deployment-oriented contributions: a three-class Siamese formulation separating no-change, changed, and geometrically inconsistent tile pairs; a pre-FPS anomaly injection strategy that aligns synthetic training with inference-time preprocessing; and a stochastic-gated Chamfer-statistics branch that complements learned embeddings with explicit geometric cues under consumer-grade hardware constraints. Evaluated on 14 controlled simulation experiments in an indoor corridor dataset, SPICD-Net achieved aggregated Precision = 0.86, Recall = 0.82, F1-score = 0.84, and Accuracy = 0.96, with zero false positives in the no-change baseline and mean inference time of 22.4 s for a 172-tile map on a single consumer GPU. Additional robustness experiments identified registration accuracy as the main operational prerequisite. A limited real-world validation in one unseen room (four scans, 67 tiles) achieved Precision = 0.583, Recall = 1.000, and F1 = 0.737. Full article

Fine-grained energy management in smart-campus buildings requires accurate load forecasting together with reliable and interpretable anomaly diagnosis. This study presents an integrated forecasting–diagnosis framework for building energy systems. Hourly energy demand is modeled using a Transformer-based sequence-to-sequence architecture, in which a domain-aware attention mechanism is introduced to separately represent historical consumption dynamics, environmental influences, and temporal regularities commonly observed in building energy use. Anomaly diagnosis is conducted through a dual-scale strategy that supports both the timely detection of abrupt abnormal events and the identification of gradual performance degradation. Short-term anomalies are detected from forecasting residuals using adaptive thresholds, while long-term anomalies are identified by comparing current residual patterns with same-season historical baselines and validating multi-window trends over a 48 h horizon. The two detection streams are jointly used to distinguish point, pattern, and composite anomalies. To support practical operation and maintenance, SHAP-based explanations are provided to interpret both energy predictions and detected anomalies. Case studies on two educational buildings from the Building Data Genome Project 2 demonstrate that the proposed framework achieves the best overall forecasting performance against both conventional baselines and stronger recent Transformer-based models, with mean absolute percentage errors of approximately 3%. The results indicate that the proposed framework provides a practical solution for data-driven energy monitoring and decision support in smart buildings. Full article

Multivariate time series anomaly detection holds critical application value in key domains such as industrial system monitoring, financial risk management, and medical surveillance. However, existing approaches face two major challenges: reconstruction-based or prediction-based models tend to adapt to anomalous patterns during training, thereby weakening the distinction between normal and abnormal samples; furthermore, the non-stationary nature of time series leads to distribution shifts between training and testing data, impairing model generalization. To address these issues, this paper proposes the TFCID model. The model innovatively leverages diffusion principles to effectively impute missing time series data while capturing significant frequency-domain features. In the temporal processing stream, an unconditional diffusion model combined with imputation masking is employed to achieve high-precision imputation of randomly missing values, effectively preventing anomalies from interfering with model training. In the frequency-domain processing stream, an amplitude-aware frequency-domain masked autoencoder is introduced to specifically capture periodic or trend-based pattern anomalies. The model mitigates distribution shift by constraining the discrepancy between temporal and frequency-domain representations via adversarial contrastive learning, and uses this discrepancy as a robust anomaly scoring metric. Experimental results on five public benchmark datasets show that TFCID significantly outperforms state-of-the-art methods in detection accuracy (F1-Score), validating its effectiveness in anomaly detection tasks. Full article

The escalating complexity of network infrastructures and the increasing sophistication of cyber threats require increasingly robust and automated Intrusion Detection Systems (IDS). This article presents a comparative investigation of the effectiveness of various Machine Learning and Deep Learning architectures in detecting network anomalies in network logs. The methodology encompassed classic supervised and ensemble algorithms, such as Random Forest and XGBoost, to sequential Deep Learning approaches (LSTM, GRU) and unsupervised models based on latent reconstruction (VAE, DeepLog). The results demonstrate that supervised approaches significantly outperformed unsupervised methods in the analyzed context. The optimized XGBoost model established a performance benchmark, achieving a Recall of 0.96 and a Precision of 0.85, thereby offering an optimal balance between detecting rare threats and minimizing false alarms. In contrast, unsupervised models revealed critical limitations, suggesting that statistical mimicry between normal and anomalous traffic hinders detection based solely on reconstruction error. Additionally, the study documents the technical interoperability challenges when attempting to integrate state-of-the-art language models, such as BERT. In conclusion, this work validates the effectiveness of Gradient Boosting algorithms and recurrent networks as viable and scalable solutions for critical network security, providing guidelines for model selection in real monitoring environments. Full article

Background: Neonatal Ebstein’s anomaly (EA) is a severe condition with significant hemodynamic instability and early myocardial dysfunction, where abnormal right-heart geometry limits conventional echocardiography and highlights the value of myocardial deformation imaging. Methods: We conducted a single-center retrospective observational study including 16 neonates with EA and 26 healthy neonates. All subjects underwent comprehensive transthoracic echocardiography during the neonatal period. Conventional two-dimensional imaging and speckle-tracking echocardiography (STE) were used to assess biventricular and biatrial myocardial deformation. Deformation parameters were compared between groups, and receiver operating characteristic (ROC) curve analysis evaluated diagnostic performance. Results: Neonates with EA demonstrated significant structural remodeling and severe biventricular and biatrial dysfunction compared with controls. Speckle-tracking showed markedly reduced right ventricular longitudinal strain (LS) in all segments (all, p < 0.001), particularly in free-wall and four-chamber views. Left ventricular (LV) global LS (GLS) was significantly reduced in neonates with EA compared with controls (−14.53% vs. −22.32%, p < 0.001), indicating early involvement of LV myocardial function in the neonatal period. Atrial reservoir, conduit, and contractile strain were severely impaired in both atria (all, p < 0.001). ROC analysis revealed excellent diagnostic accuracy, especially for LVGLS (AUC 0.919) and right atrial contractile strain (AUC 0.958). Conclusions: STE enables the early detection of extensive biventricular and biatrial myocardial dysfunction in neonatal EA, including abnormalities not fully captured by conventional echocardiographic parameters, thereby providing significant incremental diagnostic value. Full article

Generally, land seismic data acquisition in arid areas is a labor-intensive, costly, and challenging process, often hindered by challenging terrain and safety risks. To overcome these limitations, we propose the integration of autonomous Unmanned Aerial Vehicles (UAVs) into land seismic data acquisition, enabling efficient data collection in difficult, inaccessible terrain. This is a cooperative mission workflow combining a Scouting UAV for high-resolution aerial scouting, followed by the swarm deployment of an Autonomous Seismic Acquisition Device (ASAD) for seismic data recording. The cooperative system allows for precise landing and subsequent deployment of seismic sensors in optimal locations. Previously, we demonstrated the applicability of passive seismic recorded with ASAD drones to near-surface characterization. This study covers the results of a field trial, where both the ASAD and Scouting UAV systems successfully acquired high-resolution seismic data with an active source, comparable to that of a conventional seismic data acquisition system. The results show that the ASAD seismic data exhibit a slightly higher noise level due to coupling variances and the fact that geophones were hardwired into 9-sensor arrays. However, due to its single-point sensing nature, it yields a superior frequency bandwidth, making it suitable for imaging shallow anomalies. The system underwent P-wave refraction tomography modeling and accurately detected a shallow subsurface cavity, showcasing its potential for near-surface characterization and shallow geohazard identification. This heterogeneous robotic system can support seismic data acquisition by enhancing safety, improving efficiency, and streamlining equipment mobilization, while minimizing environmental footprint. Full article

Defining the near-surface signal reflecting the deep sub-surface leakage is a critical challenge in the risk assessment of geologic carbon storage (GCS) projects, often exacerbated by decoupled deep-to-shallow modeling. This study quantifies the mass distribution and phase evolution of leaked CO₂ through deep reservoir-caprocks, intermediate aquifer, and near-surface soil, thereby showing the sub-surface retention characteristics and the detectability of near-surface signals. A geological model from the deep reservoir to the soil layer was constructed to simulate CO₂ leakage through the caprock and migration into overlying strata in 1000 years. Using the simulator of GPSFLOW, this study evaluates the evolution of fluid phases and the mass distribution during the injection for 100 years and the post-injection periods. The results indicate that (1) at the moment the injection ceases, 87.43–99.06% of the CO₂ remaining within the system is retained within the reservoirs, with less than 8.42% reaching the intermediate aquifer. Remarkably, although the CO₂ ultimately reaching the near-surface soil is less than 0.00073% of the total mass retained within the system, this mass accumulation translates to a concentration anomaly with a signal-to-noise ratio of 368 relative to the background baseline. (2) Sensitivity analysis reveals that the injection rate affects the timing of fluid transport—a tenfold increase in injection rate (from 3.17 to 31.7 kg/s) accelerates the upward movement of CO₂, advancing its arrival at the near-surface by 15 years without changing the overall mass partitioning. The permeability anisotropy ratio affects CO₂ migration and phase distribution—decreasing the vertical to horizontal permeability ratio (1, 0.5, 0.25, 0.125) reduces connectivity, which delays the upward transfer and increases the amount of the aqueous CO₂. However, specifically in the soil layer, the aqueous CO₂ accumulation reveals a non-monotonic trend that peaks at an intermediate ratio of 0.25. (3) CO₂ shows a cascading distribution across formations where reservoirs provide the primary storage, and the intermediate aquifer reduces the mass available for near-surface accumulation. This attenuation effect significantly reduces the CO₂ mass that reaches the soil layer, thereby controlling the strength and duration of near-surface environmental signals. This work offers a theoretical reference for formulating near-surface monitoring strategies for CO₂ leakage in GCS. Full article

Edge intelligence is often limited by the computation–energy trade-off in resource-constrained devices. Boltzmann machines (BMs) provide strong unsupervised learning capability, yet their reliance on Gibbs sampling makes digital implementations costly in both computation and energy. In this paper, we present a voltage-controlled magnetic anisotropy magnetic tunnel junction (VCMA-MTJ)-based MRAM system that performs in-memory stochastic sampling for state generation and updates in restricted/deep Boltzmann machines (RBMs/DBMs). By exploiting the intrinsic stochastic switching of VCMA-MTJs, the proposed system achieves probabilistic sampling with an energy as low as ∼10 fJ per sample. Implemented on a microcontroller-based edge platform, it enables real-time multi-sensor anomaly detection with an F1-score of 0.9854 and stable operation. The proposed hardware–algorithm co-design achieves in situ stochastic computing and storage within a single MRAM cell, providing an ultra-low-power substrate for probabilistic inference at the edge. Full article

This study presents an exploratory methodological framework for examining structural changes in regulatory risk disclosure using sentence embeddings, multivariate anomaly detection, and explainable artificial intelligence. Prior research typically relies on dictionary-based word frequencies, tone indicators, or topic proportions to quantify risk disclosure. While these measures capture disclosure intensity, they do not directly assess whether the internal semantic organization of risk narratives has shifted relative to historical patterns. We propose a structural semantic deviation framework that represents each company–year disclosure using thematic shares and embedding-based dispersion statistics and evaluates deviations from a historical baseline through unsupervised anomaly detection. Using Item 1A Risk Factors from Wells Fargo and JPMorgan Chase surrounding the 2016 regulatory shock as a focused two-firm case study, we show that traditional lexical metrics do not clearly isolate structural breaks, whereas embedding-based semantic trajectories reveal substantial narrative reconfiguration. Isolation-based modeling provides stable and discriminative anomaly scores in this setting, and SHAP decomposition highlights semantic distance, litigation emphasis, and disclosure contraction as important drivers of deviation in 2025 out-of-sample disclosures. These findings should be interpreted as methodological evidence rather than broad population-level claims. The study demonstrates how structural semantic modeling can be operationalized in regulatory disclosure analysis and provides a transparent framework that can be extended to larger panels and cross-industry settings in future research. Full article

We formulate the Quantum Information Copy Time (QICT) framework for conserved charges under strictly local quantum dynamics and isolate its logically strongest consequence. The theorem-level core is a receiver-optimised variational speed-limit inequality: after projection away from the conserved zero mode, the copy time is bounded from below by the inverse square root of a Liouvillian-squared receiver susceptibility times a local encoding seminorm. This statement is written in a finite-volume operator framework and does not require a diffusive ansatz. We then examine what follows only after additional infrared assumptions. Under a single diffusive slow-mode hypothesis, the variational inequality reduces to the practical scaling relation used in the benchmark computations. That reduction is treated as conditional and is stress-tested numerically rather than promoted by rhetoric. Within the anomaly-free Abelian span relevant for one Standard-Model-like generation, hypercharge selection is elevated to theorem-level status; by contrast, minimal gauge-algebra uniqueness remains explicitly conditional on additional model-selection axioms. The remainder of the manuscript is organised as an explicitly documented closure programme built on top of this core. In that closure, a gauge-coded QCA construction, a microscopic benchmark for the transport normalisation, and an electroweak matching convention are combined to produce a resonance-centred Higgs-portal singlet-scalar mass band together with direct-detection, invisible-width, and relic-consistency checks. These latter results are presented as model-dependent consequences of an explicit closure ansatz rather than as deductions from locality alone. Full article

Fraud remains a critical and evolving challenge in telecommunications, costing the industry billions annually. In Mobile Virtual Network Operator (MVNO) environments, conventional supervised approaches are limited because fraud labels are scarce or delayed, and outgoing-call behavior is shaped by heterogeneous tariffs. Using a real-world MVNO dataset (9603 subscribers, 1.78 million outgoing CDRs), we derive payment-based segments and confirm statistically significant baseline differences via Kruskal–Wallis tests with Dunn post hoc pairwise comparisons and Benjamini–Hochberg correction. We propose a plan-aware calibration strategy setting interpretable thresholds using segment-wise empirical quantiles. Evaluation employs both operational metrics (activation rates and workload) and two label-free alert quality proxy metrics: multi-rule co-occurrence and activation stability (coefficient of variation). Compared to global calibration, segment-aware calibration reduces the dominant S4 rule activation (5.44% to 4.59% of user-hours) while increasing sensitivity to rare overnight patterns (F6: 0.0017% to 0.0137% of user-days). Experiments confirm improved alert quality, and the robustness of these findings is confirmed by sensitivity analysis across quantile levels and alternative segmentation schemes. Overall, segment-specific calibration yields a more balanced, interpretable, and operationally fair rule-based screening layer suitable for MVNO constraints. Full article

Traditional agricultural greenhouse environmental monitoring systems often lack effective anomaly detection mechanisms, which can lead to inaccurate environmental regulation and negatively affect plant growth. To address this issue, this paper proposes a greenhouse monitoring system integrating Zigbee and 4G communication technologies, combined with a CNN-LSTM-POT anomaly detection algorithm. The system employs a Convolutional Neural Network (CNN) to extract local spatial features from multi-source sensor data and a Long Short-Term Memory (LSTM) network to model long-term temporal dependencies. To accurately identify anomalies, the Peaks Over Threshold (POT) method from extreme value theory is applied to prediction residuals, enabling adaptive dynamic threshold determination. Experimental results show that the proposed algorithm substantially improves anomaly detection precision, prevents erroneous data from disrupting greenhouse control decisions and reduces the volume of data transmitted to the cloud platform, thereby lowering computational overhead. This work provides a reliable and efficient solution for data monitoring and precise environmental control in smart agricultural greenhouses. Full article