Search Results (1,052)

Terrestrial laser scanning (TLS) represents an efficient method for acquiring spatial data in urban environments, while the quality of resulting geometric outputs is significantly influenced by subsequent point cloud processing. This article focuses on analyzing the accuracy of automatic building footprint extraction from classified TLS point clouds, with an emphasis on the role of data cleaning methods. The study area is located in the city center of Žiar nad Hronom, where urban structures were monitored using TLS. For detailed analysis, three objects were selected—an apartment building, a garage, and an industrial building—representing different levels of geometric complexity. To simulate realistic processing conditions, classification results obtained from different software (Leica Cyclone 3DR, Trimble RealWorks, and LiDAR360) were used. Their quality was evaluated using standard metrics such as Precision, Recall, and F1-score. These classifications also served as input scenarios containing typical errors, such as point clusters, vegetation near buildings, or misclassified terrain elements. Subsequently, selected point cloud cleaning methods were applied to these datasets, specifically statistical outlier removal, noise filter, and label connected components. The accuracy of the extracted building footprints was evaluated by comparison with reference data obtained from geodetic measurements. The results show that automatic classification alone is not sufficient to achieve accurate building footprints, and that data cleaning plays a decisive role. For example, in the case of the apartment building, statistical filtering reduced the area from 1052 m² to approximately 854 m² (reference value: 706 m²) and significantly improved positional accuracy (centroid shift reduced from 0.455 m to 0.077 m). Similarly, for the industrial building, the area was reduced from 215 m² to approximately 165 m² (reference: 148 m²) while maintaining the correct number of corner points. In contrast, noise filter method proved to be less reliable, as removing up to 25–30% of points often did not lead to improvements in footprint geometry. The results highlight the importance of systematic point cloud cleaning as a key step in automated building footprint extraction and demonstrate that a properly selected combination of methods can significantly improve accuracy even in noisy datasets. The article also provides practical guidance for efficient TLS data processing in geoinformatics applications. Full article

Motor imagery (MI)-based brain–computer interfaces (BCIs) have led to great interest as a result of their potential use in neurorehabilitation, assistive robotics, and human–computer interaction. However, decoding electroencephalographic (EEG) signals with high accuracy continues to be a difficult task due to the weak signal-to-noise ratio, differences among subjects, and the complicated temporal–spectral–spatial neural dynamics. Deep learning methods recently developed, such as convolutional neural networks, recurrent architectures, graph neural networks, and adversarial transfer learning, have enhanced MI decoding performance, yet many models are still concentrating on a single representation domain or they need costly adaptation phases in terms of computation. To tackle these shortcomings, we present NeuroCrossNet, a unified tri-modal deep learning model that is able to learn the temporal, spectral, and spatial EEG features jointly for robust and calibration-free MI decoding. The suggested network combines a Temporal HyperMixer Block for capturing long-range temporal dependencies, a wavelet transformer for learning localized time–frequency representation, and a Graph Attention Network for EEG topology-aware spatial reasoning. Additionally, a Dynamic Residual Attention Gate (DRAG) has been developed to adaptively merge heterogeneous feature streams, and a compact subject-aware normalization (SAN) method enhances cross-subject generalization without the use of labeled target-domain calibration data. Our proposed model was tested following the rigorous leave-one-subject-out (LOSO) approach on BCI Competition IV-2a and High-Gamma datasets. NeuroCrossNet reached a classification accuracy of 91.30%, surpassing several strong benchmark methods, including CNN-LSTM, EEGNet, DeepConvNet, spectral CNN, and graph-based EEG decoding frameworks. Furthermore, a large number of ablation studies reveal that the integration of temporally, spectrally, and spatially complementary representations considerably boosts robustness and inter-subject consistency. Full article

The growing diversity of wireless standards and complex real-world channel effects render automatic modulation classification (AMC) increasingly challenging for spectrum monitoring and edge intelligence. However, most competitive deep-learning-based AMC networks still require ${10}^5$–${10}^6$ parameters, exceeding the memory available on resource-constrained edge platforms. We propose LGP-Net, a lightweight gated-fusion network that pairs a physics-informed expert branch with a compact temporal encoder built from depthwise separable convolution (DSConv), squeeze-and-excitation (SE) attention, and a single-layer gated recurrent unit (GRU). Specifically, unlike other dual-branch structures that directly concatenate the outputs of both pathways, this work designs a lightweight gating unit that requires no external signal-to-noise ratio (SNR) labels and adaptively reweights the two pathways according to signal-quality degradation. With fewer than 40 K parameters, a peak activation footprint of 26.00 KB and an amortised inference latency of 9.7 $\mathsf{\mu }$s per sample under GPU acceleration, LGP-Net attains 65.00% overall accuracy on RadioML 2016.10B (91.48% at 0 dB) and 62.76% on RadioML 2016.10A, placing it in a competitive accuracy–efficiency regime relative to architectures consuming 5× to 500× more parameters. These characteristics support deployment-oriented feasibility under memory-constrained edge settings and high-throughput spectrum-monitoring pipelines. Full article

Off-road semantic segmentation is challenging due to irregular terrain, vegetation clutter, class-level similarity, and ambiguous boundary annotations. Existing decoder designs often rely on compact bottlenecks that oversmooth fine structures or repeated multi-scale fusion that can amplify annotation noise and increase computational cost. To address these limitations, we propose a Cross-Scale Decoder for robust off-road semantic segmentation. The proposed decoder first stabilizes semantic representations through Global–Local Token Refinement (GLTR) on a compact bottleneck lattice. It then selectively incorporates fine-scale structural cues using Boundary-Guided Correction (BGC) and Gated Cross-Scale Interaction (GCS), avoiding dense and repeated feature fusion. In addition, uncertainty-guided class-aware point refinement focuses computation on ambiguous and low-confidence regions. Experiments on standard off-road benchmarks demonstrate that the proposed method improves segmentation accuracy and boundary consistency over existing approaches while maintaining practical inference efficiency. Full article

In terms of the detection of colorectal polyps during a colonoscopy, the accuracy of the diagnosis is key to effective prevention and treatment, and can be hindered by manual identification. Colorectal polyps are abnormal tissue growths in the colon or rectum, and their sizes, shapes and textures can make them difficult to find. Researchers have now turned to deep learning techniques and the YOLOv11 detection framework in particular to provide a method to automate the recognition and accurate identification of these abnormal growths. Specifically, the proposed method modifies the conventional YOLOv11 detection workflow by generating bounding box annotations from polyp segmentation masks, applying region-aware data preprocessing and augmentation, and training the detector under region-guided supervision to enhance localization precision and detection robustness. polyp segmentation masks are utilized to generate bounding box annotations which not only contribute exact spatial supervision but also avoid manual box labeling inconstancy. Region-aware data preprocessing and augmentation pay more attention to polyp-relevant regions and suppress background noise, which leads to clearer feature discrimination for small or irregular polyps. Additionally, region-guided supervision serves as explicit guidance for localizing objects with the anatomical polyp regions, which largely helps achieve accurate boundaries and prevent false detections. The proposed YOLOv11-based polyp detection system was tested and evaluated on the publicly available Kvasir-SEG dataset, which is comprised of annotated colonoscopy images. Enhanced data pre-processing and exhaustive training with appropriate choice of hyper-parameters fortified the reliability and useability of the model. The results confirmed high-grade results, and gave an Intersection over Union score of 0.9764, and an overall correctness rate of 99.00%, with well-balanced precision, recollection and F1-scores. Coming in with a mean Average Precision (mAP) of 0.9937 at a Intersection over Union threshold of 0.5 and 0.9935 over the full spectrum of thresholds from 0.5 to 0.95, this shows that the model is able to consistently and reliably detect polyps. The proposed system was also compared with Segment Anything Model, YOLO-Seg, and SAM2 and confirmed the efficacy of its method. Full article

In automotive crash testing, trustworthy crash-test curves are essential for reliable crashworthiness assessment, yet automated anomaly detection is difficult due to limited labeled abnormal cases, event-level data scarcity, and distribution shifts across vehicle models and sensor configurations. This paper proposes MVCA-AD (Multi-View Context Augmentation for Anomaly Detection) for single-channel crash-test curves. MVCA-AD generates multiple context-rich views using deterministic time- and frequency-domain transformations to amplify subtle anomalous patterns under limited labeled supervision. A trend-aware modulation module and cross-view attention fuse these views to improve sensitivity to critical segments such as impact spikes and gradual transitions while remaining robust to noise. Experiments on three subsets derived from physical full-scale crash tests show that MVCA-AD improves Precision, Recall, F1-score, and area under the ROC curve (AUC) over strong baselines and achieves stable performance under event-level grouped evaluation across heterogeneous head and B-pillar crash-test signals. The proposed approach supports crash-test data quality control by automatically identifying abnormal curves for downstream crashworthiness assessment workflows. Full article

Background: Complete blood count (CBC) phenotypes are routinely assigned using deterministic rule-based thresholds. While operationally efficient, such rules may lead to unstable phenotype assignments for results close to clinical cutoffs in the presence of analytical variability. Methods: We analyzed routine CBC data from a tertiary care hospital laboratory. Rule-based phenotypes for anemia subtype, white blood cell (WBC) status, and platelet (PLT) status were assigned using established laboratory thresholds. A patient-independent development and holdout split was applied. A multi-output gradient boosting model was trained to reproduce rule-based labels and provide probabilistic outputs. Phenotype stability was assessed by perturbing CBC parameters under realistic analytical noise. Instability was defined as any change in phenotype assignment across perturbations. Distances to decision boundaries were grouped into quantile-based bins. Model uncertainty was evaluated for the triage of unstable cases. Results: Phenotype instability was strongly concentrated near decision boundaries. Under medium analytical variability, samples closest to hemoglobin cutoffs exhibited the highest instability, with the highest instability in the bin closest to the cutoff, a sharp decrease in the adjacent bin, and lower instability across more distant bins. Model uncertainty was enriched among unstable cases, enabling prioritization of borderline samples while reviewing only a subset of all cases. Conclusions: Rule-based CBC phenotyping exhibits intrinsic instability near decision thresholds. Uncertainty-aware machine learning supports a practical framework to identify and prioritize borderline cases without replacing existing laboratory rules, supporting workload-controlled post-analytical decision support. Full article

Individual tree crown (ITC) detection from remotely sensed data is fundamental to forest inventory and ecological monitoring, but deep learning approaches remain constrained by limited labelled training data. We systematically evaluate three transformer detectors (the Detection Transformer (DETR), Deformable DETR, and DETR with Improved DeNoising Anchor Boxes (DINO)) paired with two backbones, ImageNet-pretrained ResNet-50 and a Masked Autoencoder (MAE) pretrained on unlabelled Canopy Height Model (CHM) data. These are benchmarked against a classical local maximum and watershed pipeline and Faster R-CNN across four test sets spanning boreal, temperate mixed-wood, and diverse North American forest types at 0.25–1.0 m resolution. Spatially held-out test regions with a one-patch buffer band eliminate sliding-window leakage; headline configurations are reported as mean ± standard deviation across three random seeds. With multi-resolution MAE pretraining, the practical lower bound for non-degenerate single-dataset transformer detection lies between ∼200 and ∼1200 patches. Without MAE pretraining, DETR fails at every dataset size we tested. Multi-dataset joint training reaches $F_1=0.84\pm 0.01$ on the boreal test set and 0.45–0.68 across the temperate-mixed-wood and NEON test sets, while Faster R-CNN narrowly wins on the smallest training pool. Standard DETR with ResNet-50 collapses regardless of the length of training schedule, but the same architecture with an MAE backbone reaches $F_1=0.83\pm 0.01$ at that schedule, showing that DETR’s reported instability is conditional on the combination of backbone initialization and training budget rather than architectural. Resolution and backbone interact: ResNet-50 wins at 0.25 m, and MAE wins at 0.5–1.0 m, consistent with the eight-pixel MAE patch-matching crown scale only at coarser resolutions. Full article

Background: Mass spectrometry imaging (MSI) enables in situ mapping of metabolite distributions within tissues, and spatial clustering is a key step for delineating metabolically distinct regions. Nevertheless, spatial clustering methods have not been systematically benchmarked for spatial metabolomics data. Methods: Here, we evaluated the effects of ion filtering and clustering method selection on clustering performance and established a dual-metric framework that jointly assesses the spatial continuity of cluster labels and inter-cluster metabolic heterogeneity. We benchmarked 30 clustering algorithms across 12 heterogeneous MSI datasets spanning three major ion sources, four mass analyzers, and multiple spatial resolutions, covering approaches from non-spatial methods to advanced spatially aware models. Results: Noise filtering markedly improved the spatial continuity of results generated by non-spatial methods (mean improvement, approximately 28%) but provided limited benefit for spatially aware methods. Across the 12 datasets, a median of only 11 methods satisfied both evaluation criteria simultaneously, whereas SSC and DRSC met the dual-metric thresholds in at least nine datasets. In the mbrain2_pos50 dataset, the top-ranked method based on the composite dual-metric score achieved 22% higher concordance between cluster assignments and cell-type annotations than the lowest-ranked method. Conclusions: Together, the proposed evaluation framework and the online platform SMcluster provide a standardized resource for benchmarking and selecting MSI clustering methods. Our results highlight the critical roles of preprocessing and method selection in determining spatial clustering performance and offer practical guidance for spatial metabolomics studies. Full article

In semi-supervised learning scenarios, the presence of limited labeled data and abundant unlabeled samples poses significant challenges to model robustness and generalization. Although the semi-supervised broad learning system (SSBLS) effectively exploits manifold structure through graph Laplacian regularization, its optimization is typically formulated under the mean square error (MSE) criterion, which is sensitive to noise and outliers. To address this limitation, this paper introduces the maximum mixture correntropy criterion (MMC) into the SSBLS framework and proposes a model termed M2C-SSBLS. By replacing the conventional MSE loss with a mixture correntropy-based objective, the proposed method enhances robustness against non-Gaussian noise and abnormal samples while preserving the computational efficiency and analytical solution property of the BLS. Furthermore, to improve representation diversity and reduce model variance, a multi-view ensemble extension, named EC-SSBLS, is proposed. This method constructs multiple feature views through a random feature subspace strategy, and independently trains an M2C-SSBLS base learner on each subspace. Finally, the predicted results of each view are fused through a voting mechanism. Experiments on benchmark UCI datasets under noise-free, 10% and 20% label noise settings demonstrate that the proposed M2C-SSBLS consistently outperforms conventional SSBLS and other advanced semi-supervised learning approaches. The ensemble extension EC-SSBLS further enhances performance, particularly in noisy environments, validating the effectiveness of combining MMC-based optimization with multi-view ensemble learning. Full article

Underwater image enhancement is essential for marine visual perception tasks. However, the highly heterogeneous optical degradations in real-world waters, the scarcity of paired training data, and the inherent dilemma for existing models in balancing long-range dependency modeling with computational overhead pose significant challenges. To address these issues, this paper proposes a prior-guided, self-supervised underwater image enhancement framework called Scene-Aware Retinex Mamba (SARM). This framework seamlessly integrates Retinex theoretical priors with state space models (SSMs) and operates without paired supervision by employing a prior-guided pseudo-labeling strategy to guide network optimization. Architecturally, SARM deeply couples the physical Retinex prior with SSM. Its core module integrates multi-color space features and leverages a 2D selective scan mechanism to achieve global context modeling with linear complexity $\mathcal{O}\left(HW\right)$, effectively removing complex color casts and suppressing non-uniform scattering noise. To further overcome the generalization bottlenecks in cross-domain underwater testing, this paper introduces a Scene-Aware Adapter (SAA), which facilitates dynamic loss scheduling and adaptive feature gating by quantifying scene-specific degradation characteristics. Comprehensive evaluations on multiple benchmark datasets, including UIEB, EUVP, and UCCS, demonstrate that SARM achieves state-of-the-art subjective and objective enhancement quality (e.g., yielding a URanker score of 2.491 and a CCF score of 35.76), while maintaining an ultra-fast inference speed of 136.52 FPS on the UIEB dataset. Furthermore, extended experiments reveal that SARM can significantly boost the performance of downstream vision tasks, validating its potential as a robust preprocessing module for various practical marine vision applications. Full article

Co-channel interference (CCI) remains a critical factor affecting link reliability in narrowband wireless systems, especially in scenarios with intensive frequency reuse, overlapping coverage, and dense terminal access. Existing interference detection methods are either computationally simple but insufficiently sensitive to short-term spectral variations, or highly accurate but dependent on labeled data and nontrivial inference resources. To address this issue, this paper proposes a lightweight CCI detection method in the FFT domain based on spectrum-jump analysis. The proposed method does not rely on absolute power growth as the primary interference indicator. Instead, it tracks the temporal inconsistency of dominant spectral-bin indices across consecutive FFT frames and converts recurrent peak-bin migration into an interference decision through a short-window counting mechanism. The method is computationally efficient, interpretable, and suitable for real-time deployment without offline model training. SDR-based measurements are combined with controlled repeated experiments to assess detector performance under varying signal-to-noise ratio (SNR), interference-to-signal ratio (ISR), carrier-frequency offset (CFO), multi-peak ambiguity, and two-path Rayleigh fading conditions. On the measured SDR record, the proposed method captures all interference-positive windows after the marked onset, while the controlled SNR/ISR experiments yield an overall detection probability of 96.0% over 250 CCI trials with no false alarms over 250 normal trials. ROC and precision–recall analyses further show that the selected threshold lies within a broad validation plateau. The results also reveal clear applicability boundaries: when the CFO approaches zero, when the interference is very weak, or when multiple stationary peaks have nearly equal power, dominant-bin migration may be weak or ambiguous. Therefore, the proposed approach is a low-complexity online detector for CCI cases that induce observable FFT-bin instability, and it can also serve as a front-end trigger for more advanced interference analysis modules. Full article

Intelligent mechanical fault diagnosis plays a key role in maintaining rotating machinery. Although data-driven unsupervised domain adaptation methods have achieved considerable progress, their industrial applications are often restricted by low-quality sensor data. Non-stationary vibration signals and background noise easily corrupt target pseudo-labels, while conventional methods focusing on global statistical matching usually neglect local structures, leading to confirmation bias under dynamic loads. To improve diagnostic reliability, we propose a Noise-Resilient Whitened Domain Adaptation (NRWDA) framework. To handle covariance fluctuations caused by changing working conditions, a Lipschitz-bounded Temporal Whitening (LTW) module is designed as a low-pass filter. An Entropy-guided Prototype Truncation (EPT) mechanism is adopted to discard ambiguous labels and better calibrate semantic centers. In addition, a Dispersion-Adaptive Contrastive Sharpening (DACS) strategy is introduced to dynamically adjust the contrastive temperature based on predictive dispersion, thus tightening decision boundaries. The proposed method is evaluated on CWRU, PU, and MFPT datasets. The PU dataset, featuring fluctuating loads and non-stationary signals, poses a strict test, yet our model maintains its stability even at a 0 dB SNR—a condition where standard approaches usually break down. During the $P0\to P3$ transfer task involving substantial radial force variations, NRWDA secures a 72.36% accuracy and surpasses established baselines. These findings confirm that our technique successfully isolates dependable diagnostic features from corrupted sensor measurements within actual industrial settings. Full article

Intrusion detection system (IDS) research relies on accurately labeled network traffic datasets; however, label quality in IDS datasets is seldom audited prior to modeling. Many publicly available IDS datasets assign ground-truth labels based on capture filenames or temporal session windows rather than per-flow inspection, a practice referred to as file-level labeling. This study identifies and corrects a systematic mislabeling instance in SDN-MG25, a CICFlowMeter-based dataset for software-defined networking (SDN)-enabled microgrid intrusion detection. IP composition analysis, which cross-references each attack-labeled flow with the documented attacker IP address, reveals that the BackgroundAttackTraffic (BAT) class, comprising 3167 flows (79.5% of all attack labels), contains no attacker-originated traffic. All BAT flows involve legitimate microgrid hosts communicating with external services during the attack capture window. Correcting this labeling error increases binary detection F1 from 0.578 to $0.956\pm 0.005$, an improvement of $+0.378$ that is 4.2 times greater than the best single modeling improvement (threshold tuning, $+0.090$). Furthermore, Confident Learning, a state-of-the-art automated label-noise detector, recovers only 8.4% of mislabeled BAT flows (recall $=0.084$, precision $=0.247$), indicating that domain-knowledge audits are essential for detecting systematic, class-level mislabeling that statistical methods cannot identify. The end-to-end pipeline Macro F1 improves from 0.749 to 0.862 after label correction. IP composition analysis is proposed as a mandatory prerequisite for IDS dataset evaluation, and a reproducible two-stage pipeline with feature-tier ablation for session confound diagnosis is provided. Full article

Real-time tracking of high-speed targets in autonomous systems requires detection and decision-making pipelines that can operate within sub-millisecond time budgets. Single Photon Avalanche Diode (SPAD) sensors are well suited for this task, offering 10 kHz Time-of-Flight (ToF) measurements with picosecond timing precision. However, processing such high-frequency time-series data with conventional deep learning models introduces computational bottlenecks that are difficult to handle on resource-constrained embedded hardware. This paper presents an ultra-lightweight, dual-head architecture built on the MiniRocket transformation algorithm, where a single shared feature extractor simultaneously feeds two independent decision pathways: one for multi-class target classification and one for 3-parameter kinematic regression covering velocity, pitch, and yaw. As a single-pixel sensor, the device provides only 1D range information; lateral 3D spatial localization is outside the scope of this work. To the best of the authors’ knowledge, this is the first application of MiniRocket to continuous kinematic estimation from high-frequency sensor data. Since collecting labeled physical flight data at these speeds is largely infeasible, a physics-based ray-casting simulation was developed to generate a 55,440-sample dataset across four 3D CAD target models under varying speed (100–450 m/s), orientation, and noise conditions. The proposed architecture achieves 98.6% classification accuracy and a velocity Mean Absolute Error (MAE) of 0.26 m/s, with orientation estimation yielding a pitch MAE of 3.47° and a yaw MAE of 2.46°—values consistent across all five cross-validation folds, indicating that the orientation performance floor is governed by the sensor’s physical angular resolution rather than by model capacity. With approximately 27,000 trainable parameters, the system completes full dual-task inference in 0.56 ms on a 16-core CPU (1785 Frames Per Second-FPS), satisfying the 1 ms real-time constraint of a 10 kHz sensor without GPU acceleration. It should be noted that the single-pixel SPAD architecture provides only 1D range-along-beam information; full 3D spatial localization is physically not extractable from a single sensor and is not addressed in this study. Full article