Search Results (635)

Fabric defect detection plays a vital role in the quality control of the textile manufacturing industry. However, it remains challenging because of defect diversity, complexity, and environmental factors. Deep learning-based methods efficiently extract visual features, improving detection accuracy and inference speed. However, most deep learning methods fail to adequately capture tiny, irregular, and blurred-edge defect features due to diverse fabric texture backgrounds and complex defect traits. To address these issues, we propose CMNet, a novel hybrid CNN–Mamba detection network with a dual-branch, heterogeneous feature-extraction architecture. First, we propose an Adaptive Weighted Adjacent Context Coordination Module (AWA-CCM) to enhance features of tiny fabric defects and edge textures while effectively suppressing interference from complex backgrounds. Second, we propose a Multi-Scale Large-Kernel Attention Mamba (MLAMamba) module to improve the network’s global representation capability for defects with variable scales and irregular shapes. Finally, we construct a Bidirectional Fusion Module (BFM) to dynamically balance local detail information and global structural information for efficient, complementary feature fusion. Experimental results on our self-developed fabric datasets, obtained using six widely adopted metrics, show that CMNet significantly outperforms state-of-the-art methods. Full article

Cyber incidents can disrupt many virtual power plant (VPP) assets through shared software and communication services. This study links preventive finance, cyber defense, dispatch, and restoration in one multi-timescale model. An attacker, a VPP operator, a bond vehicle, and a regulator interact in a partially observable stochastic game. The operator controls hardening, dispatch, isolation, and recovery. The bond provides restricted pre-event capital and releases collateral through an auditable index of service loss, control availability, network stress, and recovery delay. A risk-constrained multi-agent policy enforces power-system feasibility, investor impairment, sponsor affordability, and trigger–loss limits. Tests use transparent synthetic VPP-39 and VPP-118 portfolios and 20 out-of-sample seeds. The proposed design lowers normalized social cost to 0.691 and 0.704 and weighted basis risk to 0.065 and 0.071. It also improves critical-load continuity and restoration relative to self-insurance and three bond baselines. The VPP-118 case recovers in 11.8 h, compared with 14.3 h for the closest rule-based benchmark. Ablations separate the effects of finance and control. Removing the coupon–control link reduces verified hardening from 0.672 to 0.519. Removing the safety projection raises unsafe proposals from 0.4% to 5.9%. These results show that stochastic multi-timescale control can support adaptable and resilient VPP operation while keeping the financial mechanism within explicit risk limits. Full article

The vestibular apparatus is substantially organized before birth, whereas human equilibrium emerges only gradually through head control, sitting, standing, and independent walking. We propose an Innately Constrained Bayesian Model of Vestibular Development in which a phylogenetically shaped architecture constrains an initial model space; prenatal and postnatal experience calibrate internal models of self-motion and gravity; and developmentally gated plasticity modulates the rate of updating. We explicitly distinguish established evidence from interpretations and model-derived hypotheses. Recent human and nonhuman primate work on internal models, active self-motion, and multisensory recalibration is integrated with developmental and clinical evidence. The framework is then tested conceptually across congenital and acquired vestibular loss, altered gravity, and aging. Its proposed novelty lies not in Bayesian inference or internal models themselves, but in combining phylogenetic constraints, prenatal calibration, a qualitative transition at birth, time-dependent plasticity, and life-span context switching within a single developmental account. Finally, we operationalize the model through measurable predictions involving VOR adaptation, sensory-weighting coefficients, anticipatory postural responses, context-specific readaptation, and postural stability. The framework is intended as a falsifiable perspective rather than a definitive mechanistic account. Full article

Reliable abnormal behavior recognition from surveillance videos is hindered by the high cost of clip-level annotation, the scarcity of abnormal samples, and the context-dependent nature of behavioral semantics. Although vision–language models offer strong semantic transferability, their application under limited supervision remains susceptible to noisy pseudo-labels and confirmation bias. We propose confidence-aware semi-supervised vision–language contrastive learning (CA-VLC), which jointly exploits limited labeled videos and abundant unlabeled videos. Building on an existing CLIP-initialized temporal backbone, CA-VLC combines behavior-only and context-enriched text prototypes through confidence- and agreement-guided semantic fusion. For unlabeled videos, the model generates predictions from weakly augmented views and selects reliable pseudo-labels using entropy-based confidence estimation and class-adaptive thresholds. Detached weak-view targets then supervise strongly augmented views through confidence-weighted self-training without requiring an additional teacher network. Furthermore, cross-view consistency regularization and confidence-aware contextual alignment suppress unreliable semantic cues and improve robustness to contextual noise. Experiments on CABR50 demonstrate consistent improvements across multiple labeled-data ratios, while evaluations on CABRZ6 and UCF-101 assess prompt-based transfer to predefined target label sets without target-domain fine-tuning. With 10% labeled videos, CA-VLC achieves 84.06% Top-1 accuracy and 83.51% Macro-F1, retaining 95.47% of its fully supervised Top-1 accuracy of 88.05%, thereby demonstrating its effectiveness for label-efficient abnormal behavior recognition. Full article

This paper presents a systematic literature review of intelligent and advanced control methods for automotive wheel-slip regulation in anti-lock braking systems (ABS). Following a PRISMA-based workflow, records from IEEE Xplore, Scopus, and Engineering Village were searched for the 2014–2025 period. The search returned 5034 records; after removal of duplicates, patents, and out-of-range items, 1803 records were screened, 406 reports were assessed at full-text level, and 60 studies met the final eligibility criteria. The included studies were classified using a six-class taxonomy: fuzzy and neuro-fuzzy control, adaptive and self-tuning control, robust nonlinear control, predictive and optimization-based control, learning-assisted control, and hybrid or integrated control. The review shows that robust nonlinear and hybrid/integrated methods dominate the evidence base, while fuzzy and predictive methods remain important recurring families. Validation is still strongly simulation-weighted: many studies report braking-performance gains in slip tracking, stopping distance, chattering reduction, or road-friction robustness, but comparatively few provide hardware-in-the-loop, laboratory, or processor-level evidence. Only a very limited subset reports concrete embedded metrics such as execution time, sampling-period compliance, memory usage, or deadline margin. By combining systematic study selection, braking-specific control equations, primary-method classification, validation coding, embedded-implementation assessment, and comparative-scope analysis, this review identifies a persistent gap between algorithmic ABS performance and deployable real-time embedded feasibility. The findings motivate standardized benchmarking of intelligent braking controllers under common wheel-slip scenarios, common plant models, and common embedded timing metrics. Full article

In recent years, the use of UAVs has become increasingly widespread, and the public safety risks posed by unauthorized UAV flights have become increasingly prominent, creating an urgent need for effective detection and identification of UAV targets. However, such targets are small in size, have low contrast, and exhibit an extremely low signal-to-noise ratio; conventional detection methods generally suffer from insufficient feature discrimination, missed detections, and false alarms in complex backgrounds. To address these challenges, this paper proposes a Frequency-Guided Cross-scale Refinement Network (FGCR-Net). Based on an encoder-decoder architecture, this network achieves end-to-end collaborative optimization through cross-layer feature fusion, side-channel prediction refinement, and frequency-domain background suppression. First, a multi-path selective cross-layer fusion module (SCFM) is designed. This module employs coordinated modeling via both channel and spatial paths, supplemented by adaptive weighting with learnable coefficients, to perform differentiated selective fusion of the encoder’s fine-grained features and the decoder’s semantic features, thereby bridging the semantic gap at jump connections; Second, we designed a Cross-Scale Adaptive Fusion Enhancement Attention Module (CAFEM), which cascades multi-receptive-field hollow convolutions, strip pooling, and a bidirectional semantic guidance mechanism to perform cross-scale refinement on the side outputs of each decoder layer, thereby alleviating the issues of blurred boundaries and false alarms caused by inconsistent quality of multi-scale prediction maps and insufficient cross-layer consistency; finally, we design a Frequency-Guided Semantic Enhancement Module (FGSEM), which uses the Fast Fourier Transform (FFT) to decouple encoder features into the frequency domain. By leveraging low-frequency energy to predict the background confidence map and applying spatially selective suppression to high-frequency components, this module distinguishes, from a frequency-domain perspective, the high-frequency responses of complex backgrounds and targets that are highly similar in the spatial domain. Experiments on MSDS-UAV, a self-built multi-scenario UAV dataset for small targets, demonstrate that our method consistently outperforms existing state-of-the-art methods across multiple performance metrics, with Pixel Accuracy, Mean Intersection over Union, and Probability of Detection reaching 92.76%, 70.91%, and 92.69%, respectively; Compared to the baseline model, these three metrics improved by 1.90, 3.20, and 3.76 percentage points, respectively, fully validating the effectiveness and superiority of the proposed method. Full article

Reliable human target detection in smoke-filled environments is essential for firefighting robots and rescue perception systems. However, conventional RGB cameras are severely degraded by dense smoke, while a single infrared or thermal imaging sensor cannot simultaneously provide sufficient structural details and reliable target-related thermal information. To address these challenges, this paper proposes a local variance-guided adaptive infrared–thermal sensor fusion framework for human target detection in smoke-filled environments, aiming to alleviate smoke-induced degradation in multimodal perception through improved infrared representation and adaptive cross-modal information utilization. An improved dark channel prior-based infrared desmoking algorithm is designed, where guided filtering is employed to refine the transmission map, suppress halo artifacts, and enhance infrared image quality. Furthermore, a local variance-guided adaptive fusion strategy is proposed, which utilizes local variance as an information saliency metric to generate pixel-level adaptive modality weights for fusing desmoked infrared and thermal images. In addition, a lightweight YOLO11n detector is adopted to achieve efficient human target recognition while maintaining a favorable balance among detection accuracy, computational cost, and inference efficiency. Experimental results on the self-built dense-smoke dual-modal dataset demonstrate that the proposed framework achieves high detection performance with low model complexity and efficient detector-stage inference. The ablation results demonstrate the contribution of infrared–thermal multimodal fusion to reliable smoke perception and indicate that the proposed local variance-guided adaptive fusion strategy maintains comparable detection accuracy while providing a better detector-stage speed–accuracy balance than fixed-weight fusion. With a low parameter count, the adopted YOLO11n detector shows potential for future deployment on resource-constrained firefighting robotic platforms. Full article

In complex inshore regions, ships are often densely berthed and arbitrarily oriented, while docks, shorelines, coastal facilities, and strong scatterers produce substantial background clutter and target-like scattering responses, leading to foreground-background confusion and localization ambiguity. In practical inshore synthetic aperture radar (SAR) ship detection, training and testing images often originate from different coastal scenes. Such cross-scene domain shifts further exacerbate these challenges, resulting in degraded detection performance and reduced pseudo-label reliability during self-training. To address these issues, this study proposes a task-decoupled alignment and pseudo-label optimization framework (TAPO) for cross-scene inshore SAR ship detection with oriented bounding boxes. First, the Task-Decoupled Imbalanced Feature Alignment module (TD-IFA) separately aligns classification- and regression-related region-of-interest features, allowing foreground-background discrimination and oriented-box localization to be adapted according to their different cross-scene shifts. Imbalanced source-target alignment weights are further introduced to reduce the influence of noisy target-scene proposals. Second, the Dynamic Uncertainty-Driven Pseudo-Label Optimization module (DUD-PLO) improves self-training reliability by selecting pseudo-labels that are both confident and stable. Dynamic confidence filtering combines a fixed threshold with an adaptive fallback threshold to suppress low-confidence noisy candidates while reducing the risk of obtaining empty pseudo-label sets in difficult target scenes. Cross-view consensus retains candidates that remain consistently matched across non-geometric perturbation views, while rotated-box geometric stability, measured by rotated intersection over union (rotated IoU), and uncertainty weighting further suppress pseudo boxes with unstable locations, scales, or angles. In addition, a Kullback–Leibler divergence (KLD)-based auxiliary loss provides soft geometric consistency for oriented boxes during training. Experiments on a self-constructed cross-scene inshore SAR ship dataset show that TAPO achieves a mean average precision (mAP) of 0.439, improving over the Source-only oriented-box baseline by 0.102 and achieving the best overall performance among the compared methods. Full article

Yaks are indigenous livestock species of the Qinghai–Tibet Plateau. Due to long-term exposure to extreme environmental stressors, including high altitude, hypoxia, and low temperatures, its male reproductive system has evolved distinct adaptive strategies. Specifically, testicular weight is markedly lower than that of cattle. Nevertheless, according to reported data from different experiments, although its single ejaculate volume is lower than that of Tibetan cattle, it can still maintain comparable fresh sperm motility. These reproductive phenotypes make the yak an ideal model organism for investigating plateau adaptation and sperm energy metabolism. The present review synthesizes continuous molecular events in yak testes spanning embryonic development through senescence. It focuses on dissecting the core regulatory networks governing spermatogonial stem cell self-renewal, meiosis, and spermiogenesis. It integrates the latest advances in testicular microenvironment dynamics and epigenetic modifications. Additionally, male sterility in cattle-yak serves as a natural mutant model. Essential regulatory modules governing spermatogenesis can be inferred from its spermatogenic-arrest phenotype. Nevertheless, comparative omics analyses between yaks and cattle are confounded by seasonal variation and differing genetic backgrounds. It is therefore critical to distinguish signatures driven by genetic adaptation from those arising from environmental plasticity. Finally, this paper presents research prospects regarding how to utilize single-cell multi-omics and gene-editing technologies to deeply dissect the underlying mechanisms of plateau reproductive adaptation to provide theoretical support for improving yak fecundity and hybrid breeding. Full article

The cutting head is the primary cutting load device of the roadheader. The harsh, complex excavation environment often leads to a scarcity of labelled fault samples, significantly hindering the development of accurate fault diagnosis models. This paper proposes a novel fault diagnosis method based on the Gramian angular difference field (GADF) and an attention-enhanced transfer-learning AlexNet. The collected one-dimensional vibration signals are transformed into two-dimensional image data using GADF to capture transient impact gradients and preserve absolute temporal correlations. To overcome data limitations, the method retains the base convolutional feature extractor of an AlexNet model pre-trained on ImageNet, whilst discarding the original fully connected and classification layers. A novel classification head incorporating a multihead self-attention (MSA) mechanism is constructed to fine-tune the network specifically for the fault diagnosis task. This structural modification adaptively assigns higher weights to fault-sensitive spatial regions, significantly enhancing the model’s feature extraction focus and generalisation capability even under intense background noise. Experimental validation was conducted on a scaled cutting head fault diagnosis test bench. The results demonstrate that the proposed method outperforms other baselines across evaluation metrics, exhibiting robust recognition accuracy and stability. This effectively identifies the cutting head’s operating condition, offering a novel and practical approach for future underground fault diagnosis in coal mines. Full article

We study a conditional finite-dimensional operator construction based on two imported Recognition-Science inputs: a continuous comparison cost satisfying a multiplicative coherence axiom and equilibrium calibration, and a three-coordinate, eight-tick ledger schedule. The first input uniquely determines the reciprocal cost $J(x)=\frac{1}{2}(x+x^{-1})-1$. A cyclic Gray traversal of the configuration graph $Q_3$ fixes the cyclic ordering of an assumed eight-sample register with shift P. The additional half-cycle antiperiodicity criterion selects the odd Fourier sector $\mathcal{V}=\ker (P^4+I)$; unitarity alone does not select this sector because P is unitary on the full register. On the realification of $\mathcal{V}$, $P^2$ defines an additional complex structure, and, after assigning an independent beat duration, the principal logarithm defines one self-adjoint stroboscopic interpolation, unique only after imposing the principal-zone convention. We distinguish the exact reciprocal action from a local adapter hypothesis needed to relate its mismatch coordinates to Hilbert-space defect coordinates. For the regular shift, the rational cyclotomic decomposition identifies $\mathcal{V}$ as the minimal faithful ${\mathsf{\Phi }}_8$-block. Under the stated relative normalization of the local bridge, leading cost–defect agreement is equivalent to an isometric adapter tangent, while a positive self-adjoint single-filter commit that commutes with the beat has at most three non-negative mode factors on the minimal register; defect descent restricts them to $[0,1]$. The relaxed commit is a contraction and, in a normalized quantum interpretation, represents a trace-nonincreasing accepted branch whose complementary trace weight is recorded by a measurement outcome or environment; the in-sector component nevertheless evolves exactly and unitarily. For coupled ququart cores, adjoining the standard conjugate Weyl shift to the clock inherited from P spans the full finite-dimensional operator algebra. This proves algebraic representability of arbitrary finite-dimensional Hermitian dynamics, not a physical rule selecting their coefficients. Full article

Dynamic inductive charging (DIC) combined with hybrid energy storage systems (HESSs) and vehicle-to-grid (V2G) capabilities offers a promising pathway toward extended-range electric vehicles with grid integration benefits. However, real-time optimal energy management remains challenging due to multi-axis coil misalignment, component aging, and bidirectional power flow uncertainty. This paper proposes an adaptive digital twin driven artificial intelligence (AI) energy management framework integrating physics-informed neural networks (PINNs), soft actor critic (SAC) deep reinforcement learning, and model predictive control (MPC) for optimal power distribution among a proton exchange membrane fuel cell (PEMFC), lithium-ion battery, supercapacitor, dynamic wireless charging, and grid interface in four-wheel drive electric vehicles (4WD-EVs). The framework features: (1) a self-evolving digital twin with online learning via Elastic Weight Consolidation (EWC) updating every 50 cycles; (2) a PINN-based state estimator for battery-state estimation, with an average inference time of 1.1 ms and a worst-case latency of 2.8 ms; (3) a hierarchical SAC–MPC strategy with high-level mode selection and low-level power optimization; (4) real-time five-degree-of-freedom WPT misalignment compensation, achieving a mean efficiency of 91.5% under the evaluated dynamic lateral misalignment conditions, with a 50 mm displacement amplitude; (5) degradation-aware V2G optimization generating €582.50/year in revenue while reducing battery aging by 31.8%; and (6) comprehensive techno-economic analysis yielding a discounted payback period of approximately 5.57 years and a net present value of approximately €3777 over a 10-year horizon. Validated through 200+ hours of hardware-in-the-loop (HIL) simulation on the dSPACE/NVIDIA Jetson platform, the proposed approach achieves a 24.3% cost reduction and 31.8% lower battery degradation. The MPC controller exhibits an average execution time of 32.1 ms, a 95th-percentile latency of 44.8 ms, and a worst-case latency of 62.4 ms, while remaining within the 100-ms real-time control deadline. Results demonstrate the viability of adaptive digital twins for next-generation EVs with autonomous charging and multi-source architectures. Full article

Neuromorphic imaging sensors (event cameras) offer a promising paradigm for computational imaging and human pose estimation (HPE) under extreme illumination conditions. Nevertheless, dark-scene background activity originating from photodiode dark current and circuit thermal noise, together with hot-pixel noise, severely corrupts event streams and impedes reliable HPE in low-light scenarios. To address this issue, we propose an adaptive event denoising framework built upon a spatiotemporal Gaussian-weighted neighborhood model with a dynamic thresholding mechanism. It can effectively suppress background activity and hot-pixel noise while preserving edge and motion details critical for pose estimation. Leveraging this denoising front-end, we construct a complete dark-scene neuromorphic HPE pipeline by transferring the pre-trained MediaPipe model onto event-based time-surfaces. Quantitative and qualitative evaluations on public and self-collected datasets demonstrate that our approach outperforms state-of-the-art denoising methods with an improvement of over 20% on public benchmarks and over 30% on self-collected dark-scene data. We expect our work to pave the way toward reliable dark-scene human–robot interaction through robust neuromorphic pose estimation. Full article

Rapid weight loss (RWL) is commonly practiced among combat sports athletes to reduce body mass before competition. The Rapid Weight Loss Questionnaire (RWLQ) is a self-report instrument designed to assess weight loss behaviors and practices. This study aimed to translate and culturally adapt the RWLQ into Greek (RWLQ-GR), evaluate its psychometric properties, and examine the prevalence, magnitude, and methods of RWL among Greek-speaking combat sports athletes. The translation and cultural adaptation process followed established international guidelines. Measurement properties were examined in terms of test–retest reliability, known-groups and convergent validity. A total of 209 Greek-speaking combat sports athletes were included in the final analysis, of whom 34 completed the questionnaire twice for the test–retest reliability assessment. Overall, 92% of participants reported engaging in weight loss before competition, with a median body mass reduction of 4.5%. The RWLQ-GR demonstrated high test–retest reliability for numerical items and total score. Categorical items showed fair to excellent reliability. No systematic differences were observed between repeated administrations of the questionnaire. The questionnaire differentiated athletes according to their usual weight loss, providing supportive evidence for known-group validity. Evidence of convergent validity was limited: the RWLQ-GR total score was not associated with the EAT-26 total score and correlated weakly with the Bulimia and Food Preoccupation subscale only. Overall, these findings support the use of the RWLQ-GR for the assessment and surveillance of self-reported RWL behaviors in Greek-speaking combat sports athletes. Full article

RF fingerprints of power equipment drift over time due to environmental changes and device aging, which progressively degrades the performance of authentication systems built on fixed models. Existing incremental learning methods tend to either forget historical devices or fail to adapt to new distributions when dealing with such drift. We propose an incremental update strategy tailored to this scenario and evaluate it on a self-constructed 64-dimensional simulated drift dataset. At each update, gradient importance per feature channel is computed from the current batch, and three feature-level distillation terms (channel MSE, covariance alignment, and spatial attention) keep the new model’s representations close to the old one. The distillation strength decays exponentially with update steps, enabling strong preservation of old knowledge early and more flexible adaptation later. A small memory buffer mixes old samples into each training batch to further reinforce historical recognition. On the synthetic dataset, our method achieves higher final historical accuracy than static, fine-tuning, EWC, and LwF baselines. Ablation studies confirm that the multi-level distillation, dynamic decay, and momentum-smoothed channel weights each contribute positively. These preliminary simulation-based results indicate that the method shows promise in alleviating forgetting caused by fingerprint drift while maintaining adaptability to new fingerprints within the synthetic evaluation framework. Full article