Search Results (1,821)

Accurate fruit ripeness assessment is essential for optimizing harvest timing and ensuring postharvest quality, particularly in climacteric fruits such as peaches, which exhibit rapid and heterogeneous ripening. This study proposes a deep learning-based approach for multistage peach ripeness classification across five maturity levels (green, green-blushed, blushed, yellow-blushed, and fully yellow). Four datasets were constructed using controlled image acquisition, segmentation, data augmentation, and perceptual hashing to mitigate data leakage. The performance of AlexNet, EfficientNet-B0, and three YOLO (You Only Look Once) architectures (YOLOv8, YOLOv11, and YOLOv12) was evaluated using standard metrics, including accuracy, precision, recall, F1 score, mAP, and inference speed. Results show that YOLO-based models significantly outperform classical networks, achieving accuracies between 95.25% and 98.3% and mAP@0.5 above 98.25%, while also reducing inference time to 8.1–12.7 ms compared with 722.23 ms for AlexNet and 171.87 ms for EfficientNet-B0. In a practical sorting experiment with 214 peaches, YOLOv12 achieved 92.06% accuracy, demonstrating robust real-world performance. Misclassifications were primarily observed between adjacent ripeness stages. These findings indicate that YOLO-based models provide an effective and scalable solution for real-time fruit sorting, while the use of perceptual hashing enhances dataset reliability and model generalization for deployment in agricultural quality control systems. Full article

Addressing the challenges of privacy leakage, fragmented data silos, and high computational overhead in traditional ciphertext-policy attribute-based encryption (CP-ABE) for medical data sharing, this paper proposes an improved CP-ABE framework with outsourced decryption, integrated with consortium blockchain and the InterPlanetary File System (IPFS). The framework introduces a medical-scenario-adapted CP-ABE architecture based on a lightweight FAME design, optimizing attribute key generation and transformation key design to accommodate resource-constrained medical terminals. A hybrid encryption system is employed, combining symmetric encryption for high-efficiency processing of large medical data and CP-ABE for fine-grained access control of symmetric keys. To reduce user computational burden, a proxy-assisted secure decryption architecture is implemented, where the proxy server handles most decryption tasks while ensuring resistance to malicious proxy behavior. Furthermore, the framework provides rigorous formal security verification, achieving IND-CPA security and resilience against collusion and malicious proxy attacks. Comprehensive performance evaluations demonstrate significant improvements in key generation, encryption, and decryption efficiency, offering a better balance between security and efficiency for practical medical data sharing applications. Full article

In order to explore the influence of baffle structure on the oil–gas two-phase flow and leakage characteristics of aero-engine bearing chamber sealing systems, based on the VOF two-phase flow model, this paper systematically carried out a transient numerical simulation of the bearing chamber sealing systems with conventional configurations and baffle configurations. The oil distribution, leakage and flow evolution of the two types of configurations under different baffle heights, sealing pressure differences and rotational speeds were compared and analyzed. The results show that the higher the height of the baffle, the more obvious the accumulation effect of the lubricating oil and the greater the leakage. The increase in sealing pressure difference helps to suppress leakage and reduce leakage fluctuation. The increase in rotational speed aggravates the centrifugal effect of the lubricating oil and makes the leakage increase significantly. This paper reveals the multi-parameter coupling mechanism of the baffle structure on the leakage control of the bearing chamber sealing system, and it provides a theoretical basis for the optimal design of the bearing chamber sealing structure of the aero-engine. Full article

As digital payment infrastructures expand, the incidence of card-not-present fraud has become a major source of operational and financial risk for banks, payment processors, and merchants. In response, financial institutions increasingly rely on data-driven decision systems, yet fraudsters continuously adapt their strategies to evade conventional rule-based controls. A promising way to strengthen risk management is to model transactional data so as to uncover non-trivial, high-dimensional patterns characteristic of fraudulent behavior and to embed these models into real-time decision pipelines. In this work, we develop and compare a suite of learning-based fraud detectors, including a convolutional neural network and several machine learning classifiers, within a unified quantitative risk-management framework. The problem is formulated as a supervised classification task within a quantitative risk management framework, where the cost of missed fraud is particularly critical. The mathematical contribution is methodological rather than architectural: we design a leakage-safe and prevalence-faithful evaluation protocol for extremely imbalanced binary classification, combine cross-validated hyperparameter optimization with risk-aligned model selection based on metrics such as recall and Matthews correlation coefficient, and quantify uncertainty by bootstrap confidence intervals and paired McNemar tests. In addition, we connect statistical evaluation with deployment-time decisioning through a decision-theoretic, cost-sensitive threshold rule, showing how institution-specific false-positive and false-negative costs determine the operating point of the classifier. Because fraudulent transactions constitute only a small proportion of the total volume, we employ resampling strategies to mitigate severe class imbalance and systematically calibrate the models via cross-validated hyperparameter optimization. The empirical analysis on real transaction data shows that carefully tuned deep and ensemble methods can achieve strong fraud-detection performance, while the proposed framework clarifies which performance differences are statistically meaningful and which operating points are most suitable under institution-specific false-positive and false-negative costs. Full article

Spiral-groove dry gas seals are widely used in various rotating machinery, and their performance prediction is of great significance for structural design and operational optimization. Existing studies still face several limitations, including the limited fidelity of numerical simulations, the insufficient number of experimental samples, and the restricted generalization capability of models based on a single data source. To address these issues, this study constructed a multi-source data system integrating numerical simulation data and experimental data, and systematically compared four representative data fusion methods, namely the uncertainty-weighted fusion algorithm, TrAdaBoost, MFDNN, and CoKriging, with analysis of their applicability and predictive performance. The results show that multi-source data fusion can effectively exploit the complementary advantages of different data sources and improve the prediction accuracy of dry gas seal performance. In terms of the comparison of data fusion methods, all four methods achieved good results for the groove-depth problem; however, for the spiral-angle and groove-number problems, which exhibit stronger nonlinear characteristics, clear differences were observed among the methods. Among them, TrAdaBoost showed the best overall performance, followed by MFDNN, then CoKriging, while the uncertainty-weighted method was relatively weaker. In terms of seal performance, the influence of groove depth on seal performance was relatively direct; the spiral angle is recommended to be controlled within 10–14°, and the groove number within 12–16, so as to balance opening force and leakage rate. This study can provide a reference for the rapid performance prediction and parameter optimization of spiral-groove dry gas seals. Full article

Downhole boosting technology breaks the physical limitations of conventional surface boosting by enhancing pressure at the wellbore bottom, with micro-axial compressors serving as its core compression module. However, traditional axial compressors, when miniaturized, suffer from severe end losses and easy instability, failing to adapt to downhole space constraints and the efficient pressurization demands of low-permeability, low-pressure, and small-flow reservoirs. To address this, this study designed a compact micro-axial compressor. CFturbo was used for parametric blade design and optimization, while ANSYS CFX 2025 (with the SST turbulence model) conducted numerical simulations. A “simulation–diagnosis–optimization–validation” closed-loop strategy was adopted to adjust the blade’s leading-edge shape, camber line, and thickness distribution, combined with grid independence verification and inter-stage matching optimization. The results show that at the design speed (60,000 rpm), the compressor achieves a pressure ratio of 1.57 and an isentropic efficiency of 83.6%. It also maintains stable performance at 55,000 rpm (off-design speed), with excellent inter-stage aerodynamic matching and controllable leakage losses. This compressor meets downhole operational needs, providing technical support for developing low-permeability, low-pressure, small-flow reservoirs. Full article

In smart manufacturing, the injection molding industry faces a “data scarce environment” due to prohibitive physical trial costs. Processing recycled polypropylene (rPP) exacerbates this challenge, as traditional response surface methodology (RSM) fails to capture complex non-linear rheological behaviors induced by material variability. This study proposes a “domain-knowledge guided data augmentation framework,” integrating Taguchi experimental data (L₂₅) with Moldex3D digital twin simulations to construct a 300-sample hybrid dataset. A back-propagation neural network (BPNN) with L₂ regularization was employed for small-sample learning, providing a continuous differentiable physical mapping. To rigorously prevent neighborhood data leakage, the model was evaluated via a strict nested group-based 5-fold cross-validation. Particle swarm optimization (PSO) was coupled to overcome the local minima of gradient descent. Comparative analysis demonstrates that BPNN significantly outperforms both traditional RSM and a newly introduced Random Forest (RF) baseline, achieving a testing mean squared error (MSE) of 0.001 (±0.0002) and a testing R² of 0.95. PSO minimized the shrinkage rate to 3.079%, validated via Moldex3D digital twin simulation with a 0.19% relative error. Synergizing virtual–physical integration with robust neural computing enables superior process control precision in small-data regimes, offering small and medium-sized enterprises (SMEs) a cost-effective pathway for smart optimization. Full article

As a common brain-computer interface (BCI) paradigm, electroencephalogram (EEG)-based motor imagery provides a critical pathway for both assistive technology to (restoring communication and control) and active rehabilitation (promoting neural plasticity and functional recovery). Domain adaptation has been shown to effectively enhance the decoding performance of motor intentions for target subjects by leveraging labeled data from source subjects. However, EEG data from source subjects often contains extensive personal privacy, and the direct access to source EEG data easily leads to privacy leakage issues. An important research topic is to achieve domain adaptation without directly accessing the source subjects’ raw data. To address this challenge, a privacy-preserving source-free domain adaptation framework, termed Transformer-based SFDA with Class-balanced Multicentric Dynamic Pseudo-labeling (T-CMDP), is proposed for cross-subject motor-imagery EEG classification. This framework consists of three coupled stages. In the source model training stage, a Transformer-based encoder combined with Riemannian manifold-aware feature extraction is employed to learn transferable and discriminative EEG feature representations. In the source-free target adaptation stage, only the pretrained source model is transferred to the target domain and adapted through knowledge distillation and information maximization, without accessing raw source EEG data. In the self-supervised learning stage, class-balanced multicentric prototypes and high-confidence pseudo-label updates are introduced to progressively refine the target-domain decision boundaries. Extensive experiments on three motor-imagery EEG datasets demonstrate that the proposed T-CMDP framework consistently outperforms eleven representative baselines from traditional machine learning, deep learning, and source-free transfer approaches, achieving average accuracies of 56.85%, 76.34%, and 74.49%, respectively. These results indicate that T-CMDP effectively alleviates inter-subject EEG distribution discrepancies and ensures the privacy preserving of source subjects, thereby facilitating more reliable and practical deployment of EEG-based BCI systems. Full article

This paper presents a novel Multi-Scale Temporal Uncertainty-aware Hierarchical Adaptive Ensemble (MSTU-HAE) algorithm for intelligent ship emission monitoring and prediction in maritime environmental compliance applications. The maritime shipping industry contributes approximately 3% of global CO₂ emissions and significant amounts of nitrogen oxides and sulfur oxides, necessitating advanced predictive monitoring systems. The proposed MSTU-HAE algorithm integrates three key innovations: multi-scale temporal feature extraction using causal convolutions at short-term (5 samples), medium-term (20 samples), and long-term (60 samples) windows; gas-specific attention mechanisms that automatically weight temporal scales based on individual emission gas characteristics; and three-level hierarchical uncertainty quantification encompassing individual model uncertainty, ensemble disagreement, and regulatory compliance risk assessment. Experimental validation was conducted using emission data collected from a fishing vessel over 3 operational days (1732 original samples), augmented to 17,320 samples via controlled replication with noise injection to support model training. Rigorous temporal data splitting with 70%/15%/15% train/validation/test partitioning ensures no data leakage. Comparative analysis against six baseline methods (XGBoost, LSBoost, AdaBoost, Ridge Regression, Random Forest, and K-Nearest Neighbors) demonstrates that MSTU-HAE achieves superior average performance, with R² = 0.9670 and NSE = 0.9670 across all emission gases. This research contributes a robust, interpretable, and scalable prediction framework that advances the state of the art in maritime environmental monitoring through novel algorithmic innovations in temporal feature learning and uncertainty quantification. Full article

Machine unlearning in federated graph learning must satisfy the multi-level indistinguishability requirement of the deletion of a target node being undetectable at the level of the global model, of the unlearning client’s local model, and of every non-target client’s local model. Approximate unlearning methods that pass confidence-based audits may still leave geometric traces through embedding drift at one or more of these $K+1$ levels. We formalize this requirement, introduce a five-model threat taxonomy, and extend the Hub–Ripple embedding drift audit to global, local, and cross-client levels. Across 31,900 trials spanning five graph benchmarks, five federated unlearning methods, and four supplementary ablations (K-value, cross-edge handling, control sampling, and DP-SGD defense), we find that all approximate methods fail the following multi-level requirement: the Confidence–Embedding Gap persists at $0.12$ (versus $0.35$ centralized), cross-client leakage correlates with shared cross-edge count ($r=0.56$, $p<{10}^{-160}$), and a federated participant outperforms a white-box external auditor (AUC $0.83$ versus $0.81$). Client-level unlearning is more detectable at the global level than node-level unlearning (AUC $0.81$ versus $0.77$), contradicting the intuition that coarser deletion yields stronger privacy. FedRetrain satisfies global and local indistinguishability but exhibits residual cross-client leakage (Cross-Mean L2 AUC $=0.62\pm 0.04$) because re-aggregation itself perturbs the global parameter vector. No method evaluated achieves full multi-level indistinguishability. Supplementary studies confirm that this is a structural property of FedAvg; DP-SGD reduces Cross L2 AUC by only $0.013$ at the cost of a $79\%$ accuracy drop, and FedSage-like neighbor sharing does not change the leakage profile. Multi-level geometric auditing, spanning all $K+1$ models, is the necessary evaluation floor that any method claiming verifiable privacy compliance must satisfy. Full article

We propose a terahertz (THz) antenna-coupled wire-channel field-effect transistor—modified EdgeFET (m-EdgeFET), formed by combining single-gate FinFET and dual-side-gate EdgeFET concepts, which is used for THz detection. The proposed hybrid design was implemented on AlGaN/GaN high-electron-mobility transistor (HEMT) structures, demonstrating distinct response characteristics under 150 GHz and 300 GHz radiation at room temperature. The responsivity dependence on the channel length was determined, revealing that the peak responsivity reached up to 6.5 V/W at a gate voltage of −3 V, i.e., at a gate bias that is an order lower in magnitude than that required for EdgeFET to reach the maximum response. Meanwhile, the gate leakage current decreased by an order of magnitude (to about 1 nA) compared to a FinFET with similar geometry. The proposed geometry was shown to operate in two regimes: source-drain coupling (SD) and gate coupling (GG) of THz radiation with the transistor wire channel. The results confirm that the m-EdgeFET design is suitable for electrically controlled and fast THz detection. Full article

The mathematical characterization of non-stationary signals remains a significant challenge, particularly when impulsive components are obscured by high-dimensional noise and structural coupling. This paper proposes an application-driven mathematical methodology for a learnable discrete wavelet transform (LDWT) that combines classical multi-resolution analysis with task-optimized data-driven adaptivity. Rather than introducing entirely new foundational theory, our approach strategically relaxes constraints from orthogonal wavelet theory within the non-perfect reconstruction filter bank framework, enabling controlled spectral decomposition optimized for supervised fault diagnosis. We introduce a specialized regularization term based on the half-band property to ensure spectral complementarity and minimize cross-band correlation, while a Jacobian-based stabilization approach is formulated to ensure the convergence of filter coefficients during optimization. The proposed algorithmic architecture, LDBRFnet, features a dual-branch encoder system designed to capture the mathematical synergy between sub-band-level global statistics and time-domain local morphology. This dual-view representation effectively mitigates feature leakage and overconfidence in classification. Theoretical analysis and numerical experiments demonstrate that the learned filters satisfy the frequency-shift property and maintain robust spectral partitioning even under low signal-to-noise ratios. Validation on complex vibration datasets confirms that the framework achieves superior diagnostic accuracy (over 95.5%) and computational efficiency, reducing model parameters by 96.7% compared to state-of-the-art baselines. This work provides a generalizable mathematical approach for adaptive signal decomposition and robust pattern recognition in interdisciplinary applications. Full article

Volatile halogenated hydrocarbons (VHHs) are critical air toxic pollutants, with some ozone-depleting substances (ODSs) strictly regulated by the Montreal Protocol. However, current understanding of the pollution characteristics, sources, and health risks of atmospheric VHHs in Southwest China remains insufficient. This study performed field observations of atmospheric VHHs in summer in Mianyang, a medium-sized industrial city in the Sichuan Basin. Freon-12 (563 ± 20 ppt) and Freon-11 (264 ± 15 ppt) were the most abundant chlorofluorocarbons (CFCs); chloromethane (785 ± 261 ppt) and methylene chloride (563 ± 505 ppt) dominated among VSLSs. The mean concentration of regulated ODSs (1037 ± 33 pptv) was notably lower than unregulated very short-lived chlorinated substances (1887 ± 745 pptv), reflecting effective ODSs phase-out locally, yet enhancements relative to Northern Hemisphere background implied potential leakage from residual tanks. Methylene chloride and trichloroethylene concentrations exceeded global background levels by over 10 times, indicating strong anthropogenic industrial influences. Phased-out CFCs displayed negligible diurnal variation due to stringent emission controls, whereas unregulated VSLSs exhibited a distinct U-shaped diurnal cycle, with peaks driven by morning boundary layer dynamics and evening accumulation. Positive matrix factorization revealed that industrial sources, including electronic solvents (28.6%), industrial processes (27.8%), and solvent usage (23.7%), accounted for 80.1% of total VHHs. The total carcinogenic risk (2.3 × 10⁻⁵) surpassed the acceptable threshold (1 × 10⁻⁶), dominated by 1,2-dichloroethane, chloroform, carbon tetrachloride, and 1,2-dichloropropane. All individual compounds exhibited mean hazard quotients (HQs) below the non-carcinogenic risk threshold. The cumulative hazard index reached 1.5, suggesting combined non-carcinogenic risks to the local population. These results support VHHs health risk management and ODSs control in Southwest Chinese industrial cities. Full article

Oil and gas drilling waste drilling fluid is a complex alkaline mixture that poses risks to plants and soil ecosystems during transportation and disposal due to potential leakage. This study investigates the effects of waste drilling fluid on the growth of Trifolium pratense (L.) and Astragalus sinicus (L.) and on the soil ecosystem, aiming to provide a theoretical reference for ecological restoration of oil and gas field sites. Four gradients of waste drilling fluid stress were established by mixing 0, 50, 100, and 150 mL of waste drilling fluid into the substrate, with 0 mL serving as the control. Seed germination, morphological development, physiological, and biochemical indices of the two leguminous plants, as well as soil nutrients and enzyme activities, were analyzed, followed by a comprehensive evaluation. Waste drilling fluid stress inhibited the growth of both leguminous plants. Their physiological and biochemical parameters, such as antioxidant enzyme activities and osmotic regulatory substances, exhibited a gradually increasing trend with increasing waste drilling fluid concentration. Concurrently, waste drilling fluid stress reduced soil nutrient availability and decreased soil enzyme activities. Notably, soil nutrient content increased after planting compared to the original soil without plants. Planting these two leguminous plants can effectively alleviate the negative impacts of waste drilling fluid stress, thereby indirectly contributing to soil remediation. Full article

With the advancement of digital hydropower stations, the requirements of real-time, high-precision industrial soft measurement of key power equipment operating status are attracting more and more attention. However, it is difficult to transfer energy to the monitoring sensor in strong electromagnetic environments. In this paper, a high-efficiency, high-power-density magnetic field energy harvester is proposed for monitoring sensors in hydropower stations, which captures the energy from the magnetic flux leakage of a hydroelectric generating set. Efficient magnetic energy capture is achieved by modeling material properties and optimizing the receiver’s magnetic core parameters via a Genetic Algorithm. The theoretical analysis of charging characteristics is given, and a Maximum Power Point Tracking (MPPT) control circuit is proposed, realizing high-efficiency energy conversion. Finally, an experimental planet is built. Under 70–130 Gs power-frequency magnetic fields, the system delivers 2.8–5.1 V open-circuit voltage, 66 mW maximum load power, and 6.5 mW/cm³ power density. Full article