Search Results (752)

This paper proposes a rolling bearing fault diagnosis method based on HFMD and a dual-branch parallel network, aiming to address the issue of diagnostic accuracy being compromised by the disparity in data quality across different source domains due to sparse feature separation in rolling bearing acoustic signals. Traditional methods face challenges in feature extraction, sensitivity to noise, and difficulties in handling coupled multi-fault conditions in rolling bearing fault diagnosis. To overcome these challenges, this study first employs the HawkFish Optimization Algorithm to optimize Feature Mode Decomposition (HFMD) parameters, thereby improving modal decomposition accuracy. The optimal modal components are selected based on the minimum Residual Energy Index (REI) criterion, with their time-domain graphs and Continuous Wavelet Transform (CWT) time-frequency diagrams extracted as network inputs. Then, a dual-branch parallel network model is constructed, where the multi-scale residual structure (Res2Net) incorporating the Efficient Channel Attention (ECA) mechanism serves as the temporal branch to extract key features and suppress noise interference, while the Swin Transformer integrating multi-stage cross-scale attention (MSCSA) acts as the time-frequency branch to break through local perception bottlenecks and enhance classification performance under limited resources. Finally, the time-domain graphs and time-frequency graphs are, respectively, input into Res2Net and Swin Transformer, and the features from both branches are fused through a fully connected layer to obtain comprehensive fault diagnosis results. The research results demonstrate that the proposed method achieves 100% accuracy in open-source datasets. In the experimental data, the diagnostic accuracy of this study demonstrates significant advantages over other diagnostic models, achieving an accuracy rate of 98.5%. Under few-shot conditions, this study maintains an accuracy rate no lower than 95%, with only a 2.34% variation in accuracy. HFMD and the dual-branch parallel network exhibit remarkable stability and superiority in the field of rolling bearing fault diagnosis. Full article

To address the issues of inaccurate positioning, weak feature extraction capability, and poor cross-domain adaptability in the detection of surface defects of steel materials, this paper proposes an improved YOLOv11-EMD algorithm and integrates a multi-stage transfer learning framework to achieve high-precision, robust, and low-cost industrial defect detection. Specifically, the InnerEIoU loss function is introduced to improve the accuracy of bounding box regression, the multi-scale dilated attention (MSDA) module is integrated to enhance the multi-scale feature fusion capability, and the Cross-Stage Partial Network with 3 Convolutions and Kernel size 2 Dynamic Convolution (C3k2_DynamicConv) module is embedded to improve the expression of and adaptability to complex defects. To address the problem of performance degradation when the model migrates between different data domains, a multi-stage transfer learning framework is constructed, combining source domain pre-training and target domain fine-tuning strategies to improve the model’s generalization ability in scenarios with changing data distributions. On the comprehensive dataset constructed of NEU-DET and Severstal steel defect images, YOLOv11-EMD achieved a precision of 0.942, a recall of 0.868, and an mAP@50 of 0.949, which are 3.5%, 0.8%, and 1.6% higher than the original model, respectively. On the cross-scenario mixed dataset composed of NEU-DET and GC10-DET data, the mAP@50 was 0.799, outperforming mainstream detection algorithms. The multi-stage transfer strategy can shorten the training time by 3.2% and increase the mAP by 8.8% while maintaining accuracy. The proposed method improves the defect detection accuracy, has good generalization and engineering application potential, and is suitable for automated quality inspection tasks in diverse industrial scenarios. Full article

The primary objective of the presented study is to develop a comprehensive system for the acquisition of surface electromyographic (sEMG) data and to perform time–frequency analysis aimed at extracting discriminative features for the classification of hand gestures intended for the control of a simplified bionic hand prosthesis. The proposed system is designed to facilitate precise finger gesture execution in both prosthetic and robotic hand applications. This article outlines the methodology for multi-channel sEMG signal acquisition and processing, as well as the extraction of relevant features for gesture recognition using artificial neural networks (ANNs) and other well-established machine learning (ML) algorithms. Electromyographic signals were acquired using a prototypical LPCXpresso LPC1347 ARM Cortex M3 (NXP, Eindhoven, Holland) development board in conjunction with surface EMG sensors of the Gravity OYMotion SEN0240 type (DFRobot, Shanghai, China). Signal processing and feature extraction were carried out in the MATLAB 2024b environment, utilizing both the Fourier transform and the Hilbert–Huang transform to extract selected time–frequency characteristics of the sEMG signals. An artificial neural network (ANN) was implemented and trained within the same computational framework. The experimental protocol involved 109 healthy volunteers, each performing five predefined gestures of the right hand. The first electrode was positioned on the brachioradialis (BR) muscle, with subsequent channels arranged laterally outward from the perspective of the participant. Comprehensive analyses were conducted in the time domain, frequency domain, and time–frequency domain to evaluate signal properties and identify features relevant to gesture classification. The bionic hand prototype was fabricated using 3D printing technology with a PETG filament (Spectrum, Pęcice, Poland). Actuation of the fingers was achieved using six MG996R servo motors (TowerPro, Shenzhen, China), each with an angular range of ${180}^{\circ }$, controlled via a PCA9685 driver board (Adafruit, New York, NY, USA) connected to the main control unit. Full article

Background/Objectives: Epilepsy is a neurological disorder that severely impacts patients’ quality of life. In clinical practice, specific pharmacological and surgical interventions are tailored to distinct seizure types. The identification of the epileptogenic zone enables the implementation of surgical procedures and neuromodulation therapies. Consequently, accurate classification of seizure types and precise determination of focal epileptic signals are critical to provide clinicians with essential diagnostic insights for optimizing therapeutic strategies. Traditional machine learning approaches are constrained in their efficacy due to limited capability in autonomously extracting features. Methods: This study proposes a novel deep learning framework integrating temporal and spatial information extraction to address this limitation. Multivariate variational mode decomposition (MVMD) is employed to maintain inter-channel mode alignment during the decomposition of multi-channel epileptic signals, ensuring the synchronization of time–frequency characteristics across channels and effectively mitigating mode mixing and mode mismatch issues. Results: The Bern–Barcelona database is employed to classify focal epileptic signals, with the proposed framework achieving an accuracy of 98.85%, a sensitivity of 98.75%, and a specificity of 98.95%. For multi-class seizure type classification, the TUSZ database is utilized. Subject-dependent experiments yield an accuracy of 96.17% with a weighted F1-score of 0.962. Meanwhile, subject-independent experiments attain an accuracy of 87.97% and a weighted F1-score of 0.884. Conclusions: The proposed framework effectively integrates temporal and spatial domain information derived from multi-channel epileptic signals, thereby significantly enhancing the algorithm’s classification performance. The performance on unseen patients demonstrates robust generalization capability, indicating the potential clinical applicability in assisting neurologists with epileptic signal classification. Full article

Traffic flow prediction is a critical component of intelligent transportation systems, playing a vital role in alleviating congestion, improving road resource utilization, and supporting traffic management decisions. Although deep learning methods have made remarkable progress in this field in recent years, current studies still face challenges in modeling complex spatio-temporal dependencies, adapting to anomalous events, and generalizing to large-scale real-world scenarios. To address these issues, this paper proposes a novel traffic flow prediction model. The proposed approach simultaneously leverages temporal and frequency domain information and introduces adaptive graph convolutional layers to replace traditional graph convolutions, enabling dynamic capture of traffic network structural features. Furthermore, we design a frequency–temporal multi-head attention mechanism for effective multi-scale spatio-temporal feature extraction and develop a cross-multi-scale graph fusion strategy to enhance predictive performance. Extensive experiments on real-world datasets, PeMS and Beijing, demonstrate that our method significantly outperforms state-of-the-art (SOTA) baselines. For example, on the PeMS20 dataset, our model achieves a 53.6% lower MAE, a 12.3% lower NRMSE, and a 3.2% lower MAPE than the best existing method (STFGNN). Moreover, the proposed model achieves competitive computational efficiency and inference speed, making it well-suited for practical deployment. Full article

Pipelines play a vital role in material transportation within industrial settings. This review synthesizes detection technologies for early-stage small gas leaks from pipelines in the industrial sector, with a focus on acoustic-based methods, optical gas imaging (OGI), and multimodal fusion approaches. It encompasses detection principles, inherent challenges, mitigation strategies, and the state of the art (SOTA). Small leaks refer to low flow leakage originating from defects with apertures at millimeter or submillimeter scales, posing significant detection difficulties. Acoustic detection leverages the acoustic wave signals generated by gas leaks for non-contact monitoring, offering advantages such as rapid response and broad coverage. However, its susceptibility to environmental noise interference often triggers false alarms. This limitation can be mitigated through time-frequency analysis, multi-sensor fusion, and deep-learning algorithms—effectively enhancing leak signals, suppressing background noise, and thereby improving the system’s detection robustness and accuracy. OGI utilizes infrared imaging technology to visualize leakage gas and is applicable to the detection of various polar gases. Its primary limitations include low image resolution, low contrast, and interference from complex backgrounds. Mitigation techniques involve background subtraction, optical flow estimation, fully convolutional neural networks (FCNNs), and vision transformers (ViTs), which enhance image contrast and extract multi-scale features to boost detection precision. Multimodal fusion technology integrates data from diverse sensors, such as acoustic and optical devices. Key challenges lie in achieving spatiotemporal synchronization across multiple sensors and effectively fusing heterogeneous data streams. Current methodologies primarily utilize decision-level fusion and feature-level fusion techniques. Decision-level fusion offers high flexibility and ease of implementation but lacks inter-feature interaction; it is less effective than feature-level fusion when correlations exist between heterogeneous features. Feature-level fusion amalgamates data from different modalities during the feature extraction phase, generating a unified cross-modal representation that effectively resolves inter-modal heterogeneity. In conclusion, we posit that multimodal fusion holds significant potential for further enhancing detection accuracy beyond the capabilities of existing single-modality technologies and is poised to become a major focus of future research in this domain. Full article

Conventional rail transit train running gear fault diagnosis mainly depends on routine maintenance inspections and manual judgment. However, these approaches lack robustness under complex operational environments and elevated noise levels, rendering them inadequate for real-time performance and the rigorous accuracy standards demanded by modern rail transit systems. Furthermore, many existing deep learning–based methods suffer from inherent limitations in feature extraction or incur prohibitive computational costs when processing multivariate time series data. This study represents one of the early efforts to introduce the TimesNet time series modeling framework into the domain of fault diagnosis for rail transit train running gear. By utilizing an innovative multi-period decomposition strategy and a mechanism for reshaping one-dimensional data into two-dimensional tensors, the framework enables advanced temporal-spatial representation of time series data. Algorithm validation is performed on both the high-speed train running gear bearing fault dataset and the multi-mode fault diagnosis datasets of gearbox under variable working conditions. The TimesNet model exhibits outstanding diagnostic performance on both datasets, achieving a diagnostic accuracy of 91.7% on the high-speed train bearing fault dataset. Embedded deployment experiments demonstrate that single-sample inference is completed within 70.3 ± 5.8 ms, thereby satisfying the real-time monitoring requirement (<100 ms) with a 100% success rate over 50 consecutive tests. The two-dimensional reshaping approach inherent to TimesNet markedly enhances the capacity of the model to capture intrinsic periodic structures within multivariate time series data, presenting a novel paradigm for the intelligent fault diagnosis of complex mechanical systems in train running gears. The integrated human–machine interaction system includes a comprehensive closed-loop process encompassing detection, diagnosis, and decision-making, thereby laying a robust foundation for the continued development of train running gear predictive maintenance technologies. Full article

With the growing demand for natural and intelligent human–machine interaction in multilingual environments, automatic language identification (LID) has emerged as a crucial component in voice-enabled systems, particularly in the automotive domain. This study proposes an audio-based LID model that identifies the spoken language directly from voice input without requiring manual language selection. The model architecture leverages two types of feature extraction pipelines: a Variational Autoencoder (VAE) and a pre-trained Wav2Vec model, both used to obtain latent speech representations. These embeddings are then fed into a multi-layer perceptron (MLP)-based classifier to determine the speaker’s language among five target languages: Korean, Japanese, Chinese, Spanish, and French. The model is trained and evaluated using a dataset preprocessed into Mel-Frequency Cepstral Coefficients (MFCCs) and raw waveform inputs. Experimental results demonstrate the effectiveness of the proposed approach in achieving accurate and real-time language detection, with potential applications in in-vehicle systems, speech translation platforms, and multilingual voice assistants. By eliminating the need for predefined language settings, this work contributes to more seamless and user-friendly multilingual voice interaction systems. Full article

Road extraction from remote sensing imagery plays a critical role in applications such as autonomous driving, urban planning, and infrastructure development. Although deep learning methods have achieved notable progress, current approaches still struggle with complex backgrounds, varying road widths, and strong texture interference, often leading to fragmented road predictions or the misclassification of background regions. Given that roads typically exhibit smooth low-frequency characteristics while background clutter tends to manifest in mid- and high-frequency ranges, incorporating frequency-domain information can enhance the model’s structural perception and discrimination capabilities. To address these challenges, we propose a novel frequency-aware road extraction network, termed PWFNet, which combines frequency-domain modeling with multi-scale feature enhancement. PWFNet comprises two key modules. First, the Pyramidal Wavelet Convolution (PWC) module employs multi-scale wavelet decomposition fused with localized convolution to accurately capture road structures across various spatial resolutions. Second, the Frequency-aware Adjustment Module (FAM) partitions the Fourier spectrum into multiple frequency bands and incorporates a spatial attention mechanism to strengthen low-frequency road responses while suppressing mid- and high-frequency background noise. By integrating complementary modeling from both spatial and frequency domains, PWFNet significantly improves road continuity, edge clarity, and robustness under complex conditions. Experiments on the DeepGlobe and CHN6-CUG road datasets demonstrate that PWFNet achieves IoU improvements of 3.8% and 1.25% over the best-performing baseline methods, respectively. In addition, we conducted cross-region transfer experiments by directly applying the trained model to remote sensing images from different geographic regions and at varying resolutions to assess its generalization capability. The results demonstrate that PWFNet maintains the continuity of main and branch roads and preserves edge details in these transfer scenarios, effectively reducing false positives and missed detections. This further validates its practicality and robustness in diverse real-world environments. Full article

This study presents a gesture-based authentication system utilizing triboelectric nanogenerator (TENG) sensors. As self-powered devices capable of generating high-voltage outputs without external power, TENG sensors are well-suited for low-power IoT sensors and smart device applications. The proposed system recognizes single tap, double tap, and holding gestures. The electrical characteristics of the sensor were evaluated under varying pressure conditions, confirming a linear relationship between applied force and output voltage. These results demonstrate the sensor’s high sensitivity and precision. A threshold-based classification algorithm was developed by analyzing signal features enabling accurate gesture recognition in real time. To enhance the practicality and scalability of the system, the algorithm was further configured to automatically segment raw sensor signals into gesture intervals and assign corresponding labels. From these segments, time-domain statistical features were extracted to construct a training dataset. A random forest classifier trained on this dataset achieved a high classification accuracy of 98.15% using five-fold cross-validation. The system reduces security risks commonly associated with traditional keypad input, offering a user-friendly and reliable authentication interface. This work confirms the feasibility of TENG-based gesture recognition for smart locks, IoT authentication devices, and wearable electronics, with future improvements expected through AI-based signal processing and multi-sensor integration. Full article

Single remote sensing image dehazing aims to eliminate atmospheric scattering effects without auxiliary information. It serves as a crucial preprocessing step for enhancing the performance of downstream tasks in remote sensing images. Conventional approaches often struggle to balance haze removal and detail restoration under non-uniform haze distributions. To address this issue, we propose a Dual-domain Feature Fusion Network (DFFNet) for remote sensing image dehazing. DFFNet consists of two specialized units: the Frequency Restore Unit (FRU) and the Context Extract Unit (CEU). As haze primarily manifests as low-frequency energy in the frequency domain, the FRU effectively suppresses haze across the entire image by adaptively modulating low-frequency amplitudes. Meanwhile, to reconstruct details attenuated due to dense haze occlusion, we introduce the CEU. This unit extracts multi-scale spatial features to capture contextual information, providing structural guidance for detail reconstruction. Furthermore, we introduce the Dual-Domain Feature Fusion Module (DDFFM) to establish dependencies between features from FRU and CEU via a designed attention mechanism. This leverages spatial contextual information to guide detail reconstruction during frequency domain haze removal. Experiments on the StateHaze1k, RICE and RRSHID datasets demonstrate that DFFNet achieves competitive performance in both visual quality and quantitative metrics. Full article

Background: Efficient decoding of motor imagery (MI) electroencephalogram (EEG) signals is essential for the precise control and practical deployment of brain-computer interface (BCI) systems. Owing to the complex nonlinear characteristics of EEG signals across spatial, spectral, and temporal dimensions, efficiently extracting multidimensional discriminative features remains a key challenge to improving MI-EEG decoding performance. Methods: To address the challenge of capturing complex spatial, spectral, and temporal features in MI-EEG signals, this study proposes a multi-branch deep neural network, which jointly models these dimensions to enhance classification performance. The network takes as inputs both a three-dimensional power spectral density tensor and two-dimensional time-domain EEG signals and incorporates four complementary feature extraction branches to capture spatial, spectral, spatial-spectral joint, and temporal dynamic features, thereby enabling unified multidimensional modeling. The model was comprehensively evaluated on two widely used public MI-EEG datasets: EEG Motor Movement/Imagery Database (EEGMMIDB) and BCI Competition IV Dataset 2a (BCIIV2A). To further assess interpretability, gradient-weighted class activation mapping (Grad-CAM) was employed to visualize the spatial and spectral features prioritized by the model. Results: On the EEGMMIDB dataset, it achieved an average classification accuracy of 86.34% and a kappa coefficient of 0.829 in the five-class task. On the BCIIV2A dataset, it reached an accuracy of 83.43% and a kappa coefficient of 0.779 in the four-class task. Conclusions: These results demonstrate that the network outperforms existing state-of-the-art methods in classification performance. Furthermore, Grad-CAM visualizations identified the key spatial channels and frequency bands attended to by the model, supporting its neurophysiological interpretability. Full article

It is particularly important to ensure the safety of engineering structures, such as aerospace vehicles and wind turbines, most of which are made of composite materials. A sudden failure of the structure may happen following the accumulation of structural damage. Since they are sensitive to tiny damage and can propagate through engineering structures over a long distance, Lamb waves have been widely explored to develop highly efficient damage detection theories and methodologies. During propagation, affected by the mechanical properties of the structure, a large amount of information and features related to structural states can be reflected and transmitted by Lamb waves, including the occurrence and extent of structural damage. By analyzing the effect of damage acting on Lamb waves, a multi-scale wavelet transform analysis is adopted to extract multi-feature parameters in the time–frequency domain of the acquired signals. With the help of the nonlinear mapping ability of a neural network, a damage assessment model for composite structures is constructed to realize the evaluation of typical structural damage at different levels. The results of an experiment conducted on an epoxy–glass-fiber-reinforced plate show that the extracted multi-feature parameters of Lamb waves in the time–frequency domain are sensitive to the accumulated typical damage. The damage assessment model can properly evaluate the damage degree with satisfactory accuracy. Full article

Palmprint recognition is becoming more and more common in the fields of security authentication, mobile payment, and crime detection. Aiming at the problem of small sample size and low recognition rate of palmprint, a small-sample palmprint recognition method based on image expansion and Dynamic Model-Agnostic Meta-Learning (DMAML) is proposed. In terms of data augmentation, a multi-connected conditional generative network is designed for generating palmprints; the network is trained using a gradient-penalized hybrid loss function and a dual time-scale update rule to help the model converge stably, and the trained network is used to generate an expanded dataset of palmprints. On this basis, the palmprint feature extraction network is designed considering the frequency domain and residual inspiration to extract the palmprint feature information. The DMAML training method of the network is investigated, which establishes a multistep loss list for query ensemble loss in the inner loop. It dynamically adjusts the learning rate of the outer loop by using a combination of gradient preheating and a cosine annealing strategy in the outer loop. The experimental results show that the palmprint dataset expansion method in this paper can effectively improve the training efficiency of the palmprint recognition model, evaluated on the Tongji dataset in an N-way K-shot setting, our proposed method achieves an accuracy of 94.62% ± 0.06% in the 5-way 1-shot task and 87.52% ± 0.29% in the 10-way 1-shot task, significantly outperforming ProtoNets (90.57% ± 0.65% and 81.15% ± 0.50%, respectively). Under the 5-way 1-shot condition, there was a 4.05% improvement, and under the 10-way 1-shot condition, there was a 6.37% improvement, demonstrating the effectiveness of our method. Full article

Existing methods for diagnosing depression rely heavily on subjective evaluations, whereas electroencephalography (EEG) emerges as a promising approach for objective diagnosis due to its non-invasiveness, low cost, and high temporal resolution. However, current EEG analysis methods are constrained by volume conduction effect and class imbalance, both of which adversely affect classification performance. To address these issues, this paper proposes a multi-stage deep learning model for EEG-based depression classification, integrating a cortical feature extraction strategy (CFE), a feature attention module (FA), a graph convolutional network (GCN), and a focal adversarial domain adaptation module (FADA). Specifically, the CFE strategy reconstructs brain cortical signals using the standardized low-resolution brain electromagnetic tomography (sLORETA) algorithm and extracts both linear and nonlinear features that capture cortical activity variations. The FA module enhances feature representation through a multi-head self-attention mechanism, effectively capturing spatiotemporal relationships across distinct brain regions. Subsequently, the GCN further extracts spatiotemporal EEG features by modeling functional connectivity between brain regions. The FADA module employs Focal Loss and Gradient Reversal Layer (GRL) mechanisms to suppress domain-specific information, alleviate class imbalance, and enhance intra-class sample aggregation. Experimental validation on the publicly available PRED+CT dataset demonstrates that the proposed model achieves a classification accuracy of 85.33%, outperforming current state-of-the-art methods by 2.16%. These results suggest that the proposed model holds strong potential for improving the accuracy and reliability of EEG-based depression classification. Full article