Search Results (268)

Gas concentration in coal mine return-air corners is affected by ventilation, mining disturbance and gas drainage conditions, and it shows strong nonstationarity, local fluctuation and dynamic multi-point correlations. To improve frequency information separation, monitoring point relationship modeling, and short-term prediction accuracy, a variational mode decomposition (VMD)-based dual-branch spatiotemporal graph method is proposed. Gas concentrations from four key monitoring points are used as inputs, and the return-air corner gas concentration is taken as the output. First, the raw series are decomposed by VMD and reconstructed into low- and high-frequency components. Then, two branches are built for different frequency components. The low-frequency branch combines adaptive graph learning, graph convolution and gated recurrent units to extract global variation features, while the high-frequency branch combines graph attention and gated recurrent units to capture local disturbance features. Finally, a feature-fusion module generates multi-step predictions, and a lightweight short-term warning strategy is developed based on the predicted values. The proposed model achieves MAE, RMSE and R² values of 0.0338, 0.0471 and 0.9499 in one-step prediction, respectively, and outperforms GRU, LSTM, GCN-GRU, GAT-GRU, VMD-GRU, Informer and STGCN under three-step and six-step conditions. Cross-dataset validation and inference time analysis indicate good adaptability and online prediction potential. Full article

Within the domain of Earth observation, tracking terrestrial transitions via high-resolution optical data plays a fundamental role. Nevertheless, current methods face critical challenges, including the difficulty in collaborative modeling of local details and global features and the singularity of bi-temporal difference representation, along with insufficient cross-scale feature communication, thereby constraining both the precision and resilience of models when applied to complicated environments. To solve these problems, we propose LADENet (Locally Adaptive Mamba and Multi-scale Feature Enhancement Network), an innovative framework that synergizes CNN, Transformer, and Mamba paradigms. By leveraging customized local contextual refinement alongside sophisticated hierarchical fusion, this integration delivers highly precise and resilient detection performance. LADENet adopts a weight-sharing multi-level Transformer encoder combined with a sequence reduction mechanism to generate multi-scale global features, achieving precise alignment of bi-temporal features and global context modeling while reducing computational complexity. To realize accurate localization and local enhancement of changed regions, we design a dual spatiotemporal adaptive local feature marking module based on State-Space Scanning (SSS). This module screens high-saliency changed regions through an adaptive scanning strategy, realizes pixel-aligned spatiotemporal feature fusion via cross-temporal state-space scanning, and introduces a sliding window boundary calibration mechanism to alleviate boundary information loss caused by window segmentation. To strengthen the feature representation of changed regions, a dual-branch difference enhancement module is constructed, which collaboratively captures global change trends and fine-grained local features through an attention-enhanced difference branch and a multi-scale convolution concatenation branch, effectively suppressing background interference. To address the semantic gap between cross-scale features, a global cross-scale spatial feature fusion decoder is proposed, which balances local detail preservation and global context perception through the synergy of spatial attention and two-dimensional selective scanning, completing refined multi-scale feature fusion and spatial resolution recovery. To rigorously validate the proposed LADENet, comprehensive experiments were conducted across four widely adopted bi-temporal benchmarks: LEVIR-CD, WHU-CD, CLCD-CD, and GVLM-CD. The presented architecture establishes substantial superiority over existing cutting-edge methodologies across primary evaluation criteria. Specifically, it yields an F1-measure of 91.06% alongside an IoU of 85.28% in the LEVIR-CD tests, while registering 90.51% (F1) and 82.45% (IoU) for WHU-CD. Similarly, robust outcomes are delivered on CLCD-CD (82.15% F1, 72.83% IoU) as well as GVLM-CD (89.12% F1, 77.78% IoU). These results demonstrate that LADENet possesses excellent detection accuracy, boundary delineation capability and generalization performance in diverse and intricate bi-temporal observation environments. Full article

Respiratory diseases are a major global health challenge. However, identification of respiratory diseases is often limited by subjectivity, environmental noise and inter-clinician variability. This study presents an explainable multimodal deep learning framework for recording-level multiclass classification of respiratory audio signals. The proposed system integrates two complementary representations—a spectro-temporal encoder based on a CNN–BiLSTM-attention architecture and a handcrafted acoustic-feature encoder capturing acoustic descriptors commonly used in respiratory-audio analysis, including MFCCs, zero-crossing rate, spectral centroid, spectral bandwidth, chroma, RMS energy, and spectral rolloff features. These branches are combined through late-stage fusion to leverage both data-driven representation learning and domain-informed acoustic cues. The proposed model was trained and internally evaluated on the Asthma Detection Dataset Version 2, comprising five respiratory categories: bronchial disease, asthma, COPD, healthy, and pneumonia. Mono conversion, resampling to 16 kHz, 100–2000 Hz band-pass filtering, amplitude normalisation, fixed 4 s trimming or zero-padding, training-only augmentation, handcrafted-feature extraction, mel-spectrogram generation, quality control auditing, and stratified recording-level partitioning have been applied in the pre-processing steps. Across five repeated experiments with different random seeds, the proposed hybrid model achieved a mean held-out recording-level test accuracy of $0.9099\pm 0.0163$, balanced accuracy of $0.8936\pm 0.0152$, macro F1-score of $0.8937\pm 0.0177$, macro ROC–AUC of $0.9867\pm 0.0010$, and macro PR–AUC of $0.9489\pm 0.0044$. Conventional machine learning baseline comparisons showed that the proposed model achieved stronger internal accuracy, balanced accuracy, macro recall, macro F1-score, and macro ROC–AUC than classical machine learning algorithms trained on handcrafted acoustic features, although Random Forest remained competitive in macro PR–AUC. Ablation analysis shows that the deep spectro-temporal branch was the primary contributor to predictive performance, while the handcrafted branch provided complementary interpretable acoustic information rather than consistently improving all classification metrics. Explainability was incorporated using Grad-CAM and Integrated Gradients for spectrogram-based interpretation and SHAP for handcrafted-feature attribution. Domain-shift evaluation on the ICBHI Respiratory Sound Database and a COPD-focused cohort revealed substantial dataset shift effects, including poor healthy-case recognition on ICBHI and seed-dependent COPD recognition in the COPD-focused cohort. Identifier-aware sensitivity analyses showed lower performance than the main recording-level split, suggesting that subject-like or source-level overlap may inflate internal performance estimates. The findings should be interpreted as promising internal held-out recording-level algorithmic performance with limited external transfer, rather than evidence of readiness for clinical use. Full article

Human activity recognition has become an important component of human–computer interaction and ubiquitous computing. Among various sensing technologies, WiFi-based gesture recognition has attracted increasing attention due to its contactless nature and robustness to visual occlusion. However, environmental variations and user-specific differences often lead to significant performance degradation, particularly in cross-user scenarios. Existing methods primarily focus on learning domain-invariant representations, which may overlook user-specific characteristics that are essential for accurate recognition. To address this issue, we propose the Domain-invariant and User-Adaptive Learning Network (DUAL-Net), a dual-branch framework that jointly models domain-invariant and user-adaptive representations. Specifically, DUAL-Net incorporates a contrastive fusion learning (CFL) module with modality-specific encoders to learn complementary representations from WiFi and vision modalities. Furthermore, a spatial matrix difference (SMD)-guided cross-modal generation (CMG) module is introduced to generate user-adaptive WiFi features by incorporating structural priors derived from skeletal representations. To improve deployment efficiency, DUAL-Net adopts a two-stage learning framework, where adaptation is conducted offline to reduce online computational overhead. Experiments on the MM-Fi dataset and a self-collected dataset show that DUAL-Net achieves superior cross-user recognition performance compared with existing single-modality and multimodal methods. In addition, SMD-guided conditioning improves recognition accuracy by up to 8.79% over diffusion generation without structural guidance. Full article

The terrain of tea gardens in Yunnan is complex, and the branches and leaves of tea trees are densely interlaced. Traditional manual picking methods are labor-intensive and inefficient, and can no longer meet the needs of modern tea industry development. To achieve automatic recognition of tea buds and leaves and accurate two-dimensional localization of picking points in complex natural environments, this study constructs a lightweight and high-precision tea bud and leaf detection and keypoint localization model, YOLOv11-pose-BEH, based on the YOLOv11-pose model. Based on the original network, this model introduces the BiFPN feature fusion module to achieve efficient bidirectional transmission of multi-scale information. The EMA (Efficient Multi-scale Attention) mechanism is integrated to form the C2PSA_EMA module, improving the effect of multi-scale feature extraction. HetConv is introduced to form the C3K2_HetConv module, enhancing the extraction of local texture and edge features. Compared with the baseline network, the improved YOLOv11-pose-BEH achieved significant improvements in both tea bud object detection and picking point localization. For the tea bud object detection task, the precision, recall, mAP0.5, and F1-score increased by 7.41, 6.39, 5.28, and 6.88 percentage points, respectively. For the picking point localization task, the precision, recall, mAP0.5, and F1-score increased by 5.86, 5.83, 4.83, and 5.85 percentage points, respectively. These results indicate that the proposed model can achieve more accurate and stable tea bud and leaf detection and two-dimensional picking point localization under complex backgrounds, providing efficient and reliable technical support for the visual perception of intelligent tea-picking robots in tea gardens. Full article

To address cross-subject domain shift and insufficient complementary fusion of time-frequency information in EEG-based emotion recognition, this paper proposes a multi-task adversarial network: Path Adversarial Dual-Branch Network for EEG Emotion Recognition (PADB-Net). The model adopts a dual-branch parallel architecture for time and frequency domains, processing raw EEG waveforms and differential entropy features respectively, and extracts discriminative features using lightweight depthwise separable convolutions and channel attention. A path adversarial module is introduced for the first time in emotion recognition to align time-domain and frequency-domain feature distributions, solving the single-branch dominance problem in dual-branch fusion. Together with a domain adversarial module, the overall distributions of source and target domains as well as the internal distributions of the two modality branches are aligned within a unified framework. Experiments on a dataset containing healthy subjects and patients with major depressive disorder show that the full model significantly outperforms single-adversarial and non-adversarial baselines in accuracy, AUC, F1-score, sensitivity, and specificity, verifying the synergistic gain of the dual-adversarial mechanism. On the HybridBCI dataset, PADB-Net achieves 77.80% accuracy, 84.50% AUC, and 79.40% F1-score with only 6.45 K trainable parameters. When transferred to the public SEED dataset for three-class emotion recognition, the model attains F1-scores of 71.83% (negative), 68.99% (neutral), and 73.37% (positive), demonstrating strong cross-dataset generalizability. Full article

Non-intrusive load monitoring (NILM) is essential for smart grid demand-side management and energy conservation, yet existing methods suffer from limited feature discrimination, ambiguous identification of similar electrical appliances, and difficulty balancing model accuracy and lightweight deployment. To address these issues, this paper proposes a dual-branch lightweight load identification method fusing steady-state features and lightweight network. Firstly, V-I trajectory images are generated via standardized transformation and two-dimensional histogram logarithmic mapping, while steady-state characteristics, including active power, reactive power, trajectory area and intermediate section slope, are extracted. Then, a dual-branch network is constructed, where the visual branch adopts depthwise separable convolution and lightweight multi-head attention to mine global trajectory features, and the numerical branch uses fully connected layers to encode steady-state features; feature concatenation fusion is adopted to complete appliance classification. The experimental results on the Plug Load Appliance Identification Dataset (PLAID dataset) show that the proposed method achieves a recognition accuracy of 95.35% with only 0.17M parameters, outperforming standard and medium convolutional neural network (CNN) models. Ablation experiments verify that steady-state feature fusion effectively improves the identification accuracy of easily confused and small-sample loads. The proposed method realizes high-precision and lightweight load identification, which is suitable for edge deployment in smart meters and has practical application value for intelligent power management. Full article

Accurate remaining useful life (RUL) prediction for aero-engines is important for condition-based maintenance and safety-oriented health management. Long short-term memory (LSTM) networks are widely used for this task, but two limitations remain important in multi-sensor degradation modeling: hidden states generated over the full window are often under-utilized, and attention mechanisms may overemphasize locally fluctuating sensor readings. This paper proposes a Multi-Statistic Disentangled LSTM (MSD-LSTM) framework for aero-engine RUL prediction. The framework first applies Savitzky–Golay filtering to smooth high-frequency signal fluctuations. A hidden-state feature extraction module then combines feature-level disentangled extraction and Global Average Pooling to use the LSTM hidden-state sequence beyond the final recurrent output. In parallel, a Multi-Statistic Pooler summarizes each input window using minimum, maximum, standard deviation, and mean statistics, and its output is fused with a self-attention branch through a static-gating mechanism. On the NASA C-MAPSS benchmark, MSD-LSTM achieves RMSE values of 10.45 and 12.33 on FD001 and FD002, respectively, and ranks first in RMSE on three of the four sub-datasets and first in SCORE on two sub-datasets among the compared recent methods. Ablation and fusion analyses show that both the hidden-state extraction and statistic-guided fusion components contribute to stable RUL prediction. Full article

This paper proposes a novel forecasting model termed U-Client, which integrates parallel cross-temporal and cross-variable attention branches with an adaptive gated fusion mechanism for ultra-short-term photovoltaic (PV) power forecasting. First, meteorological features are screened using the Pearson correlation coefficient to reduce input dimensionality. Second, parallel cross-temporal and cross-variable attention branches are designed to extract long-range temporal trends and nonlinear interaction features among meteorological variables, respectively. Third, a gating mechanism is introduced to adaptively fuse the two types of features based on input conditions. Finally, a linear module is combined to generate the final forecasting results. Experiments based on measured datasets from a photovoltaic station in Ningxia, China, demonstrate that the proposed U-Client model outperforms classical models such as Long Short-Term Memory (LSTM) and Informer across all evaluation metrics for 1–4 step forecasting tasks. Ablation studies and statistical significance tests further verify the effectiveness of each component. The proposed model provides reliable support for ultra-short-term power dispatching in new-type power systems. Full article

Remote sensing object detection remains challenging because objects often appear with large scale variation, dense spatial layouts, and strong interference from complex geographical backgrounds. To address these coupled difficulties, we propose EDM-Net, an end-to-end multi-scale detector that organizes feature processing into three coordinated stages: adaptive extraction, intra-scale interaction, and cross-scale fusion. First, an efficient sparse mixture-of-experts (ES-MoE) module is embedded in the backbone to allocate scale-specific convolutional experts according to scene-level feature responses, providing a more adaptive feature basis than a single static extraction path. Second, a dynamic mixing intra-scale feature interaction (DMIFI) module is introduced into the Transformer encoder. This module combines global self-attention with dynamic spatial mixing, thereby preserving long-range context while reintroducing local two-dimensional inductive bias for dense and small objects. Third, a multi-scale synergistic attention fusion (MSAF) module aligns adjacent feature levels through parallel local and global attention branches and structural re-parameterization, reducing semantic dilution during feature aggregation. Comprehensive experiments on three large-scale remote sensing benchmark datasets, DIOR, NWPU VHR-10, and RSOD, demonstrate that EDM-Net consistently improves over the re-trained RT-DETR-R18 baseline under the same experimental protocol, attaining mAP50 scores of 83.7%, 95.6%, and 95.8% respectively. Additional ablation and scale-specific analyses indicate that the three modules contribute complementary gains, especially for small and densely distributed objects. These results suggest that coordinated extraction, interaction, and fusion can improve remote sensing object detection under complex scale and background conditions. Full article

As autonomous driving technology expands from structured urban roads to unstructured outdoor environments, precise understanding of complex terrain has become a critical requirement for ensuring safe vehicle navigation. However, outdoor environments are characterized by high dynamics, drastic illumination variations, ambiguous category boundaries, and prohibitive annotation costs, making traditional supervised learning methods that rely on large amounts of pixel-level annotations difficult to generalize. In this paper, we propose a novel dual-path bidirectional interactive encoder, termed LapDINO, that effectively combines the strong semantic generalization capability of the self-supervised foundation model DINOv3 with the multi-scale frequency analysis capacity of the Laplacian pyramid. Specifically, we leverage DINOv3 to extract global semantic features as a “semantic map”, while simultaneously obtaining multi-scale high-frequency details through Laplacian pyramid decomposition as “structural contours”. Building upon this, we design a bidirectional cross-attention fusion mechanism that enables dynamic interaction and mutual refinement between semantic information and geometric details. Furthermore, we introduce a multi-branch attention enhancement module that extracts pyramid features from three complementary perspectives. To address domain shift, we design lightweight visual adapters that enable efficient fine-tuning of the frozen DINOv3 backbone. Finally, we construct two off-road terrain segmentation datasets, VOTD and VOCD, to facilitate research in this domain. Experimental results demonstrate that the proposed method achieves state-of-the-art performance, striking an optimal balance between accuracy and computational efficiency, thereby providing a robust and efficient engineering solution for terrain perception in off-road environments. Full article

Landslides cause thousands of fatalities and billions in economic losses annually, yet reliable susceptibility mapping across heterogeneous landscapes remains challenging because conventional models assume stationary relationships between landslide occurrence and environmental controls. Ensemble methods, though promising, rely on either globally fixed aggregation weights or kernel-constrained local averaging, failing to adapt when the reliability of base models varies nonlinearly across space. To overcome this, we propose a two-stage hierarchical spatial adaptive ensemble (HSE) framework. In stage one, three complementary base learners are deployed: geographically weighted regression (GWR) for local spatial non-stationarity; a geographically optimal similarity (GOS) model, grounded in the Third Law of Geography, to represent similarity-based local dependence; and a deep neural network (DNN) for nonlinear covariate interactions. In stage two, a multi-branch attention-based network learns spatially varying fusion weights via multi-scale feature extraction, abandoning fixed weights or kernel constraints. We validate HSE on a typical landslide-prone catchment, comparing against single models (GWR, DNN, GOS). Results demonstrate that our method consistently achieves superior predictive accuracy, spatial consistency, and out-of-sample robustness. Moreover, the attention-derived spatially varying weights provide interpretable insights into where each base learner dominates, bridging predictive performance with geophysical interpretability. These findings confirm that explicitly learning spatial heterogeneity during ensemble fusion is essential for reliable landslide susceptibility mapping, with strong potential for transfer to other geospatial prediction tasks. Full article

Red tide events pose substantial threats to marine ecosystems, aquaculture, and coastal public health. Timely and accurate delineation of red tide extent from satellite imagery is therefore essential for operational monitoring and early warning. However, existing deep learning-based semantic segmentation methods generally treat multispectral bands as homogeneous inputs and do not fully exploit the domain knowledge embodied in spectral indices commonly used in traditional remote sensing analysis. To address this limitation, this study proposes a spectral-index-guided network (SIG-Net) that explicitly incorporates spectral-index priors into deep feature extraction through a dual-branch architecture. SIG-Net comprises three components: a spectral encoder based on a Mix Vision Transformer (MiT-B2) that learns spatial-spectral representations from the original Sentinel-2 bands; a lightweight CNN-based index encoder that extracts discriminative features from four spectral indices, namely the red-green index (RGI), blue-green index (BGI), normalized difference vegetation index (NDVI), and the normalized difference Noctiluca index (NDNI) proposed in this study; and a spectral-index-guided fusion (SIGF) module that adaptively integrates multi-scale features from the two branches using spatial-reduction cross-attention and a gated fusion mechanism. Experiments on a Sentinel-2 red tide dataset show that SIG-Net outperforms single-branch baselines, including U-Net, DeepLabV3+, and SegFormer, as well as naive multi-source fusion strategies. Ablation studies further confirm the contributions of the SIGF module, the gating mechanism, and the proposed NDNI to performance improvements. The proposed method provides an effective framework for integrating domain knowledge with deep learning for red tide remote sensing monitoring. Full article

To improve the accuracy and stability of short-term traffic flow prediction in complex road networks and address the limitations of existing models in modeling spatiotemporal dependencies, this paper proposes a traffic flow prediction model based on a Hypergraph Spatio-Temporal Interaction Network (HGSTIN) in the context of intelligent transportation systems. The study constructs a multi-dimensional traffic pattern input tensor by integrating three temporal scales—proximity, intra-day, and intra-week—while taking traffic flow as the prediction target and introducing average speed and lane occupancy as auxiliary features. In terms of temporal modeling, a Transformer architecture integrated with a Dynamic Tanh (DyT) mechanism is adopted to capture multi-period variations. For spatial modeling, a neighborhood hypergraph and a DTW-based semantic hypergraph are combined to enhance the representation of local and global through spatial self-attention and hypergraph neural network branches, and an adaptive feature fusion module is designed to perform adaptively weighted fusion of the outputs from the two branches. In terms of loss function design, a temporal gradient consistency loss function is proposed to enhance the robustness of predictions. Experimental results on the PEMS04 and PEMS08 datasets show that the proposed model achieves average improvements of approximately 5.15%, 1.76%, and 3.88% in MAE, RMSE, and MAPE, respectively, compared to the second-best baseline model. The model exhibits the smallest performance degradation in multi-step prediction scenarios, and the effectiveness of each module is validated through ablation studies. The findings demonstrate that HGSTIN can effectively capture the dynamic spatiotemporal characteristics of complex traffic scenarios, thereby providing high-precision prediction support for intelligent transportation systems. Full article

The operational health state of motor bearings is critical to the operational safety of harbor portal slewing cranes. However, in harsh industrial environments with strong noise and time-varying rotational speeds, existing bearing fault diagnosis methods still suffer from the problems of incomplete fault feature extraction from single-sensor signals and the excessively large size of multi-source fusion models, which makes them unable to adapt to edge deployment. To address these issues, this paper proposes a Multi-source Feature Fusion Lightweight Network (MTFL-Net) integrated with targeted structured channel pruning. First, vibration and current signals are preprocessed via differentiated time-frequency transformation and converted into 2D time-frequency images, to fully preserve transient impact and spectral fault features. Second, a multi-branch feature extraction architecture embedded with residual connections, multi-scale convolution and channel attention gating is designed, to alleviate feature degradation and adaptively enhance fault-sensitive features. Third, targeted structured channel pruning is performed on the feature extraction branches, to remove redundant channels while retaining the multi-source fusion logic and core feature extraction structure. Experiments on two public bearing datasets show that the original model achieves 99% diagnostic accuracy, and the pruned model still maintains an accuracy of 95%. The results demonstrate that MTFL-Net can significantly reduce model size and computational cost while retaining high diagnostic precision. Full article