Search Results (841)

Remote sensing acquires Earth surface information without physical contact through sensors operating at diverse spatial, spectral, and temporal resolutions. In high-resolution remote sensing imagery, objects often exhibit large scale variation, complex spatial distributions, and strong inter-class similarity, posing persistent challenges for accurate semantic segmentation. Existing methods still struggle to simultaneously preserve fine boundary details and model long-range spatial dependencies, and lack explicit mechanisms to decouple low-frequency semantic context from high-frequency structural information. To address these limitations, we propose the Multi-Perspective Information Fusion Network (MPIFNet) for remote sensing semantic segmentation, motivated by the need to integrate global context, local structures, and multi-frequency information into a unified framework. MPIFNet employs a Global and Local Mamba Block Self-Attention (GLMBSA) module to capture long-range dependencies while preserving local details, and a Double-Branch Haar Wavelet Transform (DBHWT) module to separate and enhance low- and high-frequency features. By fusing spatial, hierarchical, and frequency representations, MPIFNet learns more discriminative and robust features. Evaluations on the Vaihingen, Potsdam, and LoveDA datasets through ablation and comparative studies highlight the strong generalization of our model, yielding mIoU results of 86.03%, 88.36%, and 55.76%. Full article

Nucleotide-binding site-leucine-rich repeat (NLR) proteins are key intracellular immune receptors in plants. Integrated domains (IDs) can occasionally be fused with NLRs, contributing to their functional diversity. However, the diversity and evolutionary patterns of NLR-IDs across angiosperms remain poorly understood. In this study, we analyzed 305 angiosperm genomes and found that the proportion of NLR genes containing IDs (NLR-ID genes) ranges from 0% to 38.3%, with an average of 10.6%. A total of 1226 unique IDs were identified, some of which are widely distributed, while others are specific to particular taxa. Notably, 415 of these IDs are homologous to plant proteins targeted by pathogen effectors, suggesting their role as candidate decoys. Comparative analysis of NLR-IDs in two subfamilies—TIR-NLR (TNL) and CC-NLR (CNL)—revealed that TNL genes have a significantly higher frequency of IDs, with the C-JID and DUF3542 domains being most prevalent. N-terminal fusion of the DUF3542 domain in CNL genes correlates with the loss of the MADA motif. Our findings expand the understanding of NLR-ID diversity and provide insights into the dynamic evolution of NLR protein architecture in angiosperms. Full article

Lower limb exoskeleton systems require accurate recognition of the wearer’s movement intentions prior to action execution in order to achieve natural and smooth human–machine interaction. Surface electromyography (sEMG) signals can reflect neural activation of muscles before movement onset, making them a key physiological source for movement intention recognition. To improve sEMG-based recognition performance, this study proposes an innovative deep learning framework, IFIANet. First, a CNN–TCN-based spatiotemporal feature learning network is constructed, which efficiently models and represents multi-scale temporal–frequency features while effectively reducing model parameter complexity. Second, an IFIA (Frequency-Informed Integration Attention) module is designed to incorporate global frequency information, compensating for frequency components potentially lost during time–frequency transformations, thereby enhancing the discriminability and robustness of temporal–frequency features. Extensive ablation and comparative experiments on the publicly available MyPredict1 dataset demonstrate that the proposed framework maintains stable performance across different prediction times and achieves over 82% average recognition accuracy in within-experiments involving nine participants. The results indicate that IFIANet effectively fuses local temporal–frequency features with global frequency priors, providing an efficient and reliable approach for sEMG-based movement intention recognition and intelligent control of exoskeleton systems. Full article

Estimating biomass burning emissions remains challenging due to both the substantial spatiotemporal variability of fires and the inherent uncertainties associated with the limited overpass frequency of polar-orbiting satellites. Integrating geostationary (5–10 min, 2 km) and polar-orbiting (twice daily, 375 m) satellite observations provides a detailed characterization of active fire diurnal cycles. However, the conventional unimodal Gaussian approximation (Pol/Geo-Uni), commonly used in fire emission models, fails to accurately reproduce the diurnal patterns. This study systematically analyzed the seasonal and diurnal variation in different types of active fires (AFs) and fire radiative power (FRP) across North America (GOES-16, ABI) and East Asia (Himawari-8, AHI). In North America, forest and savanna fires exhibited high FRP and a pronounced bimodal diurnal cycle lasting 4–8 h, whereas the corresponding fire types in East Asia exhibited a shorter, unimodal pattern of 2–4 h. Agricultural fires in East Asia were predominantly small in scale with low FRP, and frequently occurred at night. We used a modified Gaussian function to estimate dry matter burned (DMB), quantitating regional emission impacts for different fire types. The fused product (VIIRS/ABI-Bi) yielded amounts of DMB in North America that was 1.22 and 1.24 times higher than that from VIIRS/ABI-Uni and GFASv1.2, respectively. In East Asia, VIIRS/AHI-Bi DMB exceeded those from VIIRS/AHI-Uni and GFASv1.2 by 1.08 and 0.94 times, with agricultural fire estimates during the fire season being 1.18–1.62 times higher. This increase was notably pronounced in eastern China, where VIIRS/AHI-Bi DMB reached 1.76 to 9.77 times higher than estimates from VIIRS/AHI-Uni, GFED5, GFED4.1s, and GFASv1.2. Overall, integrating high spatiotemporal resolution satellite fire products with regionally diurnal models can substantially improve emission estimates, particularly for frequent, small-scale fire events. Full article

Fused deposition modeling of polyether ether ketone offers distinct advantages for fabricating complex and lightweight structures. Although three principal build orientations theoretically exist for practical 3D engineering components, research on their effects remains limited, especially regarding the influence of the interaction between build orientation and printing path on mechanical performance. This study investigated the tensile and shear properties, as well as the failure mechanisms, of FDM-fabricated PEEK under the coupled effects of build orientation and printing path through mechanical testing, fracture morphology analysis, and statistical methods. The results indicate that the printing path exerts a dominant influence on anisotropic behavior, while the interaction between printing path and build orientation jointly governs the shear failure modes. Under identical printing paths, the elongation at break varied by up to twofold across different build orientations, reaching a maximum of 96%, whereas samples printed with W or T paths exhibited elongations at break below 5%. Although shear and tensile moduli remained largely consistent across build orientations, other mechanical properties demonstrated significant differences. Variations in cross-sectional dimensions induced by build orientation markedly affected tensile performance: the coupled effect of build orientation and printing path was found to render the path repetition frequency a critical factor in determining temperature uniformity within the printed region and the quality of interlayer interfaces, thereby constituting the core mechanism underlying anisotropic behavior. Furthermore, larger cross-sections re-duced tensile modulus but enhanced yield strength and elongation at break, highlight-ing the regulatory role of cross-sectional geometry on mechanical response. Based on these findings, a synergistic optimization strategy integrating printing path, build orientation, and tensile–shear performance is proposed to achieve tailored mechanical properties in FDM-fabricated PEEK components. This approach enables controlled enhancement of structural performance to meet diverse application requirements. Full article

In this study, we propose the Frequency-domain Feature Fusion Module (F3M) to address the challenges of underwater object detection, where optical degradation—particularly high-frequency attenuation and low-frequency color distortion—significantly compromises performance. We critically re-evaluate the need for strict invertibility in detection-oriented frequency modeling. Traditional wavelet-based methods incur high computational redundancy to maintain signal reconstruction, whereas F3M introduces a lightweight “Separate–Project–Fuse” paradigm. This mechanism decouples low-frequency illumination artifacts from high-frequency structural cues via spatial approximation, enabling the recovery of fine-scale details like coral textures and debris boundaries without the overhead of channel expansion. We validate F3M’s versatility by integrating it into both Convolutional Neural Networks (YOLO) and Transformer-based detectors (RT-DETR). Evaluations on the SCoralDet dataset show consistent improvements: F3M enhances the lightweight YOLO11n by 3.5% mAP₅₀ and increases RT-DETR-n’s localization accuracy (mAP_50–95) from 0.514 to 0.532. Additionally, cross-domain validation on the deep-sea TrashCan-Instance dataset shows F3M achieving comparable accuracy to the larger YOLOv8n while requiring 13% fewer parameters and 20% fewer GFLOPs. This study confirms that frequency-domain modulation provides an efficient and widely applicable enhancement for real-time underwater perception. Full article

This study aims to examine the acoustic characteristics of Mandarin vowels produced by children with cochlear implants and to explore the differences in their speech production compared with those of children with normal hearing. We propose a multiview model-based method for vowel feature analysis. This approach involves extracting and fusing formant features, Mel-frequency cepstral coefficients (MFCCs), and linear predictive coding coefficients (LPCCs) to comprehensively represent vowel articulation. We conducted k-means clustering on individual features and applied multiview clustering to the fused features. The results showed that children with cochlear implants formed discernible vowel clusters in the formant space, though with lower compactness than those of normal-hearing children. Furthermore, the MFCCs and LPCCs features revealed significant inter-group differences. Most importantly, the multiview model, utilizing fused features, achieved superior clustering performance compared to any single feature. These findings demonstrated that effective fusion of frequency domain features provided a more comprehensive representation of phonetic characteristics, offering potential value for clinical assessment and targeted speech intervention in children with hearing impairment. Full article

Multi-focus images are essential in various computer vision applications. To mitigate artifacts and information loss in multi-focus image fusion, we propose a novel algorithm based on AGPCNN and fractal dimension in the NSCT domain. The source images are decomposed into low- and high-frequency sub-bands via NSCT; the low-frequency components are fused using an averaging rule, while the high-frequency components are fused through fractal dimension and the AGPCNN model, followed by consistency verification to refine the results. Experiments on the Lytro and MFI-WHU datasets show that the proposed method outperforms existing approaches in terms of both visual quality and quantitative metrics. Furthermore, its successful application to multi-sensor and multi-modal image fusion tasks demonstrates the algorithm’s robustness and generality. Full article

Directional frequency analysis and recording (DIFAR) is a widely used sonobuoy in modern underwater acoustic monitoring and surveillance. The sonobuoy is installed in the area of interest, collects underwater data, and transmits the data to nearby aircraft for data analysis. In this process, transmission of a large volume of raw data poses significant challenges due to limited communication bandwidth. To address this problem, existing studies on autoencoder-based methods have drastically reduced amounts of information to be transmitted with moderate data reconstruction errors. However, the information bottleneck inherent in these autoencoder-based methods often leads to significant fidelity degradation. To overcome these limitations, this paper proposes a novel autoencoder method focused on the reconstruction fidelity. The proposed method operates with two key components: Gated Fusion (GF), proven critical for effectively fusing multi-scale features, and Squeeze and Excitation (SE), an adaptive Channel Attention for feature refinement. Quantitative evaluations on a realistic simulated sonobuoy dataset demonstrate that the proposed model achieves up to a 90.36% reduction in spectral mean squared error for linear frequency modulation signals compared to the baseline. Full article

The complex environment of metal mines causes significant noise interference in microseismic signals. This leads to low accuracy and high false alarm rates when using the conventional Short-Term Average/Long-Term Average (STA/LTA) method for first-arrival picking. To address these issues, this paper proposes an improved approach that combines Variational Mode Decomposition (VMD) with STA/LTA(V-STA/LTA). The proposed method selects effective mode components through multimodal decomposition. Subsequently, an energy-weighted fusion is achieved based on energy distribution characteristics to improve the accuracy of arrival time-picking. First, the microseismic signal is decomposed by VMD. The center frequencies of the Intrinsic Mode Functions (IMFs) are then calculated through Fast Fourier Transform (FFT). This helps identify and retain the effective mode components, reducing noise interference. Next, the STA/LTA method is applied to each selected mode component for first-arrival picking. Finally, the results from the different components are fused based on their energy weights for improving picking precision. In low signal-to-noise ratio (SNR) conditions, the effectiveness of the V-STA/LTA method was verified through simulation experiments and field data tests. In theoretical simulations, according to test results from multiple sets of different signal-to-noise ratios, the root mean square error (RMSE) (0.0005) and mean absolute error (MAE) (0.00055) of V-STA/LTA are significantly lower than those of STA/LTA and AIC. In actual data, the average accuracy (99.77%) is nearly 1 percentage point higher than that of the traditional STA/LTA (98.93%), improving the accuracy of microseismic signal arrival time-picking. Full article

This work presents a fused-filament fabrication (FFF) hot end that combines an unrestricted-rotation C-axis with a rectangular-slot nozzle and an induction-heated melt sleeve. The architecture replaces the popular resistive cartridge and heater block design with an external coil that induces eddy-current heating in a thin-walled sleeve, threaded to the heat break and nozzle, reducing thermal mass and eliminating wired sensors across the rotating interface. A contactless infrared thermometer targets the nozzle tip; the temperature is regulated by frequency-modulating the inverter around resonance, yielding stable control. The hot end incorporates an LPBF-manufactured nozzle, which transitions from a circular inlet to a rectangular outlet to deposit broad, low-profile strands at constant layer height while preserving lateral resolution. The concept is validated on a desktop Cartesian platform retrofitted to coordinate yaw with XY motion. A twin-printer testbed compares the proposed hot end against a stock cartridge-heated system under matched materials and environments. With PLA, the induction-heated, rotating hot end enables printing at 170 °C with defect-free flow and delivers substantial reductions in job time (22–49%) and energy per part (9–39%). These results indicate that the proposed approach is a viable route to higher-throughput, lower-specific-energy material extrusion. Full article

Accurate segmentation of lesion regions is essential for skin cancer diagnosis. As dermoscopic images of skin lesions demonstrate different sizes, diverse shapes, fuzzy boundaries, and so on, accurate segmentation still faces great challenges. To address these issues, we propose a new dermatologic image segmentation network, FFM-Net. In FFM-Net, we design a new FM block encoder based on state space models (SSMs), which integrates a low-frequency information extraction module (LEM) and an edge detail extraction module (EEM) to extract broader overall structural information and more accurate edge detail information, respectively. At the same time, we dynamically adjust the input channel ratios of the two module branches at different stages of our network, so that the model can learn the correlation relationship between the overall structure and edge detail features more effectively. Furthermore, we designed the cross-channel spatial attention (CCSA) module to improve the model’s sensitivity to channel and spatial dimensions. We deploy a multi-level feature fusion module (MFFM) at the bottleneck layer to aggregate rich multi-scale contextual representations. Finally, we conducted extensive experiments on three publicly available skin lesion segmentation datasets, ISIC2017, ISIC2018, and PH2, and the experimental results show that the FFM-Net model outperforms most existing skin lesion segmentation methods. Full article

This paper presents a comparative analysis of the influence of Fused Deposition Modeling (FDM) parameters—specifically material type, infill geometry, and density—on the vibro-acoustic characteristics of loudspeaker enclosures. The enclosures were designed as exponential horns to intensify resonance phenomena for precise evaluation. Twelve unique configurations were fabricated using three materials with distinct damping properties (PLA, ABS, wood-composite) and three internal geometries (linear, honeycomb, Gyroid). Key vibro-acoustic properties were assessed via digital signal processing of recorded audio signals, including relative frequency response and time-frequency (spectrogram) analysis, and correlated with a predictive Finite Element Analysis (FEA) model of mechanical vibrations. The study unequivocally demonstrates that a material with a high internal damping coefficient is a critical factor. The wood-composite enabled a reduction in the main resonance amplitude by approximately 4 dB compared to PLA with the same geometry, corresponding to a predicted 86% reduction in mechanical vibration. Furthermore, the results show that a synergy between a high-damping material and an advanced, energy-dissipating infill (Gyroid) is crucial for achieving high acoustic fidelity. The wood-composite with 10% Gyroid infill was identified as the optimal design, offering the most effective resonance damping and the most neutral tonal characteristic. This work provides a valuable contribution to the field by establishing a clear link between FDM parameters and acoustic outcomes, delivering practical guidelines for performance optimization in personalized audio systems. Full article

A major computational bottleneck in classifying large-scale hyperspectral images (HSI) is the mandatory data decompression prior to processing. Compressed-domain computing offers a solution by enabling deep learning on partially compressed data. However, existing compressed-domain methods are predominantly tailored for the Discrete Cosine Transform (DCT) used in natural images, while HSIs are typically compressed using the Discrete Wavelet Transform (DWT). The fundamental structural mismatch between the block-based DCT and the hierarchical DWT sub-bands presents two core challenges: how to extract features from multiple wavelet sub-bands, and how to fuse these features effectively? To address these issues, we propose a novel framework that extracts and fuses features from different DWT sub-bands directly. We design a multi-branch feature extractor with sub-band feature alignment loss that processes functionally different sub-bands in parallel, preserving the independence of each frequency feature. We then employ a sub-band cross-attention mechanism that inverts the typical attention paradigm by using the sparse, high-frequency detail sub-bands as queries to adaptively select and enhance salient features from the dense, information-rich low-frequency sub-bands. This enables a targeted fusion of global context and fine-grained structural information without data reconstruction. Experiments on three benchmark datasets demonstrate that our method achieves classification accuracy comparable to state-of-the-art spatial-domain approaches while eliminating at least 56% of the decompression overhead. Full article

The safe and stable operation of distribution networks relies on the real-time monitoring, analysis, and feedback of power quality data. However, with the continuous advancement of distribution network construction, the number of distributed power electronic devices has increased significantly, leading to frequent power quality issues such as voltage fluctuations, harmonic pollution, and three-phase unbalance in distribution terminals. Therefore, the augmentation and processing of power quality data have become crucial for ensuring the stable operation of distribution networks. Traditional methods for augmenting and processing power quality data fail to consider the differentiated characteristics of burrs in signal sequences and neglect the comprehensive consideration of both time-domain and frequency-domain features in disturbance identification. This results in the distortion of high-frequency fault information, and insufficient robustness and accuracy in identifying Power Quality Disturbance (PQD) against the complex noise background of distribution networks. In response to these issues, we propose a power quality data augmentation and processing method for distribution terminals considering high-frequency sampling. Firstly, a burr removal method of the sampling waveform based on a high-frequency filter operator is proposed. By comprehensively considering the characteristics of concavity and convexity in both burr and normal waveforms, a high-frequency filtering operator is introduced. Additional constraints and parameters are applied to suppress sequences with burr characteristics, thereby accurately eliminating burrs while preserving the key features of valid information. This approach avoids distortion of high-frequency fault information after filtering, which supports subsequent PQD identification. Secondly, a PQD identification method based on a dual-channel time–frequency feature fusion network is proposed. The PQD signals undergo an S-transform and period reconfiguration to construct matrix image features in the time–frequency domain. Finally, these features are input into a Convolutional Neural Network (CNN) and a Transformer encoder to extract highly coupled global features, which are then fused through a cross-attention mechanism. The identification results of PQD are output through a classification layer, thereby enhancing the robustness and accuracy of disturbance identification against the complex noise background of distribution networks. Simulation results demonstrate that the proposed algorithm achieves optimal burr removal and disturbance identification accuracy. Full article