Search Results (155)

Hyperspectral image (HSI) stripe noise removal is essential for downstream interpretation tasks. However, most existing methods exhibit incomplete joint modeling of spatial structures and inter-band spectral correlations, lack direction-aware modeling for stripe noise, and lack differentiated processing of high- and low-frequency components. To tackle these limitations, we propose a spatial-spectral attention-enhanced multi-level wavelet-informed network (SAMWNet). Its dual-branch module extracts spatial and spatial-spectral features from each band and its adjacent bands. Afterward, a discrete wavelet-informed progressive denoising (MDWPD) module conducts multi-level Haar wavelet decomposition and progressive reconstruction. Within this module, the low-frequency hybrid enhancement (LFHE) module preserves low-frequency spectral structures, while the high-frequency enhancement (HFME) module suppresses directional stripe artifacts in high-frequency subbands. We further adopt a composite loss function to balance pixel fidelity, noise estimation, structural similarity, and spectral consistency. Experimental results on simulated and real-world HSIs demonstrate that SAMWNet achieves competitive or superior performance compared with several representative HSI denoising methods. Full article

Small object detection in remote sensing imagery is a fundamental task for visual information extraction, yet it remains challenging due to extremely small target scales, complex backgrounds, and the loss of discriminative feature information caused by repeated downsampling. To address these issues, this paper proposes a Wavelet-Convolution and Fine-Grained Preservation Network (WCFPNet) based on YOLOv8n. Specifically, a Wavelet-Convolution Module (WCM) is introduced into the backbone to decompose feature maps into low- and high-frequency sub-bands, thereby enhancing structural feature modeling and preserving subtle target details. To compensate for the weakened fine-grained information after repeated downsampling, an Enhanced Spatial Pyramid Pooling-Fast (ESPPF) module is embedded at the end of the backbone to strengthen multi-scale contextual aggregation. In addition, an Enhanced Feature Pyramid Network (EFPN) is designed in the neck to facilitate the propagation of shallow and intermediate fine-grained features to high-level semantic features through cross-level fusion and the Convolutional Block Attention Module (CBAM). Experiments on the NWPU VHR-10 dataset show that WCFPNet achieves 0.879 mAP@0.5 and 0.515 mAP@0.5:0.95, outperforming YOLOv8n by 1.7 and 2.5 percentage points, respectively. Moreover, the proposed WCFPNet achieves a competitive performance compared with several representative detectors while maintaining moderate model complexity. These results demonstrate the effectiveness of WCFPNet in challenging remote sensing scenes characterized by complex backgrounds, dense object distributions, and weak textures. Full article

We study whether the semantic category of a hidden display terminal can be inferred from a wall-mediated indirect observation when the display remains outside the camera field of view under a controlled and calibrated scene configuration. This setting provides a security-motivated feasibility test for indirect optical semantic leakage, but it remains challenging for two reasons. First, indirect propagation makes the wall pattern dominated by the occluder contour, while category-bearing evidence survives only as weak radiometric variations, making stable extraction difficult. Second, even after front-end recovery, low-frequency support is relatively stable, whereas the mid- and high-frequency details required for class separation remain weak and distortion-prone; as a result, the classifier may drift toward dominant but weakly informative coarse-grained patterns and fail to consistently accumulate fine-grained discriminative cues. We propose SCISA-Net, which combines scene-constrained inversion with multi-stage Haar-subband attention to reorganize indirect observations, compensate residual feature degradation, and aggregate class-relevant subband evidence. Experiments on a paired 31-class benchmark show stable recognition, robustness to illumination attenuation and ambient background interference, matched scene-operator re-parameterization capability, and clear degradation when key inverse or subband components are disrupted. These results support the feasibility of category-level semantic inference from calibrated wall-mediated indirect observations. Full article

Shortwave infrared (SWIR) imagery plays an important role in land–water boundary delineation, coastal monitoring, and complex aquatic environment observation. However, the spatial resolution of SWIR bands is usually lower than that of visible bands, which limits their capability to represent fine-scale targets and boundary structures. To address this problem, this study proposes MLE-ResUNet, a SWIR image super-resolution method that integrates along-track oversampling with visible-light-guided deep learning. The proposed method first exploits dual-view SWIR observations with sub-pixel displacement generated by increasing the sampling line rate in the push-broom imaging process. A maximum likelihood estimation (MLE)-based physical prior module is then introduced to transform multi-view degraded observations into a physically consistent latent high-resolution prior. Finally, high-resolution visible images are used to provide edge, texture, and structural guidance, and a ResUNet-based network is employed for multi-source feature fusion and residual reconstruction. Based on multi-region measured data acquired by the LHRSI (Lightweight High-Resolution Spectral Imager) payload onboard the BlueCarbon-1A satellite, a SWIR super-resolution dataset covering typical urban, farmland, and coastal scenarios was constructed. Comparative experiments were conducted against PCA, BDSD, PanNet, GPPNN, and two additional lightweight-guided deep learning baselines, namely LGPConv and a CANConv-style visible-guided baseline. The results show that MLE-ResUNet achieves the best performance across different scenarios and consistently outperforms the comparison methods in terms of SSIM, SAM, ERGAS, and Q-index. The proposed method effectively enhances spatial detail recovery while maintaining favorable spectral consistency. Ablation experiments further demonstrate that both along-track oversampling information and the MLE-based physical prior contribute to improved reconstruction quality and more stable training convergence. These findings indicate that the proposed method can enhance fine-scale SWIR observation capability without substantially increasing hardware complexity, providing an effective technical solution for shoreline identification, land–water boundary extraction, and complex surface target monitoring. Full article

Fatigue monitoring and detection are crucial for improving efficiency and safety due to their influence on reducing cognitive and physical performance that may result in safety-related incidents. This paper proposes a filter bank-based approach that decomposes electroencephalography (EEG) signals into delta, theta, alpha, beta, and gamma sub-bands for feature extraction to enhance fatigue detection using a wearable EEG-based brain–computer interface (BCI). The study utilized a publicly available EEG dataset from 40 participants collected with a dry-EEG headband while performing two cognitive tasks: a Cognitive Vigilance Task (CVT) and a Multi-Modal Integration Task (MMIT). The data was previously investigated for stress detection on the MMIT. In this study, we investigate fatigue detection on the CVT. Subjects who were not fatigued post-CVT were iteratively removed. Two models were trained with five models to classify the fatigued state from the non-fatigued state, one using features extracted from a broadband filter approach and the other from the proposed filter bank approach. Leave-one-subject-out cross-validation yielded accuracies of 75.8% $\pm$10.4% (95% confidence interval) from the broadband filter approach, and 86.4% $\pm$8.3% (95% confidence interval) from the proposed filter bank approach, yielding an overall increase of 10.6%. These results demonstrate the potential of filter bank-based feature extraction for fatigue detection in wearable EEG-based BCI systems. Full article

Optical object detection under fog-induced atmospheric degradation remains a challenging problem for terrestrial sensing and monitoring systems. Atmospheric scattering reduces image contrast and attenuates high-frequency edge and texture features that are important for precise object localization, while standard downsampling in convolutional neural networks (CNNs) further amplifies this information loss during feature extraction. Existing spatial-domain methods largely improve pixel appearance or feature refinement without explicitly preserving fog-weakened high-frequency edge and texture features during feature extraction. To address this issue, we propose WFSCA-YOLO, a frequency-aware and feature-preserving detection framework with cross-domain fusion between frequency-domain details and spatial semantic responses. The framework introduces the Wavelet-driven Frequency–spatial Co-awareness Block (WFSCA-Block) into YOLOv8, where the Discrete Wavelet Transform (DWT) is used to decompose feature maps into multi-directional high-frequency subbands and preserve high-frequency edge and texture features degraded by atmospheric scattering. A Cross-Domain Feature Selector (CDFS) is further designed to adaptively recalibrate the fusion of frequency-domain details and spatial semantic responses under varying visibility conditions. Experiments on synthetic and real-world degraded optical benchmarks from near-ground scenes, namely Foggy Cityscapes and RTTS, show that WFSCA-YOLO consistently outperforms representative state-of-the-art methods, achieving 50.3% mAP@50 on Foggy Cityscapes (2.1 percentage points above the baseline) and a mean mAP@50 of 79.28% on RTTS over three independent runs. Under a unified FP32 batch-1 inference benchmark, WFSCA-YOLO runs at 134.76 FPS on an RTX 4090D, indicating real-time capability with only a slight latency increase relative to the YOLOv8-s baseline. These results indicate that preserving high-frequency edge and texture features is an effective strategy for robust perception under degraded visibility and offers practical potential for terrestrial sensing and monitoring platforms. Full article

The unilateral limb motor imagery paradigm can effectively address the cognitive dissociation problem among multiple limbs and provide strong technical support for extending the functionality of external devices. However, feature mining and accurate decoding of unilateral limb movements remain challenging. In this study, we propose a feature mining method that combines automatic frequency band boundary localization with regularized common spatial pattern (AFBBL-RCSP), and employ a pinball-loss-based twin support vector machine (Pin-UTSVM) to decode EEG signals corresponding to reaching, turning, and grasping movements. First, multiple optimal frequency band boundaries were identified for each subject using AFBBL. Then, regularized spatial features were extracted from each sub-band, and all features were reduced using Fisher’s discriminant analysis. Finally, the Pin-UTSVM classifier was used to categorize the three types of movement data. The results show that, compared with CSP and RCSP feature mining methods using the fixed 8–30 Hz band, the proposed method improves decoding accuracy by 9.52% and 3.89%, respectively. Compared with fixed single-band feature mining methods based on the α band, β band, and α + β band, the proposed method improves accuracy by 5.56%, 3.89%, and 3.73%, respectively. In addition, compared with existing unilateral limb decoding methods based on temporal-spatial features, temporal-frequency features, and temporal-spatial-temporal-frequency fusion CNN features, the proposed method improves decoding accuracy by 34.93%, 34.09%, and 28.11%, respectively. These results suggest that the proposed AFBBL-RCSP method is effective for unilateral limb motor imagery EEG decoding. Full article

Unmanned aerial vehicle (UAV) imagery offers a promising alternative to manual and vehicle-based inspection for highway pavement distress detection, but the high-angle perspective reduces the relative size and feature richness of small distresses and amplifies aliasing during downsampling, limiting the accuracy of existing detectors. To address these problems, this paper proposes an improved YOLOv8 algorithm with four coordinated modifications: (i) a Feature-Focusing Diffusion Pyramid Network (FFDPN) that replaces the conventional PAN to strengthen multi-scale feature fusion and preserve fine-grained details; (ii) an Information Interaction Detection Head (IIDH) that replaces the decoupled dual-branch head, sharing interaction features between the classification and regression branches via deformable convolution (DCNv2) to reduce parameters while improving task synergy; (iii) an Edge Information Extraction Module (EIEM) placed at the front of the backbone, which uses Sobel-based gradient response plus max-pooling to inject low-level edge priors; and (iv) a WaveletPool downsampling operator that decomposes features into LL/LH/HL/HH sub-bands to suppress aliasing of small-scale distresses. Experiments on 3408 UAV images of four distress categories (transverse, longitudinal, and alligator cracks and potholes) show that the improved model reaches 93.7% Precision, 89.6% Recall, and 96.0% mAP@0.50—a 12.2 percentage-point gain over YOLOv8n—while using only 2.41 × 10⁶ parameters and outperforming Faster R-CNN, DETR, YOLOv7-tiny, YOLOv9, YOLOv10n, YOLOv11n, and YOLO-World on the same benchmark. The model eliminates the duplicate and missed detections observed in baselines, at a moderate cost in FPS (30.3 vs. 57.1 for YOLOv8n). Full article

Magnetic resonance imaging (MRI) remains one of the most important tests for diagnosing and monitoring various diseases. In recent years, machine learning methods have been widely applied to automate MRI analysis. It supports decision-making by predicting disease and highlighting relevant regions. However, the proper use of feature extraction methods can improve the performance of the model. This paper proposes a WaveletFusion architecture that combines a two-dimensional Haar wavelet decomposition with a convolutional neural network (CNN) for classification. The approach was demonstrated on the Brain Tumor MRI dataset and further examined on the Br35H :: Brain Tumor Detection 2020 (Br35H). The model decomposes each MRI slice into approximation and directional detail subbands and fuses multi-scale wavelet features within the convolutional pipeline. To evaluate the effect of decomposition depth, WaveletFusion variants from one to eight levels were compared with a Baseline CNN model under the same training protocol. The results showed that performance improved progressively with increasing decomposition depth up to level 7, whereas the 8-level configuration consistently declined, indicating that excessive decomposition introduces information loss and over-compression in the deepest approximation pathway. The best-performing configuration, which outperformed both the Baseline CNN and the WaveletFusion variations in five independent runs, was the 7-level WaveletFusion model, achieving a test accuracy of 0.94 ± 0.01 and test macro-F1 of 0.93 ± 0.02. A similar tendency was observed on the Br35H dataset, where the 7-level model achieved a 0.97 ± 0.01 test accuracy and 0.97 ± 0.01 test macro-F1, while the 8-level configuration remained weaker on both datasets. These results show that multi-scale wavelet fusion can improve Brain Tumor MRI classification while maintaining a compact model size and a fair comparison setting, and that the decomposition depth must be selected carefully. Full article

High-quality maritime infrared images are crucial for accurate object detection, classification, and segmentation in maritime environments. However, maritime infrared images are often degraded by various types of noise, including non-uniform noise and detector non-uniformity-induced fixed-pattern noise (e.g., vertical stripe noise), which pose significant challenges for the aforementioned high-level vision tasks. A novel network, termed SWDNet (Spatial–Wavelet Joint Denoising Network), is proposed to jointly model spatial- and wavelet-domain features, enabling the effective enhancement of maritime infrared image quality while preserving fine image details. Two parallel sub-networks with distinct architectures are employed to extract complementary information for maritime infrared image denoising. In the upper branch, hierarchical spatial attention aggregation (HSAA) modules are employed at multiple scales to extract spatial features and adaptively assign importance weights to different spatial locations. The lower branch employs a Haar-based DWT for sub-band decomposition, a pixel-grouped self-attention module for boundary refinement, and parallel multi-scale horizontal convolutions to suppress vertical stripe noise in the HL sub-band. Finally, the directional edge enhancement (DEE) module employs learnable Sobel operators in conjunction with multi-layer convolutions to effectively extract and enhance directional edge features. Experimental results demonstrate that, compared with state-of-the-art methods, the proposed SWDNet achieves superior denoising performance on both synthetic and real maritime infrared datasets. Full article

Accurate photovoltaic (PV) power forecasting is crucial for grid stability and the integration of renewable energy. To address the multiscale, nonlinear characteristics of PV power series and the limitations of traditional methods in dynamic error compensation, a novel hybrid forecasting framework is proposed, integrating Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Convolutional Neural Network-Bidirectional Long Short-Term Memory (CNN-BiLSTM), and a Simplified Deep Q-Network (SDQN). The framework first decomposes the power series into subcomponents across different frequency bands via CEEMDAN. Subsequently, dedicated CNN-BiLSTM sub-models are employed in parallel to extract spatiotemporal features from each component. Finally, an SDQN agent is introduced to perform real-time error compensation. Validation based on operational data from a PV plant in Ningxia, China, demonstrates that the proposed framework achieves RMSE, MAE, MAPE, and R² values of 0.4463, 0.1256, 1.2814%, and 92.58%, respectively, significantly outperforming benchmark models. Specifically, the CEEMDAN decomposition effectively mitigates mode mixing. The CNN-BiLSTM as the base predictor reduces RMSE by 25.04–65.68% compared to mainstream models. Furthermore, the SDQN compensation mechanism delivers an additional 24.5% reduction in prediction error. The proposed approach thus constitutes a high-precision, adaptive solution for PV power forecasting. Full article

Adversarial examples are crucial tools for assessing the robustness of deep neural networks (DNNs) and revealing potential security vulnerabilities. Adversarial example generation methods based on Generative Adversarial Networks (GANs) have made significant progress in generating image adversarial examples, but still suffer from insufficient sparsity and transferability. To address these issues, this study proposes a novel semi-white-box untargeted adversarial example generation method named Wavelet-AdvGAN, with an explicit threat model defined as follows. Specifically, the attack is strictly untargeted without predefined target categories, aiming solely to mislead DNNs into classifying adversarial examples into any category other than the original label. It adopts a semi-white-box setting where attackers are denied access to the target model’s private information. Regarding the generator’s information dependence, the training phase only utilizes public resources (i.e., the target model’s public architecture and CIFAR-10 public training data), while the test phase generates adversarial examples through one-step feedforward of clean images without interacting with the target model. The method incorporates a Frequency Sub-band Difference (FSD) module and a Wavelet Transform Local Feature (WTLF) extraction module, evaluating the differences between original and adversarial examples from the frequency domain perspective. This approach constrains the magnitude of perturbations, reinforces feature regions, and further enhances the attack effectiveness, thereby improving the sparsity and transferability of adversarial examples. Experimental results demonstrate that the Wavelet-AdvGAN method achieves an average increase of 1.26% in attack success rates under two defense strategies—data augmentation and adversarial training. Additionally, the adversarial transferability improves by an average of 2.7%. Moreover, the proposed method exhibits a lower $l_0$ norm, indicating better perturbation sparsity. Consequently, it effectively evaluates the robustness of deep neural networks. Full article

Electroencephalography (EEG)-based brain–computer interface (BCI) mimics the brain’s intrinsic information-processing mechanisms by translating neural oscillations into actionable commands. In motor imagery (MI) BCI, imagined movements evoke characteristic patterns over the sensorimotor cortex, forming a biomimetic channel through which internal motor intentions are decoded. However, this biomimetic interaction is highly vulnerable to signal degradation, particularly in mobile or low-resource environments where low sampling frequencies obscure these MI-related oscillations. To address this limitation, we propose a robust MI classification framework that integrates spatial, spectral, and temporal dynamics through a filter bank common spatial pattern with time segmentation (FBCSP-TS). This framework classifies motor imagery tasks into four classes (left hand, right hand, foot, and tongue), segments EEG signals into overlapping time domains, and extracts frequency-specific spatial features across multiple subbands. Segment-level predictions are combined via soft voting, reflecting the brain’s distributed integration of information and enhancing resilience to transient noise and localized artifacts. Experiments performed on BCI Competition IV datasets 2a (250 Hz) and 1 (100 Hz) demonstrate that FBCSP-TS outperforms CSP and FBCSP. A paired t-test confirms that accuracy at 110 Hz is not significantly different from that at 250 Hz (p < 0.05), supporting the robustness of the proposed framework. Optimal temporal parameters (window length = 3.5 s, moving length = 0.5 s) further stabilize transient-signal capture and improve SNR. External validation yielded a mean accuracy of 0.809 ± 0.092 and Cohen’s kappa of 0.619 ± 0.184, confirming strong generalizability. By preserving MI-relevant neural patterns under degraded conditions, this framework advances practical, biomimetic BCI suitable for wearable and real-world deployment. 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

Current lake and reservoir water body extraction algorithms are confronted with two critical challenges: (1) design dependency on specific geographical features, leading to constrained cross-regional adaptability (e.g., the JRC Global Water Body Dataset achieves ~90% overall accuracy globally, while the ESA WorldCover 2020 reaches ~92% for water body classification, both showing degraded performance in complex karst terrains); (2) information loss due to cloud occlusion, compromising dynamic monitoring accuracy. To address these limitations, this study presents a multi-feature fusion and multi-level hierarchical extraction algorithm for lake and reservoir water bodies, leveraging the Google Earth Engine (GEE) cloud platform and Sentinel-2 multispectral imagery in the karst landscape of Bijie City. The proposed method integrates the Automated Water Extraction Index (AWEIsh) and Modified Normalized Difference Water Index (MNDWI) for initial water body extraction, followed by a comprehensive fusion of multi-source data—including Normalized Difference Vegetation Index (NDVI), Normalized Difference Built-up Index (NDBI), Normalized Difference Red-Edge Index (NDREI), Sentinel-2 B8/B9 spectral bands, and Digital Elevation Model (DEM). This strategy hierarchically mitigates vegetation shadows, topographic shadows, and artificial feature non-water targets. A temporal flood frequency algorithm is employed to restore cloud-occluded water bodies, complemented by morphological filtering to exclude non-target water features (e.g., rivers and canals). Experimental validation using high-resolution reference data demonstrates that the algorithm achieves an overall extraction accuracy exceeding 96% in Bijie City, effectively suppressing dark object interference (e.g., false positives due to topographic and anthropogenic features) while preserving water body boundary integrity. Compared with single-index methods (e.g., MNDWI), this method reduces false positive rates caused by building shadows and terrain shadows by 15–20%, and improves the IoU (Intersection over Union) by 6–13% in typical karst sub-regions. This research provides a universal technical framework for large-scale dynamic monitoring of lakes and reservoirs, particularly addressing the challenges of regional adaptability and cloud compositing in karst environments. Full article