Search Results (1,514)

Converter stations are pivotal in high-voltage direct current (HVDC) systems, enabling power conversion between an alternating current (AC) and a direct current (DC) while ensuring efficient and stable energy transmission. Fault detection in converter stations is crucial for maintaining their reliability and operational safety. This paper focuses on image-based detection of five common faults: metal corrosion, discoloration of desiccant in breathers, insulator breakage, hanging foreign objects, and valve cooling water leakage. Despite advancements in deep learning, existing detection methods face two major challenges: limited model generalization due to diverse and complex backgrounds in converter station environments and sparse supervision signals caused by the high cost of collecting labeled images for certain faults. To overcome these issues, we propose InvMOE, a novel fault detection algorithm with two core components: (1) invariant representation learning, which captures task-relevant features and mitigates background noise interference, and (2) multi-task training using a mixture of experts (MOE) framework to adaptively optimize feature learning across tasks and address label sparsity. Experimental results on real-world datasets demonstrate that InvMOE achieves superior generalization performance and significantly improves detection accuracy for tasks with limited samples, such as valve cooling water leakage. This work provides a robust and scalable approach for enhancing fault detection in converter stations. Full article

Clustering is a powerful and efficient technique for pattern recognition which improves classification accuracy. In machine learning, it is a useful unsupervised learning approach due to its simplicity and efficiency for clustering applications. The curse of dimensionality poses a significant challenge as the volume of data increases with rapid technological advancement. It makes traditional methods of analysis inefficient. Sparse clustering is essential for efficiently processing and analyzing large-scale, high-dimensional data. They are designed to handle and process sparse data efficiently since most elements are zero or lack information. In data science and engineering applications, they play a vital role in taking advantage of the natural sparsity in data to save computational resources and time. Motivated by recent sparse k-means and k-means++ algorithms, we propose two novel Lasso-based k-means++ (Lasso-KM++) clustering algorithms, Lasso-KM1++ and Lasso-KM2++, which incorporate Lasso regularization to enhance feature selection and clustering accuracy. Both Lasso-KM++ algorithms can shrink the irrelevant features towards zero, and select relevant features effectively by exploring better clustering structures for datasets. We use numerous synthetic and real datasets to compare the proposed Lasso-KM++ with k-means, k-means++ and sparse k-means algorithms based on the six performance measures of accuracy rate, Rand index, normalized mutual information, Jaccard index, Fowlkes–Mallows index, and running time. The results and comparisons show that the proposed Lasso-KM++ clustering algorithms actually improve both the speed and the accuracy. They demonstrate that our proposed Lasso-KM++ algorithms, especially for Lasso-KM2++, outperform existing methods in terms of efficiency and clustering accuracy. Full article

Helmet-wearing detection for electric vehicle riders is essential for traffic safety, yet existing detection models often suffer from high target occlusion and low detection accuracy in complex road environments. To address these issues, this paper proposes YOLO-CBF, an improved YOLOv7-based detection network. The proposed model integrates coordinate convolution to enhance spatial information perception, optimizes the Focal EIOU loss function, and incorporates the BiFormer dynamic sparse attention mechanism to achieve more efficient computation and dynamic content perception. These enhancements enable the model to extract key features more effectively, improving detection precision. Experimental results show that YOLO-CBF achieves an average mAP of 95.6% for helmet-wearing detection in various scenarios, outperforming the original YOLOv7 by 4%. Additionally, YOLO-CBF demonstrates superior performance compared to other mainstream object detection models, achieving accurate and reliable helmet detection for electric vehicle riders. Full article

In cognitive radio, compressive spectrum sensing (CSS) is critical for efficient wideband spectrum sensing (WSS). However, traditional reconstruction algorithms exhibit suboptimal performance, and conventional WSS methods fail to fully capture the inherent structural information of wideband spectrum signals. Moreover, most existing deep learning-based approaches fail to effectively exploit the sparse structures of wideband spectrum signals, resulting in limited reconstruction performance. To overcome these limitations, we propose BEISTA-Net, a deep learning-based framework for reconstructing compressed wideband signals. BEISTA-Net integrates the iterative shrinkage-thresholding algorithm (ISTA) with deep learning, thereby extracting and enhancing the block sparsity features of wideband spectrum signals, which significantly improves reconstruction accuracy. Next, we propose BSWSS-Net, a lightweight network that efficiently leverages the sparse features of the reconstructed signal to enhance WSS performance. By jointly employing BEISTA-Net and BSWSS-Net, the challenges in CSS are effectively addressed. Extensive numerical experiments demonstrate that our proposed CSS method achieves state-of-the-art performance across both low and high signal-to-noise ratio scenarios. Full article

This paper analyzes two main categories of feature selection: filter methods (such as minimum redundancy maximum relevance, CHI2, Kruskal–Wallis, and ANOVA) and embedded methods (such as alternating direction method of multipliers (BP_ADMM), least absolute shrinkage and selection operator, and orthogonal matching pursuit). The mathematical foundations of feature selection methods inspired by compressed detection are presented, highlighting how the principles of sparse signal recovery can be applied to identify the most relevant features. The results have been obtained using two biomedical databases. The used algorithms have, as their starting point, the notion of sparsity, but the version implemented and tested in this work is adapted for feature selection. The experimental results show that BP_ADMM achieves the highest classification accuracy (77% for arrhythmia_database and 100% for oncological_database), surpassing both the full feature set and the other methods tested in this study, which makes it the optimal feature selection option. The analysis shows that embedded methods strike a balance between accuracy and efficiency by selecting features during the model training, unlike filtering methods, which ignore feature interactions. Although more accurate, embedded methods are slower and depend on the chosen algorithm. Although less comprehensive than wrapper methods, they offer a strong trade-off between speed and performance when computational resources allow for it. Full article

The availability of high-quality and ample synthetic aperture radar (SAR) image datasets is crucial for understanding and recognizing target characteristics. However, in practical applications, the limited availability of SAR target images significantly impedes the advancement of SAR interpretation methodologies. In this study, we introduce a Generative Adversarial Network (GAN)-based approach designed to manipulate the target azimuth angle with few samples, thereby generating high-quality target images with adjustable angle ranges. The proposed method consists of three modules: a generative fusion local module conditioned on image features, a controllable angle generation module based on sparse representation, and an angle discrimination module based on scattering point extraction. Consequently, the generative modules fuse semantically aligned features from different images to produce diverse SAR samples, whereas the angle synthesis module constructs target images within a specified angle range. The discriminative module comprises a similarity discriminator to distinguish between authentic and synthetic images to ensure the image quality, and an angle discriminator to verify that generated images conform to the specified range of the azimuth angle. Combining these modules, the proposed methodology is capable of generating azimuth angle-controllable target images using only a limited number of support samples. The effectiveness of the proposed method is not only verified through various quality metrics, but also examined through the enhanced distinguishability of target recognition methods. In our experiments, we achieved SAR image generation within a given angle range on two datasets. In terms of generated image quality, our method has significant advantages over other methods in metrics such as FID and SSIM. Specifically, the FID was reduced by up to 0.37, and the SSIM was increased by up to 0.46. In the target recognition experiments, after augmenting the data, the accuracy improved by 6.16% and 3.29% under two different pitch angles, respectively. This demonstrates that our method has great advantages in the SAR image generation task, and the research content is of great value. Full article

Sparse autoencoders have recently produced dictionaries of high-dimensional vectors corresponding to the universe of concepts represented by large language models. We find that this concept universe has interesting structure at three levels: (1) The “atomic” small-scale structure contains “crystals” whose faces are parallelograms or trapezoids, generalizing well-known examples such as (man:woman::king:queen). We find that the quality of such parallelograms and associated function vectors improves greatly when projecting out global distractor directions such as word length, which is efficiently performed with linear discriminant analysis. (2) The “brain” intermediate-scale structure has significant spatial modularity; for example, math and code features form a “lobe” akin to functional lobes seen in neural fMRI images. We quantify the spatial locality of these lobes with multiple metrics and find that clusters of co-occurring features, at coarse enough scale, also cluster together spatially far more than one would expect if feature geometry were random. (3) The “galaxy”-scale large-scale structure of the feature point cloud is not isotropic, but instead has a power law of eigenvalues with steepest slope in middle layers. We also quantify how the clustering entropy depends on the layer. Full article

Autonomous navigation is key to improving efficiency and addressing labor shortages in the fruit industry. Semi-structured orchards, with straight tree rows, dense weeds, thick canopies, and varying light conditions, pose challenges for tree identification and navigation line extraction. Traditional 3D lidars suffer from a narrow vertical FoV, sparse point clouds, and high costs. Furthermore, most lidar-based tree-row-detection algorithms struggle to extract high-quality navigation lines in scenarios with thin trunks and dense foliage occlusion. To address these challenges, we developed a 3D perception system using a servo motor to control the rolling motion of a 2D lidar, constructing 3D point clouds with a wide vertical FoV and high resolution. In addition, a method for trunk feature point extraction and tree row line detection for autonomous navigation has been proposed, based on trunk geometric features and RANSAC. Outdoor tests demonstrate the system’s effectiveness. At speeds of 0.2 m/s and 0.5 m/s, the average distance errors are 0.023 m and 0.016 m, respectively, while the average angular errors are 0.272° and 0.146°. This low-cost solution overcomes traditional lidar-based navigation method limitations, making it promising for autonomous navigation in semi-structured orchards. Full article

This study introduces an enhanced ORB-SLAM3 algorithm to address the limitations of traditional visual SLAM systems in feature extraction and localization accuracy within the challenging terrains of open-pit mining environments. It also tackles the issue of sparse point cloud maps for mobile robot navigation. By combining point-line features with a Micro-Electro-Mechanical System (MEMS) Inertial Measurement Unit (IMU), the algorithm improves the feature matching’s reliability, particularly in low-texture areas. The method integrates dense point cloud mapping and an octree structure, optimizing both navigation and path planning while reducing storage demands and improving query efficiency. The experimental results using the TUM dataset and conducting tests in a simulated open-pit mining environment show that the proposed algorithm reduces the absolute trajectory error by 44.33% and the relative trajectory error by 14.34% compared to the ORB-SLAM3. The algorithm generates high-precision dense point cloud maps and uses an octree structure for efficient 3D spatial representation. In simulated open-pit mining scenarios, the dense mapping outperforms at reconstructing complex terrains, especially in low-texture gravel and uneven surfaces. These results highlight the robustness and practical applicability of the algorithm in dynamic and challenging environments, such as open-pit mining. Full article

In this paper, we present a novel color edge detection method that integrates collaborative filtering with multiscale gradient fusion. The Block-Matching and 3D (BM3D) filter is utilized to enhance sparse representations in the transform domain, effectively reducing noise. The multiscale gradient fusion technique compensates for the loss of detail in single-scale edge detection, thereby improving both edge resolution and overall quality. RGB images from the dataset are converted into the XYZ color space through mathematical transformations. The Colored Block-Matching and 3D (CBM3D) filter is applied to the sparse images to reduce noise. Next, the vector gradients of the color image and anisotropic Gaussian directional derivatives for two scale parameters are computed. These are then averaged pixel-by-pixel to generate a refined edge strength map. To enhance the edge features, the image undergoes normalization and non-maximum suppression. This is followed by edge contour extraction using double-thresholding and a novel morphological refinement technique. Experimental results on the edge detection dataset demonstrate that the proposed method offers robust noise resistance and superior edge quality, outperforming traditional methods such as Color Sobel, Color Canny, SE, and Color AGDD, as evidenced by performance metrics including the PR curve, AUC, PSNR, MSE, and FOM. Full article

The vertical profiles of PM_2.5 chemical components are crucial for tracing pollution development, determining causes, and improving air quality. Yet, previous studies only yielded transient and sparse results due to technological limitations. Comprehensive analysis of component vertical distribution across an entire boundary layer remains challenging. Here, we provided a first-ever vertical–temporal continuous dataset of aerosol component concentrations, including sulfate (SO₄²⁻), ammonium (NH₄⁺), nitrate (NO₃⁻), organic matter (OM), and black carbon (BC), using ground-based remote sensing retrieval. The retrieved dataset showed high correlations with in situ chemical observation, with all components exceeding 0.75 and some surpassing 0.90. Using the Beijing 2022 Winter Paralympics as an example, we observed distinct vertical patterns and responses to meteorology and emissions of different components under strictly controlled conditions. During the Paralympics, the emissions contribution (51.12%) surpassed meteorology (48.88%), except SO₄²⁻ and NO₃⁻. Inorganics showed high-altitude transport features, while organics were surface-concentrated, with high-altitude inorganic(organic) concentrations 1.19(0.56) times higher than those near the surface. SO₄²⁻ peaked at 919 m and 1516 m, NH₄⁺ and NO₃⁻ showed an additional peak near 300–500 m, influenced by surface sources and secondary generation. The inorganics exhibited a transport-holding–sinking–resurging process, with NO₃⁻ reaching higher and sinking more. By contrast, organic components massified near 200 m, with a slight increase in high-altitude transport by time. The dispersion of all components driven by a north-westerly wind started 5 h earlier at high altitudes than near the surface, marking the end of the process. The insights gleaned highlight regional inorganic impacts and local organic impacts under the coupling of emission control and meteorology, thus offering helpful guidance for source attribution and targeted control policies. Full article

The increasing popularity of short videos has presented sellers with fresh opportunities to craft video advertisements that incorporate diverse modal information, with each modality potentially having a different influence on consumer engagement. Understanding which information is most important in attracting consumers can provide theoretical support to researchers. However, the dimensionality of the multimodal features of short video advertisements is often higher than the available data, posing specific difficulties in data analysis. Therefore, designing a multimodal analysis framework is needed to comprehensively extract and reduce the dimensionality of the different modal features of short video advertisements, thus analyzing which modal features are more important for consumer engagement. In this study, we chose TikTok as the research subject, and employed deep learning and machine learning techniques to extract features from short video advertisements, encompassing visual, acoustic, title, and speech text features. Subsequently, we introduced a method based on mixed-regularization sparse representation to select variables. Ultimately, we utilized multiblock partial least squares regression to regress the selected variables alongside additional scalar variables to calculate the block importance. The empirical analysis results indicate that visual and speech text features are the key factors influencing consumer engagement, providing theoretical support for subsequent research and offering practical insights for marketers. Full article

Immunotherapy has shown limited efficacy in prostate cancer, largely due to low tumor immunogenicity, sparse tumor-infiltrating lymphocytes, and a suppressive microenvironment. Recent therapeutic strategies aim to boost immune responses and counteract immunosuppressive factors through interventions such as immune checkpoint inhibitors, immunogenic cell death-inducing therapies, and the targeted blockade of pathways like that of transforming growth factor-β. Vaccine-based approaches, potent immune adjuvants, and engineered chimeric antigen receptor (CAR) T cells are also being investigated to overcome local immune inhibitory signals. Advancements in imaging, multi-omic profiling, and liquid biopsies offer promising avenues for real-time monitoring, better patient selection, and precision treatment. This review provides an overview of the key immunosuppressive features of prostate cancer, current immunotherapeutic modalities, and emerging strategies to transform “cold” tumors into more responsive “hot” targets. By integrating these approaches, we may achieve more durable clinical benefits for patients with advanced or metastatic prostate cancer. Full article

To improve the accuracy and robustness of bearing remaining useful life (RUL) prediction, this paper proposes a bearing RUL prediction method based on PELT state segmentation and time–frequency analysis, incorporating the Informer model for time-series modeling. First, the PELT (Pruned Exact Linear Time) algorithm is used to segment the vibration signals over the full life cycle of the bearing, accurately identifying critical degradation states and optimizing the stage division of the degradation process. Next, wavelet transform is applied to perform time–frequency analysis on the vibration signals, generating time–frequency spectrograms to comprehensively extract features in both the time and frequency domains. Finally, the extracted time–frequency features are used as input to predict the bearing RUL using the Informer model. As an efficient time-series prediction model, the Informer excels at handling long time series by leveraging a sparse self-attention mechanism to effectively capture the long-term dependencies in the signals. Experiments conducted on a publicly available dataset and comparisons with traditional methods demonstrate that the proposed method offers significant advantages in terms of prediction accuracy, computational efficiency, and robustness, making it more suitable for bearing health assessment and RUL prediction under complex working conditions. Full article

Background and Clinical Significance: Since the prevalence of fungal lung infections is increasing, certain agents, such as Cladosporium spp., have emerged as unexpected causes. Cladosporium spp. fungi are ubiquitous in environments such as soil, fruits, and wine corks; they are a part of the normal human skin flora; and they are known respiratory allergens. Case Presentation: A patient with a history of post-COVID-19 syndrome and AIDS presented with lung pathology indicative of an invasive fungal infection. The initial histopathological examination revealed numerous yeast-like cells with narrow-based budding, which led to a mistaken diagnosis of cryptococcosis. However, further detailed examination revealed sparse hyphae in the lung tissue, suggesting a more complex fungal infection. Molecular analyses and sequence BLAST alignment were performed, ultimately identifying the infectious agent as “Cladosporium species novum”, a rare cause of invasive pulmonary cladosporiasis. Conclusions: Invasive pulmonary cladosporiasis is a rare condition, and the morphological features of the fungus alone were insufficient to establish a correct diagnosis. A comprehensive pathohistological and molecular approach with bioinformatics tools is essential for the correct identification of rare and potentially life-threatening fungal pathogens in immunocompromised patients. Full article