Search Results (14,003)

High-quality image restoration (IR) is a fundamental task in computer vision, aiming to recover a clear image from its degraded version. Prevailing methods typically employ a static inference pipeline, neglecting the spatial variability of image content and degradation, which makes it difficult for them to adaptively handle complex and diverse restoration scenarios. To address this issue, we propose a novel adaptive image restoration framework named Hybrid Convolutional Transformer with Dynamic Prompting (HCTDP). Our approach introduces two key architectural innovations: a Spatially Aware Dynamic Prompt Head Attention (SADPHA) module, which performs fine-grained local restoration by generating spatially variant prompts through real-time analysis of image content and a Gated Skip-Connection (GSC) module that refines multi-scale feature flow using efficient channel attention. To guide the network in generating more visually plausible results, the framework is optimized with a hybrid objective function that combines a pixel-wise L1 loss and a feature-level perceptual loss. Extensive experiments on multiple public benchmarks, including image deraining, dehazing, and denoising, demonstrate that our proposed HCTDP exhibits superior performance in both quantitative and qualitative evaluations, validating the effectiveness of the adaptive restoration framework while utilizing fewer parameters than key competitors. Full article

This work reports the performance evaluation of a newly developed small-animal positron emission tomography (PET) system based on lutetium-yttrium oxyorthosilicate (LYSO) crystals and multi-pixel photon counter (MPPC). Performance was evaluated, including spatial resolution, system sensitivity, energy resolution, scatter fraction (SF), noise–equivalent count rate (NECR), micro-Derenzo phantom imaging, and in vivo imaging of mice and rats. The system achieved a tangential spatial resolution of 0.9 mm in the axial direction at a quarter axial offset using the three-dimensional ordered-subsets expectation maximization (3D OSEM) reconstruction algorithm. The peak sensitivity was 8.74% within a 200–750 keV energy window, with an average energy resolution of 12.5%. Scatter fractions were 12.9% and 30.0% for mouse- and rat-like phantoms, respectively. The NECR reached 878.7 kcps at 57.6 MBq for the mouse phantom and 421.4 kcps at 63.2 MBq for the rat phantom. High-resolution phantom and in vivo images confirmed the system’s capability for quantitative, high-sensitivity small-animal imaging, demonstrating its potential for preclinical molecular imaging studies. Full article

Nondestructive quality detection of characteristic fruits is essential for ensuring nutritional value, economic viability, and consumer safety in global supply chains, yet traditional destructive methods compromise sample integrity and scalability. Visible and near-infrared (Vis/NIR) spectroscopy offers a transformative solution by enabling rapid, non-invasive multi-attribute quantification through molecular overtone vibrations. This review examines recent advancements in Vis/NIR-based fruit quality detection, encompassing fundamental principles, system configurations, and detection strategies calibrated to fruit biophysical properties. Firstly, optical mechanisms and system architectures (portable, online, vehicle-mounted) are compared, emphasizing their compatibility with fruit structural complexity. Then, critical challenges arising from fruit-specific characteristics—such as rind thickness, pit interference, and spatial heterogeneity—are analyzed, highlighting their impact on spectral accuracy. Applications across diverse fruit categories (pitted, thin-rinded, and thick-rinded) are systematically reviewed, with case studies demonstrating the robust prediction of key quality indices. Subsequently, considerations in model development and validation are presented. Finally, persistent limitations in model transferability and environmental adaptability are discussed, proposing future research directions centered on integrating hyperspectral imaging, AI-driven calibration transfer, standardized spectral databases, and miniaturized, field-deployable sensors. Collectively, these methodological breakthroughs will pave the way for autonomous, next-generation quality assessment platforms, revolutionizing postharvest management for characteristic fruits. Full article

Precise online measurement of large structural components is urgently needed in modern manufacturing and intelligent construction, requiring a measurement range over 1 m, near-millimeter accuracy, second-level measurement speed, and adaptability to complex environments. In this paper, three mainstream measurement technologies, namely the image method, line laser scanning method, and structured light method, are comparatively analyzed. The structured light method exhibits remarkable comprehensive advantages in terms of accuracy and speed; however, it suffers from the issue of occlusion during contour measurement. To tackle this problem, multi-camera stitching is employed, wherein the accuracy of camera calibration plays a crucial role in determining the quality of point cloud stitching. Focusing on the cable tightening scenario of meter-diameter cables in cable-stayed bridges, this study develops a contour measurement system based on the collaboration of multiple structured light cameras. Measurement indicators are optimized through modeling analysis, system construction, and performance verification. During verification, four structured light scanners were adopted, and measurements were repeated 11 times for the test workpieces. Experimental results demonstrate that although the current measurement errors have not yet been stably controlled within the millimeter level, this research provides technical exploration and practical experience for high-precision measurement in the field of intelligent construction, thus laying a solid foundation for subsequent accuracy improvement. Full article

Due to the inherent characteristics of synthetic aperture radar (SAR) imaging, variations in target orientation, and the challenges posed by non-cooperative targets (i.e., targets without cooperative transponders or external markers), limited viewpoint coverage results in a small-sample problem that severely constrains the application of deep learning to SAR image interpretation and target recognition. To address this issue, this paper proposes a multi-target, multi-view SAR image generation method based on conditional information and StyleGAN2, designed to generate high-quality, angle-controllable SAR images of typical targets from limited samples. The proposed framework consists of an angle encoder, a generator, and a discriminator. The angle encoder employs a sinusoidal encoding scheme that combines sine and cosine functions to address the discontinuity inherent in one-hot angle encoding, thereby enabling precise angle control. Moreover, the integration of SimAM and IAAM attention mechanisms enhances image quality, facilitates accurate angle control, and improves the network’s generalization to untrained angles. Experiments conducted on a self-constructed dataset of typical civilian targets and the SAMPLE subset of the MSTAR dataset demonstrate that the proposed method outperforms existing baselines in terms of structural fidelity and feature distribution consistency. The generated images achieve a minimum FID of 6.541 and a maximum MS-SSIM of 0.907, while target recognition accuracy improves by 6.03% and 7.14%, respectively. These results validate the feasibility and effectiveness of the proposed approach for SAR image generation and target recognition tasks. Full article

Modern semantic segmentation models require extensive pixel-level annotations, creating a significant barrier to practical deployment as labeling a single image can take hours of human effort. Active learning offers a promising way to reduce annotation costs through intelligent sample selection. However, existing methods rely on poorly calibrated confidence estimates, making uncertainty quantification unreliable. We introduce Conformal Risk Controlled Active Learning (CRC-AL), a novel framework that provides statistical guarantees on uncertainty quantification for semantic segmentation, in contrast to heuristic approaches. CRC-AL calibrates class-specific thresholds via conformal risk control, transforming softmax outputs into multi-class prediction sets with formal guarantees. From these sets, our approach derives complementary uncertainty representations: risk maps highlighting uncertain regions and class co-occurrence embeddings capturing semantic confusions. A physics-inspired selection algorithm leverages these representations with a barycenter-based distance metric that balances uncertainty and diversity. Experiments on Cityscapes and PascalVOC2012 show CRC-AL consistently outperforms baseline methods, achieving 95% of fully supervised performance with only 30% of labeled data, making semantic segmentation more practical under limited annotation budgets. Full article

Transmission line icing poses a significant natural disaster threat to power grid security. This paper systematically reviews recent advances in the understanding of icing mechanisms, intelligent detection, and prevention technologies, while providing perspectives on future development directions. In mechanistic research, although a multi-physics coupling framework has been established, characterization of dynamic evolution over complex terrain and coupled physical mechanisms remains inadequate. Detection technology is undergoing a paradigm shift from traditional contact measurements to non-contact intelligent perception. Visual systems based on UAVs and fixed platforms have achieved breakthroughs in ice recognition and thickness retrieval, yet their performance remains constrained by image quality, data scale, and edge computing capabilities. Anti-/de-icing technologies have evolved into an integrated system combining active intervention and passive defense: DC de-icing (particularly MMC-based topologies) has become the mainstream active solution for high-voltage lines due to its high efficiency and low energy consumption; superhydrophobic coatings, photothermal functional coatings, and expanded-diameter conductors show promising potential but face challenges in durability, environmental adaptability, and costs. Future development relies on the deep integration of mechanistic research, intelligent perception, and active prevention technologies. Through multidisciplinary innovation, key technologies such as digital twins, photo-electro-thermal collaborative response systems, and intelligent self-healing materials will be advanced, with the ultimate goal of comprehensively enhancing power grid resilience under extreme climate conditions. Full article

High-definition maps (HDMaps) are critical for safe autonomy on structured roads. Yet traditional production—relying on dedicated mapping fleets and manual quality control—is costly and slow, impeding large-scale, frequent updates. Recently, standard-definition maps (SDMaps) derived from remote sensing have been adopted as priors to support HDMap perception, lowering cost but struggling with subtle urban changes and localization drift. We propose Loop-MapNet, a self-evolving, multimodal, closed-loop mapping framework. Loop-MapNet effectively leverages surround-view images, LiDAR point clouds, and SDMaps; it fuses multi-scale vision via a weighted BiFPN, and couples PointPillars BEV and SDMap topology encoders for cross-modal sensing. A Transformer-based bidirectional adaptive cross-attention aligns SDMap with online perception, enabling robust fusion under heterogeneity. We further introduce a confidence-guided masked autoencoder (CG-MAE) that leverages confidence and probabilistic distillation to both capture implicit SDMap priors and enhance the detailed representation of low-confidence HDMap regions. With spatiotemporal consistency checks, Loop-MapNet incrementally updates SDMaps to form a perception–mapping–update loop, compensating remote-sensing latency and enabling online map optimization. On nuScenes, within 120 m, Loop-MapNet attains 61.05% mIoU, surpassing the best baseline by 0.77%. Under extreme localization errors, it maintains 60.46% mIoU, improving robustness by 2.77%; CG-MAE pre-training raises accuracy in low-confidence regions by 1.72%. These results demonstrate advantages in fusion and robustness, moving beyond one-way prior injection and enabling HDMap–SDMap co-evolution for closed-loop autonomy and rapid SDMap refresh from remote sensing. Full article

Diabetic Macular Edema (DME) poses a significant threat to vision, often leading to permanent blindness if not detected and addressed swiftly. Existing manual diagnostic methods are arduous and inconsistent, highlighting the pressing necessity for automated, accurate, and personalized solutions. This study presents a novel methodology for diagnosing DME and categorizing choroidal neovascularization (CNV), drusen, and normal conditions from fundus images through the application of transfer learning models and bio-inspired optimization methodologies. The methodology utilizes advanced transfer learning architectures, including VGG16, VGG19, ResNet50, EfficientNetB7, EfficientNetV2-S, InceptionV3, and InceptionResNetV2, for analyzing both binary and multi-class Optical Coherence Tomography (OCT) datasets. We combined the OCT datasets OCT2017 and OCTC8 to create a new dataset for our study. The parameters, including learning rate, batch size, and dropout layer of the fully connected network, are further adjusted using the bio-inspired Particle Swarm Optimization (PSO) method, in conjunction with thorough preprocessing. Explainable AI approaches, especially Shapley additive explanations (SHAP), provide transparent insights into the model’s decision-making processes. Experimental findings demonstrate that our bio-inspired optimized transfer learning Inception V3 significantly surpasses conventional deep learning techniques for DME classification, as evidenced by enhanced metrics including the accuracy, precision, recall, F1-score, misclassification rate, Matthew’s correlation coefficient, intersection over union, and kappa coefficient for both binary and multi-class scenarios. The accuracy achieved is approximately 98% in binary classification and roughly 90% in multi-class classification with the Inception V3 model. The integration of contemporary transfer learning architectures with nature-inspired PSO enhances diagnostic precision to approximately 95% in multi-class classification, while also improving interpretability and reliability, which are crucial for clinical implementation. This research promotes the advancement of more precise, personalized, and timely diagnostic and therapeutic strategies for Diabetic Macular Edema, aiming to avert vision loss and improve patient outcomes. Full article

The benefit of neoadjuvant chemotherapy (NAC) over upfront surgery (UFS) for resectable pancreatic ductal adenocarcinoma (PDAC) is increasingly recognized, yet prognostic biomarkers remain undefined. We evaluated tumor–stroma ratio (TSR), β-catenin (β-CTN) expression, and tumor budding (TB) in 84 resected PDACs (35 NAC, 49 UFS) using digital image analysis of multi-cytokeratin (m-CK) and β-CTN immunohistochemistry. TSR was defined as the proportion of malignant epithelial area within the tumor, and the β-CTN/m-CK index as the ratio of β-CTN to m-CK immunoreactivity in tumor tissue relative to intralobular ducts. TB was significantly less frequent in NAC than UFS (p = 0.003), suggesting that NAC may indirectly modulate epithelial–mesenchymal transition, with TB regarded as its morphological correlate. In the NAC cohort, low TSR was associated with more favorable histological response (Evans IIa/IIb, median 7%; Evans I, 16%; p = 0.009), likely reflecting NAC-induced tumor shrinkage with relative stromal predominance. In multivariable analysis, low β-CTN/m-CK index (<0.5) predicted shorter relapse-free survival in both NAC (HR = 2.516, p = 0.043) and UFS (HR = 2.230, p = 0.025) subgroups. High TSR (≥13%) was associated with shorter cancer-specific survival (HR = 2.414, p = 0.034) in the overall cohort, indicating prognostic value complementing its association with NAC response. These results identify the β-CTN/m-CK index and TSR as prognostic biomarkers in PDAC. Full article

Accurate depth image inpainting in complex urban environments remains a critical challenge due to occlusions, reflections, and sensor limitations, which often result in significant data loss. We propose a hybrid deep learning framework that explicitly combines local and global modelling through Convolutional Neural Networks (CNNs) and Transformer modules. The model employs a multi-branch parallel architecture, where the CNN branch captures fine-grained local textures and edges, while the Transformer branch models global semantic structures and long-range dependencies. We introduce an optimized attention mechanism, Agent Attention, which differs from existing efficient/linear attention methods by using learnable proxy tokens tailored for urban scene categories (e.g., façades, sky, ground). A content-guided dynamic fusion module adaptively combines multi-scale features to enhance structural alignment and texture recovery. The frame-work is trained with a composite loss function incorporating pixel accuracy, perceptual similarity, adversarial realism, and structural consistency. Extensive experiments on the Paris StreetView dataset demonstrate that the proposed method achieves state-of-the-art performance, outperforming existing approaches in PSNR, SSIM, and LPIPS metrics. The study highlights the potential of multi-scale modeling for urban depth inpainting and discusses challenges in real-world deployment, ethical considerations, and future directions for multimodal integration. Full article

Current drone navigation systems primarily rely on Global Navigation Satellite Systems (GNSSs), but their signals are susceptible to interference, spoofing, or suppression in complex environments, leading to degraded positioning performance or even failure. To enhance the positioning accuracy and robustness of low-altitude drones in satellite-denied environments, this paper investigates an absolute visual localization solution. This method achieves precise localization by matching real-time images with reference images that have absolute position information. To address the issue of insufficient feature generalization capability due to the complex and variable nature of ground scenes, a visual-based image retrieval algorithm is proposed, which utilizes a fusion of shallow spatial features and deep semantic features, combined with generalized average pooling to enhance feature representation capabilities. To tackle the registration errors caused by differences in perspective and scale between images, an image registration algorithm based on cyclic consistency matching is designed, incorporating a reprojection error loss function, a multi-scale feature fusion mechanism, and a structural reparameterization strategy to improve matching accuracy and inference efficiency. Based on the above methods, a hierarchical absolute visual localization system is constructed, achieving coarse localization through image retrieval and fine localization through image registration, while also integrating IMU prior correction and a sliding window update strategy to mitigate the effects of scale and rotation differences. The system is implemented on the ROS platform and experimentally validated in a real-world environment. The results show that the localization success rates for the h, s, v, and w trajectories are 95.02%, 64.50%, 64.84%, and 91.09%, respectively. Compared to similar algorithms, it demonstrates higher accuracy and better adaptability to complex scenarios. These results indicate that the proposed technology can achieve high-precision and robust absolute visual localization without the need for initial conditions, highlighting its potential for application in GNSS-denied environments. Full article

With the development of infrared imaging technology and the integration of intelligent algorithms, the realization of non-contact, dynamic and real-time detection of SF6 gas leakage based on infrared video has been a significant research direction. However, the existing real-time detection algorithms exhibit low accuracy in detecting SF6 leakage and are susceptible to noise, which makes it difficult to meet the actual needs of engineering. To address this problem, this paper proposes a real-time SF6 leakage detection method, VGEC-Net, based on multi-channel fusion and spatiotemporal feature extraction. The proposed method first employs the ViBe-GMM algorithm to extract foreground masks, which are then fused with infrared images to construct a dual-channel input. In the backbone network, a CE-Net structure—integrating CBAM and ECA-Net—is combined with the P3D network to achieve efficient spatiotemporal feature extraction. A Feature Pyramid Network (FPN) and a temporal Transformer module are further integrated to enhance multi-scale feature representation and temporal modeling, thereby significantly improving the detection performance for small-scale targets. Experimental results demonstrate that VGEC-Net achieves a mean average precision (mAP) of 61.7% on the dataset used in this study, with a mAP@50 of 87.3%, which represents a significant improvement over existing methods. These results validate the effectiveness and advancement of the proposed method for infrared video-based gas leakage detection. Furthermore, the model achieves 78.2 frames per second (FPS) during inference, demonstrating good real-time processing capability while maintaining high detection accuracy, exhibiting strong application potential. Full article

Al-Zn-Mg-Cu alloys are widely used as heat-treatable ultra-high-strength materials in aerospace structural applications. While conventional single-stage aging enables high strength, advanced performance demands call for precise microstructural control via multi-stage aging. In this study, we employ a combination of scanning transmission electron microscopy (STEM), energy-dispersive X-ray spectroscopy (EDS), and X-ray diffraction (XRD) to investigate the microstructural evolution and its correlation with mechanical properties of AA7150 aluminum alloy subjected to two-step aging treatments, following a 6 h pre-aging at 120 °C. Through atomic-scale STEM imaging along the [110]_Al zone axis, we systematically characterize the precipitation behavior of GPII zones, η′ phases, and equilibrium η phases both within the grains and at grain boundaries under varying secondary aging (SA) conditions. Our results reveal that increasing the SA temperature from 140 °C to 180 °C leads to coarsening and reduced number density of intragranular precipitates, while promoting the continuous and coarse precipitation of η phases along grain boundaries, accompanied by a widening of the precipitation-free zone (PFZ). Notably, SA at 160 °C induces the formation of fine, uniformly dispersed nanoscale η′ precipitates in the alloy, as confirmed by XRD phase analysis. Aging at this temperature markedly enhances the mechanical properties, achieving an ultimate tensile strength (UTS) of 613 MPa and a yield strength (YS) of 598 MPa, while presenting an exceptionally broad peak-aging plateau. Owing to this feature, a moderate extension of the SA duration does not reduce strength and can further improve ductility, increasing the elongation (EL) to 14.26%. These results demonstrate a novel two-step heat-treatment strategy that simultaneously achieves ultra-high strength and excellent ductility, highlighting the critical role of advanced electron microscopy in elucidating phase-transformation pathways that inform microstructure-guided alloy design and processing. Full article

Due to their technical complexity, multi-step procedures, and low efficiency, traditional monitoring techniques have struggled to meet the rapid development of safety management on construction sites in assessing the safety of cuplok scaffolding. Therefore, this study applied image recognition technology to the safety monitoring of cuplok scaffold systems. A recognition model for identifying member shapes in images of cuplok scaffolds was proposed. Combined with a judgment criterion established based on the energy method, the safety state of the scaffold system was evaluated, ultimately forming an image recognition-based technique for detecting the safety performance of cuplok scaffolds. Experimental studies on a reduced-scale model demonstrated that the proposed method achieved an accuracy and efficiency of 80% in both recognition and judgment. The results indicated that this method enables rapid and efficient safety performance monitoring of cuplok scaffolding, holding significant practical implications for improving monitoring efficiency. Full article