Search Results (5,142)

Point cloud semantic segmentation is a key technology for applications such as autonomous driving, robotics, and virtual reality. Current approaches are heavily reliant on local relative coordinates and simplistic attention mechanisms to aggregate neighborhood information. This often leads to an ineffective joint representation of geometric perturbations and feature variations, coupled with a lack of adaptive selection for salient features during context fusion. On this basis, we propose LSSCC-Net, a novel segmentation framework based on LACV-Net. First, the spatial-feature dynamic aggregation module is designed to fuse offset information by symmetric interaction between spatial positions and feature channels, thus supplementing local structural information. Second, a dual-dimensional attention mechanism (spatial and channel) is introduced to symmetrically deploy attention modules in both the encoder and decoder, prioritizing salient information extraction. Finally, Lovász-Softmax Loss is used as an auxiliary loss to optimize the training objective. The proposed method is evaluated on two public benchmark datasets. The mIoU on the Toronto3D and S3DIS datasets is 83.6% and 65.2%, respectively. Compared with the baseline LACV-Net, LSSCC-Net showed notable improvements in challenging categories: the IoU for “road mark” and “fence” on Toronto3D increased by 3.6% and 8.1%, respectively. These results indicate that LSSCC-Net more accurately characterizes complex boundaries and fine-grained structures, enhancing segmentation capabilities for small-scale targets and category boundaries. Full article

Small-object semantic segmentation remains challenging in urban driving scenes due to limited pixel occupancy, blurred boundaries, and the constraints imposed by lightweight deployment. To address these issues, this paper presents a lightweight semantic segmentation framework that enhances boundary awareness and contextual representation while maintaining computational efficiency. The proposed method integrates an edge-supervised boundary gating module to emphasize object boundaries, an efficient multi-scale context aggregation strategy to mitigate scale variation, and a lightweight feature enhancement mechanism for effective feature fusion. Edge supervision is introduced as an auxiliary regularization signal and does not require additional manual annotations. Extensive experiments conducted on multiple benchmark datasets, including Cityscapes, CamVid, PASCAL VOC 2012, and IDDA, demonstrate that the proposed framework consistently improves segmentation performance, particularly for small-object categories, while preserving a favorable balance between accuracy and efficiency. Full article

Current approaches to dynamic sign language recognition commonly rely on dense landmark representations, which impose high computational cost and hinder real-time deployment on resource-constrained devices. To address this limitation, this work proposes a computationally efficient framework for real-time dynamic Mexican Sign Language (MSL) recognition based on a multimodal minimal angular-geometry representation. Instead of processing complete landmark sets (e.g., MediaPipe Holistic with up to 468 keypoints), the proposed method encodes the relational geometry of the hands, face, and upper body into a compact set of 28 invariant internal angular descriptors. This representation substantially reduces feature dimensionality and computational complexity while preserving linguistically relevant manual and non-manual information required for grammatical and semantic discrimination in MSL. A real-time end-to-end pipeline is developed, comprising multimodal landmark extraction, angular feature computation, and temporal modeling using a Bidirectional Long Short-Term Memory (BiLSTM) network. The system is evaluated on a custom dataset of dynamic MSL gestures acquired under controlled real-time conditions. Experimental results demonstrate that the proposed approach achieves 99% accuracy and 99% macro F1-score, matching state-of-the-art performance while using fewer features dramatically. The compactness, interpretability, and efficiency of the minimal angular descriptor make the proposed system suitable for real-time deployment on low-cost devices, contributing toward more accessible and inclusive sign language recognition technologies. Full article

This study proposes a lightweight framework for transferring pretrained land cover classification architectures without additional training. The system utilizes French IGN imagery and Korean UAV and aerial imagery. It employs FLAIR U-Net models with ResNet34 and MiTB5 backbones, along with the AI-HUB U-Net. The implementation consists of four sequential stages. First, we perform class mapping between heterogeneous schemes and unify coordinate systems. Second, a quadratic polynomial regression equation is constructed. This formula uses multispectral band statistics as hyperparameters and class-wise IoU as the dependent variable. Third, optimal parameters are identified using the stationary point condition of Response Surface Methodology (RSM). Fourth, the final land cover map is generated by fusing class-wise optimal results at the pixel level. Experimental results show that optimization is typically completed within 60 inferences. This procedure achieves IoU improvements of up to 67.86 percentage points compared to the baseline. For automated application, these optimized values from a source domain are successfully transferred to target areas. This includes transfers between high-altitude mountainous and low-lying coastal territories via proportional mapping. This capability demonstrates cross-regional and cross-platform generalization between ResNet34 and MiTB5. Statistical validation confirmed that the performance surface followed a systematic quadratic response. Adjusted R² values ranged from 0.706 to 0.999, with all p-values below 0.001. Consequently, the performance function is universally applicable across diverse geographic zones, spectral distributions, spatial resolutions, sensors, neural networks, and land cover classes. This approach achieves more than a 4000-fold reduction in computational resources compared to full model training, using only 32 to 150 tiles. Furthermore, the proposed technique demonstrates 10–74× superior resource efficiency (resource consumption per unit error reduction) over prior transfer learning schemes. Finally, this study presents a practical solution for inference and performance optimization of land cover semantic segmentation on standard commodity CPUs, while maintaining equivalent or superior IoU. Full article

To address the issues of inadequate performance and excessive interaction costs when handling large-scale and complex-shaped change areas with existing interaction forms, this paper proposes RSICDNet, an interactive change detection (ICD) model with regional scribble interaction. In this framework, regional scribble interaction is introduced for the first time to provide rich spatial prior information for accurate ICD. Specifically, RSICDNet first employs an interaction processing network to extract interactive features, and subsequently utilizes the High-Resolution Network (HRNet) backbone to extract features from bi-temporal remote sensing images concatenated along the channel dimension. To effectively integrate these two information streams, an Interaction Fusion and Refinement Module (IFRM) is proposed, which injects the spatial priors from the interactive features into the high-level semantic features. Finally, an Object Contextual Representation (OCR) module is applied to further refine feature representations, and a lightweight segmentation head is used to generate final change map. Furthermore, a human–computer ICD application has been developed based on RSICDNet, significantly enhancing its potential for practical deployment. To validate the effectiveness of the proposed RSICDNet, extensive experiments are conducted against mainstream interactive deep learning models on the WHU-CD, LEVIR-CD, and CLCD datasets. The quantitative results demonstrate that RSICDNet achieves optimal Number of Interactions (NoI) metrics across all three datasets. Specifically, its NoI80 values reach 1.15, 1.45, and 3.42 on the WHU-CD, LEVIR-CD, and CLCD datasets, respectively. The qualitative results confirm a clear advantage for RSICDNet, which consistently delivers visually superior outcomes using the same or often fewer interactions. Full article

The deployment of wireless seismic nodal systems necessitates the efficient identification of optimal locations for sensor installation, considering factors such as ground stability and the absence of interference. Semantic segmentation of satellite imagery has advanced significantly, and its application to this specific task remains unexplored. This work presents a baseline empirical evaluation of the U-Net architecture for the semantic segmentation of surfaces applicable for seismic sensor installation. We utilize a novel dataset of Sentinel-2 multispectral images, specifically labeled for this purpose. The study investigates the impact of pretrained encoders (EfficientNetB2, Cross-Stage Partial Darknet53—CSPDarknet53, and Multi-Axis Vision Transformer—MAxViT), different combinations of Sentinel-2 spectral bands (Red, Green, Blue (RGB), RGB+Near Infrared (NIR), 10-bands with 10 and 20 m/pix spatial resolution, full 13-band), and a technique for improving small object segmentation by modifying the input convolutional layer stride. Experimental results demonstrate that the CSPDarknet53 encoder generally outperforms the others (IoU = 0.534, Precision = 0.716, Recall = 0.635). The combination of RGB and Near-Infrared bands (10 m/pixel resolution) yielded the most robust performance across most configurations. Reducing the input stride from 2 to 1 proved beneficial for segmenting small linear objects like roads. The findings establish a baseline for this novel task and provide practical insights for optimizing deep learning models in the context of automated seismic nodal network installation planning. Full article

With the rise in experience economy and the popularization of digital technology, user-generated content (UGC) has become a core data source for understanding tourist needs and evaluating the service quality of venues. As a landmark venue that combines science education, interactive experience, and landscape viewing, the service quality of the Macau Science Center directly affects tourists’ travel experience and word-of-mouth dissemination. However, existing studies mostly rely on traditional questionnaire surveys and lack multi-technology collaborative analysis. In order to accurately identify the factors affecting satisfaction, this study uses 788 valid UGC data from five major platforms, namely Google Maps reviews, TripAdvisor, Sina Weibo, Xiaohongshu (Rednote), and Ctrip, from January 2023 to November 2025. It integrates word frequency analysis, semantic network analysis, latent Dirichlet allocation (LDA) topic modeling, and Valence Aware Dictionary and sEntiment Reasoner (VADER) sentiment computing to construct a systematic research framework. The study found that (1) the core attention dimensions of users cover the needs of parent–child and family visits, exhibitions and interactive experiences, ticketing and consumption services, surrounding environment and landscape, emotional evaluation, and recommendation intention. (2) The keyword association network has gradually developed from a loose network in the early stage to a comprehensive experience-dense network. (3) LDA analysis identified five main potential demand themes: comprehensive visiting experience and scenario integration, parent–child interaction and characteristic scenario experience, core venue facilities and ticketing services, visiting value and emotional evaluation, and transportation and surrounding landscapes. (4) User emotions were predominantly positive, accounting for 82.7%, while negative emotions were concentrated in local service details, and the emotional scores showed a fluctuating upward trend. This study provides targeted suggestions for the service optimization of the Macau Science Center and also provides a methodological reference for UGC-driven research in similar cultural venues. Full article

Automatic Identification System (AIS) signals are critical for maritime traffic monitoring and collision avoidance. In satellite reception environments, signal collisions occur frequently due to large coverage areas and high ship density, severely degrading decoding performance. We propose a dual-branch deep learning architecture that combines precise boundary detection with segment-level classification to address this collision problem. The network employs a multi-scale convolutional backbone that feeds two specialized branches: one detects collision boundaries with sample-level precision, while the other provides semantic context through segment classification. We developed a satellite AIS dataset generation framework that simulates realistic collision scenarios including multiple ships, Doppler effects, and channel impairments. The trained model achieves 96% collision detection accuracy on simulated data. Validation on real satellite recordings demonstrates that our method retains 99.4% of valid position reports compared to direct decoding of the original signal. Controlled experiments show that intelligent collision removal outperforms random segment exclusion by 6.4 percentage points, confirming the effectiveness of our approach. Full article

Optical imagery is often contaminated by clouds to varying degrees, which greatly affects the interpretation and analysis of images. Synthetic Aperture Radar (SAR) possesses the characteristic of penetrating clouds and mist, and a common strategy in SAR-assisted cloud removal involves fusing SAR and optical data and leveraging deep learning networks to reconstruct cloud-free optical imagery. However, these methods do not fully consider the characteristics of the frequency domain when processing feature integration, resulting in blurred edges of the generated cloudless optical images. Therefore, an adaptive frequency domain feature reconstruction-based cloud removal method is proposed to solve the problem. The proposed method comprises four key sequential stages. First, shallow features are extracted by fusing optical and SAR images. Second, a Transformer-based encoder captures multi-scale semantic features. Subsequently, the Frequency Domain Decoupling Module (FDDM) is employed. Utilizing a Dynamic Mask Generation mechanism, it explicitly decomposes features into low-frequency structures and high-frequency details, effectively suppressing cloud interference while preserving surface textures. Finally, robust information interaction is facilitated by the Cross-Frequency Reconstruction Module (CFRM) via transposed cross-attention, ensuring precise fusion and reconstruction. Experimental evaluation on the M3R-CR dataset confirms that the proposed approach achieves the best results on all four evaluated metrics, surpassing the performance of the eight other State-of-the-Art methods. It has demonstrated its effectiveness and advanced capabilities in the task of SAR-optical fusion for cloud removal. Full article

The advent of sixth-generation (6G) imposes stringent demands on wireless networks, while traditional 2D rigid reconfigurable intelligent surfaces (RISs) face bottlenecks in regulatory freedom and scenario adaptability. To address this, stacked intelligent metasurfaces (SIMs) have emerged. This paper presents a systematic review of SIM technology. It first elaborates on the SIM multi-layer stacked architecture and wave-domain signal-processing principles, which overcome the spatial constraints of conventional RISs. Then, it analyzes challenges, including beamforming and channel estimation for SIM, and explores its application prospects in key 6G scenarios such as integrated sensing and communication (ISAC), low earth orbit (LEO) satellite communication, semantic communication, and UAV communication, as well as future trends like integration with machine learning and nonlinear devices. Finally, it summarizes the open challenges in low-complexity design, modeling and optimization, and performance evaluation, aiming to provide insights to promote the large-scale adoption of SIM in next-generation wireless communications. Full article

Oriented object detection in remote sensing images remains a challenging task due to arbitrary target rotations, extreme scale variations, and complex backgrounds. However, current rotated detectors still face several limitations: insufficient orientation-sensitive feature representation, feature misalignment for rotated proposals, and unstable optimization of rotation parameters. To address these issues, this paper proposes an enhanced Rotation-Sensitive Feature Pyramid Network (RSFPN) framework. Building upon the effective Oriented R-CNN paradigm, we introduce three novel core components: (1) a Dynamic Adaptive Feature Pyramid Network (DAFPN) that enables bidirectional multi-scale feature fusion through semantic-guided upsampling and structure-enhanced downsampling paths; (2) an Angle-Aware Collaborative Attention (AACA) module that incorporates orientation priors to guide feature refinement; (3) a Geometrically Consistent Multi-Task Loss (GC-MTL) that unifies the regression of rotation parameters with periodic smoothing and adaptive weight mechanisms. Comprehensive experiments on the DOTA-v1.0 and HRSC2016 benchmarks show that our RSFPN achieves superior performance. It attains a state-of-the-art mAP of 77.42% on DOTA-v1.0 and 91.85% on HRSC2016, while maintaining efficient inference at 14.5 FPS, demonstrating a favorable accuracy-efficiency trade-off. Visual analysis confirms that our method produces concentrated, rotation-aware feature responses and effectively suppresses background interference. The proposed approach provides a robust solution for detecting multi-oriented objects in high-resolution remote sensing imagery, with significant practical value for urban planning, environmental monitoring, and security applications. Full article

Real-timenetwork protocol data are indispensable for network security analysis. However, the rapid evolution of protocol standards poses significant challenges to automated parsing and dynamic extensibility. While artificial intelligence (AI) techniques offer potential solutions, they often introduce semantic ambiguities and inconsistent results, thereby undermining parsing precision. To overcome these limitations, we propose PMDL (Protocol Model Description Language), a general-purpose protocol description language. PMDL abstracts protocols into structured sets of fields and attributes, enabling precise and unambiguous specification of protocol syntax and semantics. Based on PMDL descriptions, our execution engine dynamically instantiates and loads protocol templates on the fly, achieving accurate, automated, and dynamically extensible parsing of network traffic. We evaluate PMDL against representative tools such as Wireshark and Kelai, as well as approaches such as Nail and BIND. Experimental results demonstrate that PMDL provides concise yet expressive protocol specifications, and the execution engine achieves superior parsing throughput. Furthermore, performance evaluation using real-world HTTP, MySQL, and DNS traffic from a campus network confirms that our system robustly meets the throughput requirements of large-scale security analysis. Full article

The high computational demand of dense voxel representations severely limits current vision-centric 3D semantic occupancy prediction methods, despite their capacity for granular scene understanding. This challenge is particularly acute in safety-critical applications like autonomous driving, where accurately perceiving dynamic instances often takes precedence over capturing the static background. This paper challenges the paradigm of dense prediction for such instance-focused tasks. We introduce the LSON-IP, a framework that strategically avoids the computational expense of dense 3D grids. LSON-IP operates on a sparse set of 3D instance queries, which are initialized directly from multi-view 2D images. These queries are then refined by our novel Sparse Instance Aggregator (SIA), an attention-based module. The SIA incorporates rich multi-view features while simultaneously modeling inter-query relationships to construct coherent object representations. Furthermore, to obviate the need for costly 3D annotations, we pioneer a Differentiable Sparse Rendering (DSR) technique. DSR innovatively defines a continuous field from the sparse voxel output, establishing a differentiable bridge between our sparse 3D representation and 2D supervision signals through volume rendering. Extensive experiments on major autonomous driving benchmarks, including SemanticKITTI and nuScenes, validate our approach. LSON-IP achieves strong performance on key dynamic instance categories and competitive overall semantic completion, all while reducing computational overhead by over 60% compared to dense baselines. Our work thus paves the way for efficient, high-fidelity instance-aware 3D perception. Full article

The use of resource-constrained Low-Power and Lossy Networks (LLNs), where the IPv6 Routing Protocol for LLNs (RPL) is the de facto routing standard, has increased due to the Internet of Things’ (IoT) explosive growth. Because of the dynamic nature of IoT deployments and the lack of in-protocol security, RPL is still quite susceptible to routing-layer attacks like Blackhole, Lowered Rank, version number manipulation, and Flooding despite its lightweight architecture. Lightweight, data-driven intrusion detection methods are necessary since traditional cryptographic countermeasures are frequently unfeasible for LLNs. However, the lack of RPL-specific control-plane semantics in current cybersecurity datasets restricts the use of machine learning (ML) for practical anomaly identification. In order to close this gap, this work models both static and mobile networks under benign and adversarial settings by creating a novel, large-scale multiclass RPL attack dataset using Contiki-NG’s Cooja simulator. To record detailed packet-level and control-plane activity including DODAG Information Object (DIO), DODAG Information Solicitation (DIS), and Destination Advertisement Object (DAO) message statistics along with forwarding and dropping patterns and objective-function fluctuations, a protocol-aware feature extraction pipeline is developed. This dataset is used to evaluate fifteen classifiers, including Logistic Regression (LR), Support Vector Machine (SVM), Decision Tree (DT), k-Nearest Neighbors (KNN), Random Forest (RF), Extra Trees (ET), Gradient Boosting (GB), AdaBoost (AB), and XGBoost (XGB) and several ensemble strategies like soft/hard voting, stacking, and bagging, as part of a comprehensive ML-based detection system. Numerous tests show that ensemble approaches offer better generalization and prediction performance. With overfitting gaps less than 0.006 and low cross-validation variance, the Soft Voting Classifier obtains the greatest accuracy of 99.47%, closely followed by XGBoost with 99.45% and Random Forest with 99.44%. Full article

Weed competition represents a primary constraint in sesame production, causing substantial yield losses typically ranging from 18 to 68% under inadequate control measures. Precise crop–weed discrimination remains challenging due to morphological similarities, complex field conditions, and vegetation overlapping. To address these issues, we developed Context-Aware Feature-Enhanced Detection Transformer (CAFE-DETR), an enhanced Real-Time Detection Transformer (RT-DETR) architecture optimized for sesame–weed identification. First, the C2f with a Unified Attention-Gating (C2f-UAG) module integrates unified head attention with convolutional gating mechanisms to enhance morphological discrimination capabilities. Second, the Hierarchical Context-Adaptive Fusion Network (HCAF-Net) incorporates hierarchical context extraction and spatial–channel enhancement to achieve multi-scale feature representation. Furthermore, the Polarized Linear Spatial Multi-scale Fusion Network (PLSM-Encoder) reduces computational complexity from O(N²) to O(N) through polarized linear attention while maintaining global semantic modeling. Additionally, the Focaler-MPDIoU loss function improves localization accuracy through point distance constraints and adaptive sample focusing. Experimental results on the sesame–weed dataset demonstrate that CAFE-DETR achieves 90.0% precision, 89.5% mAP₅₀, and 59.5% mAP_50–95, representing improvements of 13.07%, 4.92%, and 2.06% above the baseline RT-DETR, respectively, while reducing computational cost by 23.73% (43.4 GFLOPs) and parameter count by 10.55% (17.8 M). These results suggest that CAFE-DETR is a viable alternative for implementation in intelligent spraying systems and precision agriculture platforms. Notably, this study lacks external validation, cross-dataset testing, and field trials, which limits the generalizability of the model to diverse real-world agricultural scenarios. Full article