Search Results (4,637)

Lane detection is a crucial perception task for autonomous driving, but existing methods often struggle with spatial information loss, feature upsampling artifacts, and prediction discontinuities under complex scenarios such as occlusions or poor lighting. To address these limitations, this paper proposes ACL-Net, an end-to-end lane detection network integrating attention mechanisms and context enhancement based on the Cross Layer Refinement Network framework. First, a coordinate attention module is embedded at the output of the backbone network to recalibrate spatial position information and mitigate depth-induced detail loss. Second, the feature pyramid network is reconstructed utilizing a dynamic upsampling operator and an additional bottom-up pathway to prevent edge distortion and preserve fine-grained geometric features. Finally, a lane-aware atrous spatial pyramid pooling module with asymmetric convolutions is designed to aggregate multi-scale global context, effectively reconnecting fragmented lane lines caused by visual occlusions. Extensive experiments on the TuSimple and CULane datasets demonstrate the superiority of the proposed approach. ACL-Net achieves an accuracy of 96.98% on TuSimple and a total F1-measure of 80.34% on CULane, outperforming the baseline Cross Layer Refinement Network while maintaining a real-time inference speed of 61.90 FPS. The results indicate that ACL-Net significantly improves the utilization of geometric features and exhibits enhanced robustness in challenging road conditions, including severe occlusions, nighttime, and large-curvature curves. Full article

This study analyzes domestic tourism mobility in Spain using aggregated and anonymized mobile phone data, with a particular focus on the outbound market of the municipality of Madrid and its territorial redistribution between September 2019 and September 2025. Using experimental statistics from the National Institute of Statistics (INE), a monthly series of origin–destination flows to all Spanish municipalities was constructed, harmonizing the municipal database and incorporating intensive indicators to improve inter-territorial comparability. The spatiotemporal dynamics were integrated into a Space–Time Cube (monthly resolution), and Emerging Hot Spot Analysis (EHSA) was applied to classify the persistence, intensification, or attenuation of high- and low-intensity clusters. Additionally, the grouping of time series allowed for the identification of seasonal patterns associated with coastal, urban, and nearby inland destinations. The results show: (i) a synchronous disruption in the spring of 2020 linked to COVID-19; (ii) a staggered recovery beginning in 2021, consolidating in 2023–2025; and (iii) a dual structural pattern, with a strong concentration of volumes in large urban and coastal hubs, along with high relative intensities in small municipalities in the ring surrounding Madrid. EHSA identifies intensifying hotspots in established coastal systems (Costa del Sol and Costa Blanca) and cooling or attenuated dynamics in parts of the inland region, consistent with the reconfiguration of the “tourism radius” following the pandemic. Limitations arising from statistical confidentiality and the representativeness of the source are discussed, and future research directions are proposed based on the integration of the information with expenditure and transportation data and on spatiotemporal modeling to support destination planning and management. Full article

Attention Deficit Hyperactivity Disorder (ADHD) is a heterogeneous neurodevelopmental disorder whose diagnosis is mainly based on behavioral assessment and is often delayed due to clinical complexity and limited availability of specialists. Resting-state functional magnetic resonance imaging (rs-fMRI) provides a valuable source of information for supporting automated and objective diagnosis. However, existing studies often do not fully capture the complex interactions of functional connectivity between different brain regions. To address this limitation, this work proposes a graph-based deep learning framework for ADHD classification from rs-fMRI that combines functional connectivity modeling with graph representation learning. The approach used Phase-Locking Value (PLV)-based connectivity estimation and Graph Sample and Aggregate (GraphSAGE) to jointly capture regional brain activity and inter-regional interactions in a scalable and efficient manner. GraphSAGE improves robustness to noise and inter-subject variability by aggregating information from stable local graph neighborhoods. This integration allows the model to learn discriminative connectivity-aware representations while remaining robust to signal variability and adaptable to multi-site data. The proposed framework was evaluated on the publicly available ADHD-200 dataset across multiple acquisition sites as well as on a combined multi-site dataset. The results indicate consistent performance across individual sites and on the combined dataset. The model achieved an Accuracy of 0.89, an AUC of 0.96, and a Specificity of 0.96 on the combined dataset, outperforming several existing methods in this setting. By integrating PLV-based connectivity with GraphSAGE learning, the approach provides an effective and scalable solution for automated ADHD classification from rs-fMRI data, contributing to data-driven approaches for the analysis of neurodevelopmental disorders. Full article

Underwater object detection is important for ocean exploration and marine applications. However, underwater images are often degraded by absorption, scattering, and background interference, which weaken object contours, blur boundaries, and obscure fine texture details, thereby increasing the difficulty of detecting small objects and objects with large shape variations. To address these challenges, we propose WEC-UOD, an underwater object detector that improves structure-sensitive representation learning and multi-scale feature fusion within the detector, without relying on a separate image enhancement stage. In the backbone, the Wavelet–Edge Collaboration (WEC) module first uses wavelet-subband guidance to compensate for degraded structural and texture information and then applies edge-guided spatial correction to refine object boundaries and local geometry. In the neck, the Scale-Selective Fusion (SSF) module adaptively selects informative responses from branches with different receptive fields and further suppresses background interference through channel and spatial recalibration. Experiments on RUOD and DUO show that WEC-UOD achieves mAP@0.5 scores of 87.4% and 86.9%, respectively, consistently outperforming the YOLOv11s baseline. These results demonstrate the effectiveness of combining structural enhancement with selective multi-scale aggregation for underwater object detection. Full article

Background/objectives: This study presents a mechanistic assessment of an anti-HSPG2 monoclonal antibody (AM6) as an antibody–drug conjugate (ADC) carrier in vitro. Methods: Using live-cell confocal imaging with pathway inhibitors, we qualitatively characterized AM6 internalization and trafficking and compared linker/payload configurations for intracellular delivery and in vitro cytotoxicity. Results: AM6 exhibited rapid cellular entry in MDA-MB-231-LM2 cells, with contributions from clathrin-mediated endocytosis and macropinocytosis, followed by accumulation in endo-lysosomal compartments. Consistent with these trafficking observations, AM6 ADCs bearing cleavable linkers and a potent payload (MMAE) produced more pronounced antiproliferative effects in MDA-MB-231-LM2 and other HSPG2-positive tumor cells than non-cleavable constructs, whereas doxorubicin-based ADCs showed limited activity and greater aggregation risk. Conclusions: Overall, the data inform linker/payload selection and highlight considerations for future work, including quantitative internalization, antigen-negative or knockdown controls, and in vivo pharmacology. Full article

Smart building services are increasingly realised through combinations of sensors, actuators, communication infrastructures, software platforms, analytics, and artificial intelligence-based functions. These configurations enable adaptive control, real-time monitoring, contextual automation, predictive support, user interaction, and cross-domain coordination across heating, ventilation, air conditioning, lighting, energy management, security and access control, water management, and user-centric comfort services. At the same time, the European Union Smart Readiness Indicator provides a formal basis for assessing building smartness through technical domains, service functionalities, and multidimensional impact criteria. A systematic basis for translating real-world descriptions of smart building services and their enabling technology stacks into Smart Readiness Indicator-aligned assessment inputs remains underdeveloped. A PRISMA ScR informed review was conducted to identify principal smart building service domains, synthesise their core functionalities, and reconstruct the digital technologies through which these functionalities are realised. The synthesis shows that heating, ventilation, and air conditioning and lighting provide comparatively direct translation pathways to formal Smart Readiness Indicator domains, while energy management operates mainly as a supervisory and cross-domain layer. Security and access control, water management, and several user-centric services contribute meaningfully to building smartness but often show partial or extended formal correspondence. Monitoring and control emerge as a central cross-cutting layer because many higher-order smart building capabilities are expressed through visibility, supervision, orchestration, and digital representation. Building on this review, a methodological framework is established for translating smart building services into Smart Readiness Indicator-aligned assessments. The procedure uses the smart building service instance as the unit of analysis and links service identification, functionality formulation, technology stack reconstruction, formal domain correspondence, impact profiling, maturity classification, and building-level aggregation. This enables heterogeneous service descriptions to be converted into structured readiness profiles while preserving the distinction between operational functionality, enabling technology, formal assessment correspondence, and multidimensional impact contribution. Application of the framework to the IoT Building Cloud platform shows that a substantial share of smart building capability may derive from supervisory digital infrastructure rather than from isolated end-use control alone. The resulting readiness profile is characterised by strong representation in monitoring and control, information to occupants and operators, and maintenance awareness, together with more selective contributions to indoor environmental control and limited flexibility-related capability. The proposed framework supports Smart Readiness Indicator-aligned pre-assessment, comparative analysis, design stage reasoning, and digital tool development by providing a transparent bridge between smart building service descriptions and formal assessment-oriented interpretation. Full article

Deep learning methods have significantly improved hyperspectral image (HSI) classification by exploiting hierarchical feature learning to integrate spatial and spectral information, thus significantly improving classification accuracy. Nevertheless, current deep learning approaches (such as CNNs, Transformers and Mamba) still face three major challenges: inadequate mitigation of spectral redundancy, high computational costs associated with global modeling, and the loss of two-dimensional spatial structure during sequential processing. To address these issues, we propose LMamba, a task-oriented hybrid framework that combines multi-scale convolutional filtering with local-context-conditioned state space modeling for hyperspectral image classification. Rather than introducing a fundamentally new SSM formulation, LMamba focuses on adapting the input-dependent parameter projection of Mamba to HSI data by injecting local 2D neighborhood context into the generation of selective SSM parameters. This design enables the state space module to better preserve spatial continuity while maintaining linear-complexity sequence modeling. The framework consists of two core components. First, the Multi-scale Aggregation and Compression Block (MACB) employs parallel grouped convolutions with varying kernel sizes to capture spatial features at multiple scales while simultaneously reducing spectral redundancy through channel compression. Second, the Locally Guided 2D Scanning Mechanism replaces conventional unidirectional 1D scanning with a context-aware 2D scanning strategy, thereby preserving structural continuity and enhancing feature representation by integrating local neighborhood spatial information into state transitions. Validation on three prominent HSI datasets demonstrates that LMamba consistently outperforms state-of-the-art methods based on CNNs, Transformers, and SSMs as measured by overall accuracy (OA), average accuracy (AA), and the Kappa coefficient. In summary, LMamba provides an efficient and accurate HSI classification framework under the considered benchmark settings, and its compact complexity and low-sample robustness suggest potential usefulness for practical HSI analysis. Full article

Transitioning to a circular economy (CE) is now a strategic priority for countries to decouple economic growth from environmental degradation. However, in developing contexts, the readiness of environmental resource sectors to adopt CE principles is unknown due to a lack of data and uneven institutional capacity. This study presents the first regional baseline assessment of circularity readiness in Vietnam’s environmental resource sectors, focusing on land, mining, water and waste. A five-dimensional readiness framework (policy, resource management, innovation, business, awareness) was developed and applied across Vietnam’s six ecological–economic regions. A Delphi process with 12 experts was conducted in three rounds to capture and refine expert judgments, supplemented by triangulated proxy indicators (e.g., plastic recycling rates, wastewater treatment coverage). Readiness scores were aggregated at dimension and regional levels and analyzed using radar charts, heatmaps and hierarchical clustering. Results showed significant regional disparities. The Southeast (SE) and Red River Delta (RRD) have high readiness due to clearer policy frameworks, stronger institutions and more dynamic business ecosystems. The Northern Midlands and Mountains (NMM) and Central Highlands (CH) have low readiness due to infrastructural gaps, weak innovation and limited public engagement. The Mekong Delta (MD) and North Central Coast (NCC) have medium readiness, reflecting partial progress but uneven implementation. The study made three contributions: (1) a new context-specific framework for CE readiness in environmental resource sectors; (2) the value of expert-based, proxy-informed methods in data-scarce contexts; and (3) a policy roadmap for different regional readiness levels. Findings suggest that the CE should be integrated into resource planning, regional observatories should be established and CE-related research and development (R&D) should receive investment. Future research should move towards standardized quantitative indicators and predictive models to track how readiness changes under policy interventions. Full article

This paper investigates solar energy production forecasting at a monthly temporal resolution using a pooled neural network framework applied to the Chikalov photovoltaic systems in southwestern Bulgaria. The study considers several related PV installations with unequal time-series lengths and formulates the forecasting task as one-step-ahead prediction of the next monthly total energy yield, measured in kWh, in a global pooled setting. Two complementary neural architectures are compared: a multilayer perceptron (MLP), which serves as a nonlinear feed-forward benchmark based on lagged observations and seasonal descriptors, and a gated recurrent unit (GRU), which explicitly models sequential temporal dependence. In both cases, seasonality is represented through cyclical calendar encodings, while model selection is performed by chronological hyperparameter search using a separate validation block. Forecast accuracy is assessed by RMSE, MAE, coefficient of determination (R²), MAPE, and sMAPE, and uncertainty is quantified through validation residual prediction intervals. The results show that the MLP achieves stronger validation performance, whereas the GRU provides better final out-of-sample generalization after refitting on the combined training and validation data. For both architectures, the best configurations are obtained with a 12-month input horizon, indicating that one full annual cycle contains the most informative memory for forecasting monthly aggregated photovoltaic energy yield in the considered dataset. After refitting on the combined training and validation data, the GRU achieved the best final out-of-sample performance, with RMSE = 296.38 kWh, MAE = 213.16 kWh, R² = 0.9231, MAPE = 7.52%, and sMAPE = 7.49%. Overall, the findings demonstrate that pooled neural modeling is an effective framework for monthly PV production forecasting and can provide practically useful support for sustainable energy planning, monitoring, and optimization. Full article

Network approaches are commonly used to study mutualistic interactions between frugivorous birds and plants at the community level. However, most fruit–bird networks aggregate individual data and rely on species-level traits, often overlooking intraspecific variation. Here, we downscale a fruit–bird network to the individual level to evaluate intraspecific diet variation and individual specialization in the four main frugivorous passerine species of an island community. Fruit consumption was identified from fecal samples collected from mist-netted birds and individuals’ diets were modeled with a Bayesian approach. Intraspecific diet variation was quantified using the E and NODF indices, individual specialization using the Psi index, and clustering of individuals sharing fruit resources using the Cws index. We detected low intraspecific diet variation and individuals’ diets were not nested. Individual specialization was mainly related to recapture of individuals and weakly related to phenotypic traits. Clustering mainly involved heterospecific individuals whose diets matched plant fruiting phenology during the capture period. Accordingly, future community-level studies addressing the role of mutualistic interactions in biodiversity maintenance may benefit from integrating network approaches with complementary information on interindividual and interspecific competition. Full article

Concrete crack segmentation in bridge health monitoring is crucial for ensuring the safety and longevity of infrastructure. However, this task is complicated by challenges such as weak contrast, background interference, and multi-scale crack structures, which hinder traditional methods’ accuracy. This study introduces a novel Boundary-Sensitive Hybrid Attention Network (BSA-Net) designed to address these issues by combining a hierarchical Transformer encoder (Hiera-A), a multi-scale context module (Light-ASPP), and a boundary-aware decoder (BAD). The hierarchical encoder effectively captures multi-scale features, while Light-ASPP enhances the network’s ability to aggregate contextual information with minimal computational cost, making it suitable for large-scale applications. The dual-branch decoder explicitly decouples the learning of semantic segmentation and boundary prediction, ensuring more accurate boundary detection and crack continuity. The extensive experiments on multiple benchmark datasets demonstrate that BSA-Net consistently outperforms existing crack detection models, particularly in complex, noisy environments. The model achieves competitive performance in terms of segmentation accuracy, boundary clarity, and recall rates, particularly for fine-scale and weak contrast cracks. The results indicate that BSA-Net not only enhances the performance of crack segmentation in real-world conditions but also provides a scalable and reliable solution for automated infrastructure monitoring and defect detection. Full article

Safety gear detection in complex industrial environments faces challenges such as strong background interference, multi-scale spatial perturbations, and the loss of small target features. Furthermore, existing attention-based object detection methods often struggle to balance fine-grained feature retention with background noise suppression. To address these issues, this paper proposes AGCA-YOLOv9, a lightweight object detection model (9.77 M parameters and 39.6 GFLOPs). The core contribution is the Adaptive Gated Coordinate Attention (AGCA) module integrated into the GELAN backbone. Unlike standard coordinate attention mechanisms, AGCA employs a dual-path hybrid pooling strategy combined with an adaptive gated weight fusion mechanism. This design dynamically regulates the synergy between global semantic information and local salient textures, differentiating it from traditional linear feature aggregation. Consequently, it effectively suppresses false detections caused by visually isomorphic backgrounds, such as dense steel frames, while enhancing the representation of distant tiny targets. Validation on the Safety Helmet and Reflective Jacket dataset and the Helmet-Vest-Belt dataset shows that, compared to the YOLOv9s baseline, AGCA-YOLOv9 increases the mAP@50:95 on the Safety Helmet and Reflective Jacket dataset by 0.6% (reaching 80.9%) and the recall rate by 0.4% (reaching 91.9%). Specifically, the mAP@50:95 for the safety helmet category improves by 0.8%. On the Helmet-Vest-Belt dataset, the mAP@50:95 increases by 1.5% (reaching 60.5%). The single-image inference time is 4.6 ms. These results indicate that the proposed algorithm achieves a practical trade-off between detection accuracy and real-time processing speed, demonstrating its potential for safety compliance monitoring in industrial scenarios. Full article

Remote sensing image super-resolution (RSISR) has gained significant attention in recent years due to its critical role in enhancing image analysis capabilities. While existing methods often focus on nonlocal feature extraction, they frequently overlook the importance of local information integration. Moreover, many methods reconstruct images by introducing more complex structures, which poses a challenge to resource-limited devices. To address these issues, we present a local–nonlocal dual-branch feature complementary fusion network (DFCFNet) featuring two key components: a lightweight dual-branch feature aggregation (DBFA) module and an Efficient Feed-Forward Network (EFFN). The DBFA employs a dual-branch structure comprising a Focused Local Feature Branch (FLFB) with novel Partial Convolution Channel Mixers for localized pattern modeling and a Non-Focal Exploration Branch (NFEB) utilizing global variance analysis for comprehensive feature extraction. This dual-branch design enables simultaneous capture of local and global contextual information. The EFFN is designed to further refine the features of the DBFA output in order to make full use of the detailed information of the image. Extensive experimental results show that the proposed DFCFNet reconstructs optimally on remote sensing datasets and is also optimal in terms of computational efficiency and network complexity. The framework’s versatility is further confirmed through successful adaptation to natural image SR tasks, showing consistent performance improvements across five standard datasets. Full article

LiDAR point clouds provide accurate and direct representations of spatial locations and geometric structures of objects in 3D space, making them essential for applications such as target recognition in autonomous driving and 3D reconstruction in smart cities. However, large-scale point clouds pose challenges, including massive data volume, uneven density distribution, and complex object structures. Existing point-based and graph-based semantic segmentation networks often suffer from limitations such as loss of local contextual information, over-reliance on local graph construction, and insufficient modeling of relationships between neighboring points. To address these issues, we propose PG-Net, a novel network that integrates discrete point distribution features with local graph structural information. The framework includes: (1) a point branch equipped with a Local Adaptive Feature Augmentation (LAFA) module to extract efficient local features; (2) a graph branch featuring a Dynamic Graph Feature Aggregation (DGFA) module, which explicitly models relationships among points in local graphs and adaptively balances a point’s intrinsic features with its neighborhood context; and (3) fuses local features from both branches, allowing their complementary strengths to enhance feature representation, a process further promoted by a New Aggregation Loss Function. Experiments on the Toronto3D and S3DIS datasets show that PG-Net achieves overall accuracy (OA) of 97.69% and 89.87%, and mean Intersection-over-Union (mIoU) of 83.51% and 73.22%, respectively. Comparative and ablation studies against advanced methods such as RandLA-Net, BAAF-Net, and LACV-Net demonstrate the effectiveness and robustness of our approach. By jointly exploiting discrete point distribution and local graph structural relationships, PG-Net effectively leverages the complementary strengths of its dual-branch design, offering a reliable solution for efficient and accurate large-scale point cloud semantic segmentation. Full article

In video synthetic aperture radar (Video SAR), target motion causes defocusing, making it impossible to determine the target’s real-time position using reflected echoes. However, the shadows formed by the target occluding ground reflections can accurately characterize the target’s real-time position. To address challenges such as varying shadow scales, low contrast with the moving background, and susceptibility to clutter interference, this paper proposes a shadow detection network called DSENet to enhance the detail and semantic features of shadows. First, to enhance shadow features and reduce sampling loss during backbone network feature extraction, we design a detailed information enhancement (DIE) module to achieve lossless downsampling and effectively preserve the detailed features of the shadowed target. Second, we propose a semantic spatial feature aggregation (SSFA) module to enhance global semantic space feature extraction, improve the contextual feature representation of the target’s shadow region, and provide robust semantic space prior information for the model. Finally, we designed a detailed semantic fusion (DSF) module to improve the neck network’s ability to fuse shadow details and semantic features in video SAR images, further enhancing the model’s localization performance for target shadow features and achieving accurate localization of moving targets in video SAR. Comparative and ablation experiments validate the effectiveness and superiority of the proposed method. Experimental results on the Sandia National Laboratories (SNL) public dataset demonstrate that DSENet is efficient and performs excellently, achieving a P of 92.4% and an F1 score of 83.1%. Full article