Search Results (12,706)

Building damage and collapse are the main sources of serious casualties and financial losses during earthquakes, which are among the most destructive natural disasters that endanger human life and property. Therefore, quick and precise post-earthquake building damage assessment is essential for risk assessment and emergency action. Convolutional neural networks (CNNs) primarily concentrate on local features and frequently ignore global contextual information within and across buildings, despite the fact that deep learning-based techniques allow automated damage identification. Transformer-based approaches, on the other hand, are good at capturing global dependencies, but their large memory and processing costs restrict their usefulness. As a result, existing networks still struggle to achieve an effective balance between accuracy and efficiency. To address this issue, this study proposes a lightweight and efficient network for post-earthquake building damage assessment. Specifically, we develop a two-stage method based on EfficientViM with an encoder–decoder architecture. In the encoder, Mamba is introduced to extract multi-scale change features with long-range dependencies, leveraging the state space model to preserve global modeling capability while significantly reducing computational complexity. In the decoder, two lightweight modules are designed to further enhance discriminative capability and computational efficiency. The network finally outputs building localization and pixel-level building damage, respectively. Experiments were conducted on four earthquake events from the BRIGHT dataset using a three-for-training and one-for-testing cross-event rotation evaluation strategy. The results demonstrate that LEViM-Net requires only 30.94 M parameters and 27.10 G FLOPs. In addition, for the Türkiye earthquake event, the proposed method achieves an F1 score of 80.49%, an overall accuracy (OA) of 88.17%, and a mean intersection over union (mIoU) of 49.73%. The proposed model enables efficient remote-sensing-based mapping of macroscopic and image-visible building damage, providing timely support for early-stage emergency response. Full article

Open-vocabulary semantic segmentation (OVSS) aims to achieve pixel-level object segmentation guided by arbitrary natural language descriptions. Although pre-trained vision–language models (VLMs) have significantly advanced the development of OVSS, their reliance on the Vision Transformer (ViT) architecture imposes a fundamental constraint on dense prediction. Specifically, the absence of hierarchical downsampling in ViT-based VLM results in single-scale representations that trade spatial localization for global semantics. To address these issues, this paper proposes a hierarchical boundary-constrained decoding network for OVSS, called CLIP-HBD. Our approach leverages VLM semantic priors to reconstruct multi-scale features and introduces a boundary-constrained decoding strategy to refine edge details. Specifically, CLIP-HBD leverages a ConvNeXt-based backbone alongside a hierarchical adaptation mechanism to fuse multi-layer VLM features, generating a comprehensive multi-scale representation. To address the issue of boundary inaccuracy, we perform explicit boundary prediction based on multi-scale representations, where the resulting boundary maps are subsequently transformed into structural constraints to steer the decoder’s focus toward boundary regions. By integrating structural constraints with hierarchical features, the decoding process effectively maintains semantic consistency and restores precise object boundaries. Extensive experiments demonstrate that CLIP-HBD achieves superior performance in both segmentation precision and boundary quality across multiple benchmarks. Full article

Against the global backdrop of carbon neutrality and the green transition of the construction sector, modern timber-framed buildings have emerged as a core enabler of sustainable construction. However, a systematic synthesis of research on indoor hygrothermal environments and thermal comfort in such buildings remains lacking, and the underlying coupling mechanisms—as well as pathways for performance optimization—are still insufficiently understood. To address these gaps, this study aims to systematically characterize and evaluate the performance features of indoor thermal and moisture environments in modern timber buildings, and to identify the key influencing factors and their underlying mechanisms. In accordance with the PRISMA 2020 guidelines for systematic reviews, this study identified and analyzed 203 high-quality peer-reviewed publications retrieved from three major academic databases, covering the period 2010–2025. Specifically, the literature search was conducted across the Web of Science, Scopus, and the China National Knowledge Infrastructure (CNKI), and visualization analysis was performed using VOSviewer 1.6.20 software. The results indicate that timber-framed buildings exhibit distinctive indoor hygrothermal characteristics: rapid temperature response, strong humidity buffering capacity, and superior thermal insulation performance compared with concrete structures, enabling indoor relative humidity to remain stably within the thermally comfortable range. Nevertheless, challenges persist, including summer overheating and elevated risks of mold growth under hot-humid conditions. Furthermore, the PMV model demonstrates significant predictive deviation for thermal comfort in timber-framed buildings; its application thus requires calibration incorporating both the hygrothermal properties of timber materials and occupants’ psychological adaptation. This study synthesizes the current state of research, identifies key influencing factors, and proposes climate-responsive optimization strategies to advance the development of robust thermal comfort models and support the low-energy, high-comfort design of timber-framed buildings. Full article

Accurate identification of key nodes in complex networks is vital for optimizing system robustness and controlling information spread. Existing centrality metrics struggle to balance the continuous extraction of global topological features with the fine-grained perception of local structures, while traditional heuristic algorithms also face severe resolution limitations. To address these issues, this paper proposes a node importance evaluation method based on fractional-order topological propagation and local information entropy (FSEC). This method overcomes the limitations of discrete integer-order propagation inherent in traditional graph walks. It constructs a continuous fractional-order topological propagation operator within the spectral graph theory framework. This enables the smooth projection of node degree features into the global topological space, thereby yielding high-order global impact factors. Furthermore, an information theory mechanism is introduced to quantify the probability distribution of a node’s information contribution within its local neighborhood. The local structural information entropy is then calculated to reflect the node’s asymmetric control over micro-level information flow. Deliberate attack simulations were conducted on nine real-world networks and three types of artificial network models. The results show that the proposed FSEC algorithm significantly outperforms baseline algorithms like Autoencoder and Graph Neural Network (AGNN), Degree Centrality, k-shell, PageRank, and Mixed Degree Decomposition (MDD) in degrading the largest connected component (LCC) and global network efficiency (NE). The proposed method also achieves the minimum Area Under the Curve (AUC) values globally. Its monotonicity is slightly lower than that of AGNN but superior to all other baseline algorithms. In addition, SIR simulations further confirm the effectiveness of the FSEC method. This approach successfully resolves the ranking tie problem among nodes in the same topological layer. Full article

Occlusion is a critical challenge in person re-identification (ReID), as partial visibility severely degrades feature discriminability and matching reliability. To address this issue, we propose a novel framework termed Visibility-Guided and Occlusion-Simulated Learning (VGOSL) for robust person ReID. The framework consists of two key components: a part-aware visibility modeling (PVM) module and an occlusion box simulation (OBS) module. The PVM module explicitly estimates part-level visibility reliability and adaptively reweights local features to guide global representation learning, enabling the model to emphasize informative regions while suppressing occluded ones. Meanwhile, the OBS module introduces structured occlusion box simulation during training to enhance robustness against realistic obstruction patterns through multi-branch supervision. Extensive experiments on Occluded-DukeMTMC, DukeMTMC-reID, Market-1501, Partial-ReID, and MSMT17 demonstrate that the proposed framework achieves competitive performance under both occluded and holistic settings. The source code has been publicly released on GitHub. Full article

Autonomous wheelchair navigation requires both reliable global guidance and safe local interaction with the environment, typically addressed using separate perception and control strategies. This paper presents a unified vision-based control framework that combines learning-based corridor following with image-based obstacle avoidance under a common visual servoing perspective. This work provides a unified interpretation of learning-based and analytical control as complementary realizations of visual servoing. A convolutional neural network (CNN) is employed to directly predict steering commands from monocular images, enabling robust corridor following without explicit feature extraction. In parallel, obstacle avoidance is formulated as an image-based visual servoing (IBVS) task, where detected obstacles are represented as image features and regulated toward safe regions. A supervisory control strategy coordinates these components by prioritizing safety-critical avoidance when necessary, while maintaining nominal navigation otherwise. The system is implemented using a single monocular camera and deployed on a low-cost embedded platform. Experimental results demonstrate that the CNN-based module maintains stable performance under challenging visual conditions, while the IBVS controller provides predictable and reliable avoidance behavior. The proposed framework highlights the complementary roles of learning-based and analytical visual servoing, offering a practical and scalable solution for assistive autonomous mobility. Full article

Infrared (IR) object detection on unmanned aerial vehicle (UAV) platforms is fundamentally challenged by low signal-to-noise ratios and extremely tight onboard computational budgets. Conventional CNNs lack sufficient global context, while Transformers suffer from quadratic complexity, hindering real-time deployment. To address these bottlenecks, we propose EfficientIR-Det, a lightweight end-to-end detector featuring a holistic optimization of the backbone, encoder, and sampling mechanisms. Specifically, we design a Partial Star Network (PSN) backbone that achieves implicit high-dimensional feature expansion via element-wise multiplication to amplify weak IR signals with minimal redundancy. Furthermore, a Hierarchical Mamba (HiMamba) encoder leverages selective state-space modeling to provide linear-complexity global enhancement with superior hardware efficiency. To refine cross-scale representations, we introduce an Adaptive Gated Sampling (AGS) module and a Hierarchical Sampling Strategy (HSS) to optimize feature fusion and sampling budget allocation toward dim-small targets. On HIT-UAV, EfficientIR-Det achieves 88.4% mAP@0.5, outperforming the RT-DETR-R18 baseline by 3.3 points while reducing FLOPs and parameters by 48.9% and 44.2%, respectively. On the larger-scale DroneVehicle dataset, it consistently leads with a 74.1% mAP@0.5 and a high inference speed of 140.8 FPS. Our results offer a promising research scheme for robust, real-time infrared perception on edge-constrained UAV platforms. Full article

Scene graph generation (SGG) provides a structured representation for visual understanding. However, most existing methods are designed to optimize global triplet recall rather than retrieve relation instances specified by a user query. In query-driven visual relation retrieval, two major challenges arise: the target relation must compete with a highly redundant candidate space, and query semantics are not incorporated before relation classification. To address these challenges, we propose a Query-Driven Candidate Relation Screening (QCRS) module, which injects query semantics into the candidate screening process. Specifically, QCRS encodes the query and candidate visual relation features, and then filters query-relevant candidates through relevance scoring. By reducing interference from irrelevant candidates and avoiding redundant computation, QCRS improves the final exact triplet hit performance and enhances the interpretability of query-specific relations, thereby facilitating query-driven visual relation retrieval. Built upon the strong EGTR baseline, QCRS learns query relevance to prioritize relation instances matching the target query, enabling precise triplet retrieval. Extensive ablation studies and analyses on the VG150 benchmark validate the effectiveness of the proposed approach: when integrated with EGTR, QCRS improves PairR@50 from 61.52% to 80.06% and ETR@50 from 30.54% to 47.07%, achieving absolute gains of over 16 percentage points in both correct object pair retention and end-to-end target relation retrieval performance. Full article

Accurate direction-of-arrival (DOA) estimation with high resolution is fundamental to many array sensing applications. In practice, however, sparse arrays with missing sensors and snapshot-limited observations often lead to incomplete and noisy second-order statistics, which substantially degrades the performance of conventional eigendecomposition-based estimators. In this paper, we propose a two-stage Transformer framework for sparse-array DOA estimation that explicitly separates correlation recovery from angle inference. The first stage operates in the correlation domain and learns to reconstruct a clean and complete correlation vector from partially observed measurements using masking-aware tokenization and global-context modeling. The recovered representation can be further converted into a structured covariance matrix, providing an interpretable interface to classical signal processing back-ends. Based on the recovered features, the second stage adopts a Transformer regressor to directly predict multi-source DOAs. Extensive simulations on a large-scale dataset with SNRs from −5 to 10 dB and various snapshot numbers demonstrate that the proposed method delivers robust accuracy and improved stability in low-SNR and snapshot-limited regimes, while maintaining competitive performance at higher SNRs. Additional evaluations with an ESPRIT back-end further confirm that the recovery-based covariance yields more reliable DOA estimation than conventional difference–coarray processing, with particularly evident gains under challenging noise conditions. Full article

Change detection in open-pit mining areas from remote sensing imagery is of great importance for mining supervision, ecological monitoring, and restoration planning. Nevertheless, mining-related changes usually exhibit multi-scale patterns, irregular boundaries, and fragmented spatial distributions, which make accurate detection difficult. Existing CNN- and Transformer-based methods often cannot effectively balance global context perception and local detail preservation, resulting in incomplete boundary extraction and insufficient sensitivity to subtle changes. To overcome these limitations, we propose GLMECD-Net, a Global–Local Multi-scale Cross-fusion Enhanced Change Detection Network for remote sensing image change detection in open-pit mining areas. Specifically, a Siamese encoder is used to extract hierarchical bi-temporal features, while a Global–Local Feature Mixing Embedding (GLME) module is introduced to jointly capture long-range contextual information and local spatial details. Furthermore, multi-scale feature aggregation and cross-temporal feature fusion are employed to improve change representation and boundary recovery. Experimental results on mining area datasets show that the proposed method achieves 71.66% Precision, 83.78% OA, 77.53% F1-score, and 53.82% IoU. The results demonstrate that GLMECD-Net provides effective and robust performance for detecting complex and subtle changes in open-pit mining areas. Full article

With the continuous growth of global population and increasing pressure on food security, the transformation toward precise and intelligent agricultural production has become an inevitable trend. In this context, accurate identification of field weeds is crucial for improving crop yields and reducing agricultural inputs. However, agricultural Internet of Things (IoT) edge devices are generally subject to strict constraints in terms of power consumption, storage, and real-time performance. Existing lightweight convolutional neural networks often struggle to simultaneously achieve high accuracy and low resource consumption for fine-grained weed identification tasks. To address this challenge, this paper proposes a hardware aware lightweight convolutional neural network named Weed-Aware Ghost Enhanced Network (WAGENet) for microcontroller deployment. The network synergistically integrates Ghost low-cost feature generation, Mobile Inverted Bottleneck Convolution (MBConv) for deep semantic extraction, Squeeze and Excitation (SE) and Coordinate Attention (CA) dual attention mechanisms for channel space joint calibration, and Atrous Spatial Pyramid Pooling (ASPP) for multi-scale context fusion. It constructs a progressive feature abstraction system from shallow textures to high-level semantics. On the public DeepWeeds dataset, WAGENet achieves 95.71% classification accuracy and 93.80% F1 score with only 0.163 M parameters and 2.43 × 10⁸ multiply accumulate operations (MACC), attaining a parameter efficiency of 587.19%/M and significantly outperforming existing mainstream lightweight models. The model has been successfully deployed on the STM32H7B3I microcontroller development board, achieving a single inference latency of 94.63 ms, an internal Flash footprint of only 686.95 KiB, and a single inference energy consumption of 41.45 mJ. Experimental results demonstrate that WAGENet achieves a trade off among accuracy, latency, and energy consumption under strict resource constraints, providing a reproducible microcontroller deployment paradigm for battery powered field robots, drones, and other agricultural IoT edge devices. Full article

Accurate extrinsic calibration among multiple heterogeneous Light Detection and Ranging (LiDAR) sensors is essential for autonomous vehicle perception systems, yet remains challenging in distributed topologies where overlap exists only between adjacent sensor pairs. Existing methods often assume a central LiDAR with direct field-of-view overlap to all others and suffer from error accumulation in sequential pairwise registration. This paper presents a targetless, motionless multi-LiDAR extrinsic calibration framework that is topology-agnostic and resolves error accumulation through global optimization. The method integrates (1) Random Sample Consensus (RANSAC)-based planar patch extraction with a dual-criterion normal-guided matching strategy, (2) robust coarse alignment via TEASER++, and (3) pose graph optimization with analytically derived edge weights from Generalized Iterative Closest Point (GICP) covariance matrices. The use of structural planar primitives rather than local point descriptors overcomes density-dependent matching failures inherent to heterogeneous sensor pairs, while global pose graph optimization eliminates the cumulative error propagation of sequential pairwise approaches. Validation is performed on three distinct real-world configurations: a six-LiDAR autonomous port truck (ring topology), the four-LiDAR EDGAR research vehicle (distributed topology), and a three-LiDAR benchmark from the OpenCalib toolbox. The proposed method consistently outperforms state-of-the-art baselines, achieving 0.021 m translation Root Mean Square Error (RMSE) and 0.36° rotation RMSE on the port dataset, with full calibration completed in under 2 s on CPU—enabling rapid in-situ recalibration without requiring dedicated facilities or vehicle motion. Full article

Deep reinforcement learning often suffers from inefficient exploration, which is commonly addressed by introducing an auxiliary model that assigns intrinsic rewards when the agent encounters novel scenarios. However, such approaches increase training complexity and computational overhead. This paper proposes an Attribution-Guided Reinforcement Learning (AGRL) framework that exploits real-time attribution analysis to guide exploration in autonomous driving decision-making. The proposed method is built upon the Kolmogorov–Arnold-Network-based Interpretable Deep Reinforcement Learning (KAN-IDRL) framework. Specifically, action-wise attribution patterns are computed online, and perturbations are applied to the state inputs to measure attribution sensitivity. The resulting attribution-sensitivity signal identifies actions whose decision rationales are more locally responsive to state changes, and these actions are therefore preferentially explored. In addition, local attribution results collected from a pretrained interpretable policy are aggregated into global feature-importance scores, which are then used to initialize a trainable prior attention gate in a Prior-Attention-Enhanced Kolmogorov–Arnold Network (PAE-KAN). This design allows the policy to incorporate attribution-derived prior knowledge while maintaining sufficient adaptability for task-specific learning. Experiments across multiple autonomous driving scenarios demonstrate that the proposed AGRL framework achieves faster convergence and competitive final performance compared with representative baseline methods. These findings indicate that attribution information can be transformed from a post hoc interpretability tool into an effective guidance signal for improving reinforcement learning. Full article

Visual similarity analysis of ethnic brocades is valuable for image retrieval, style comparison, and digital archiving in cultural heritage informatics. However, although deep neural networks provide powerful visual representations, their encoded similarity structures are often difficult to interpret. This study presents an interpretable deep feature framework for analyzing inter-ethnic visual similarity in brocade images from ten minority groups in Southwest China. Four convolutional neural network backbones, including AlexNet, VGG-16, ResNet-18, and an SE-enhanced ResNet-18 (SResNet-18), were first evaluated to identify a reliable feature extractor. The best-performing model was then used to construct deep feature-based similarity and distance relationships among ethnic categories. To interpret this structure, five handcrafted descriptor types, namely color, texture, geometric, local-structure, and frequency-domain features, were compared with the deep feature similarity matrix using Spearman correlation analysis and weighted descriptor fusion. Experimental results showed that SResNet-18 achieved the best classification performance, with an accuracy of 95.15% and an F1-score of 95.14%. Among the handcrafted descriptors, color showed the strongest correspondence with the RGB-based deep similarity structure ($r=0.643$), followed by local-structure descriptors ($r=0.416$), whereas classical texture descriptors showed near-zero correspondence ($r=-0.063$). The optimal weighted fusion further improved the correlation to $r=0.731$. These findings suggest that the SResNet-18 feature space is more strongly associated with color composition and local motif organization than with the specific grayscale texture, global geometric, or frequency-domain descriptors used in this study. The proposed framework provides an interpretable approach for understanding deep visual similarity in cultural heritage images and offers methodological support for pattern-based retrieval, comparative style analysis, and digital documentation. Full article

Background/Objectives: Preclinical studies of head and neck squamous cell carcinoma (HNSCC) commonly use subcutaneous heterotopic (flank) tumor models for simplicity; however, orthotopic models may better reflect the native tumor environment. Direct comparisons of the tumor immune microenvironments (TIME) and tumor-draining lymph nodes (tdLNs) between these models remain limited. Better understanding of site-specific immune differences could improve model selection and interpretation of translational HNSCC studies. Methods: ROC1 tumors were established in murine heterotopic and orthotopic sites, followed by assessment of tumor growth kinetics, survival, and the tumor microenvironment. Immune composition of tumors, blood, tdLNs, and spleen was evaluated at three tumor progression timepoints using multiparameter spectral flow cytometry. Results: Heterotopic and orthotopic tumor models showed similar growth kinetics and survival. Immune profiling revealed increased infiltration of CD3⁺ T-cells, natural killer (NK) cells, and myeloid populations in both models. Heterotopic tumors were enriched in dendritic cells (DCs), plasmacytoid DCs, and monocytic myeloid-derived suppressor cells (M-MDSCs), whereas orthotopic tumors showed increased macrophages, granulocytic MDSCs, and M-MDSCs. Despite temporal variation, both TIMEs were dominated by macrophages, DCs, and CD3⁺ T-cells. Late-stage heterotopic tumors contained more CD4⁺ T-cells. Reduced T-cell cytotoxicity (PD-1, CD107a) and increased immune checkpoint expression across myeloid cells indicated an immunosuppressive TIME. Systemically, effector cells were preserved despite suppressive cell trafficking, and tdLNs in both models exhibited immunosuppressive PD-L1 expression. Conclusions: Heterotopic and orthotopic ROC1 tumors share key immune features, but site-specific differences in the TIME and tdLNs reveal tissue-dependent regulation. These local effects align with systemic changes, supporting global tumor-associated immunosuppression. Full article