Search Results (4,348)

Cordulegaster heros is an endemic species for Central and Southeastern Europe and one of the species protected under the European Union’s Habitats Directive. To adequately protect this species and its habitats, it is of crucial importance to have detailed information on its distribution, habitat requirements and potential threats to its survival. The main aims of this study were to record Cordulegaster heros geographical and altitudinal distribution and habitat requirements (stream width, fine substrate content and habitat shading), along with the importance of protected area network in its conservation and threats to its habitats in Croatia. To achieve those aims, we investigated 201 perennial and intermittent streams across three biogeographical regions (Continental, Alpine, Mediterranean). Additionally, in a small-scale study conducted in streams located within a protected area, we assessed the species’ relationship with water quality. According to the current results, Cordulegaster heros was confirmed to reproduce in 44 perennial streams in the Continental and Alpine regions, with a significantly higher number of sites and species’ abundance recorded in the Continental region. As the species was not recorded in the Mediterranean region, its occurrence there remains unverified. The species occurred at an altitudinal range between 150 and 665 m a.s.l., with 77% of the sites being between 150 and 350 m a.s.l. It was mostly documented in streams with widths up to 250 cm, fine sediment content up to 30%, and high habitat shading (>75%). A small-scale assessment of its relationship with water parameters within a protected area revealed a significant correlation with higher concentration of oxygen and lower conductivity, confirming its requirements for clean and well-oxygenated habitats. Approximately 57% of the sites where this species was recorded are within the protected area network. However, because most known occurrences are concentrated within only one area, the Continental region, along with the rather low population densities and anthropogenic threats (e.g., deforestation, hydro-morphological alterations) present at 43% of those streams, further monitoring activities are necessary. The presented results provide a basis for further monitoring of Cordulegaster heros and its habitats in Croatia. Full article

Detecting small objects in unmanned aerial vehicle (UAV) imagery remains a challenging task due to the limited target scale, cluttered backgrounds, severe occlusion, and motion blur commonly observed in dynamic aerial environments. This study presents STAIR-DETR, a real-time synergistic detection framework derived from RT-DETR, featuring comprehensive enhancements in feature extraction, resolution transformation, and detection head design. A Statistical Feature Attention (SFA) module is incorporated into the neck to replace the original AIFI, enabling token-level statistical modeling that strengthens fine-grained feature representation while effectively suppressing background interference. The backbone is reinforced with a Diverse Semantic Enhancement Block (DSEB), which employs multi-branch pathways and dynamic convolution to enrich semantic expressiveness without sacrificing spatial precision. To mitigate information loss during scale transformation, an Adaptive Scale Transformation Operator (ASTO) is proposed by integrating Context-Guided Downsampling (CGD) and Dynamic Sampling (DySample), achieving context-aware compression and content-adaptive reconstruction across resolutions. In addition, a high-resolution P2 detection head is introduced to leverage shallow-layer features for accurate classification and localization of extremely small targets. Extensive experiments conducted on the VisDrone2019 dataset demonstrate that STAIR-DETR attains 41.7% mAP@50 and 23.4% mAP@50:95, outperforming contemporary state-of-the-art (SOTA) detectors while maintaining real-time inference efficiency. These results confirm the effectiveness and robustness of STAIR-DETR for precise small object detection in complex UAV-based imaging scenarios. Full article

Multimodal emotion recognition (MER) often relies on single-scale representations that fail to capture the hierarchical structure of emotional signals. This paper proposes a Dual Routing Mixture-of-Experts (MoE) model that dynamically selects between local (fine-grained) and global (contextual) representations extracted from speech and text encoders. The framework first obtains local–global embeddings using WavLM and RoBERTa, then employs a scale-aware routing mechanism to activate the most informative expert before bidirectional cross-attention fusion. Experiments on the IEMOCAP dataset show that the proposed model achieves stable performance across all folds, reaching an average unweighted accuracy (UA) of 75.27% and weighted accuracy (WA) of 74.09%. The model consistently outperforms single-scale baselines and simple concatenation methods, confirming the importance of dynamic multi-scale cue selection. Ablation studies highlight that neither local-only nor global-only representations are sufficient, while routing behavior analysis reveals emotion-dependent scale preferences—such as strong reliance on local acoustic cues for anger and global contextual cues for low-arousal emotions. These findings demonstrate that emotional expressions are inherently multi-scale and that scale-aware expert activation provides a principled approach beyond conventional single-scale fusion. Full article

Sentiment classification is a key technique for analyzing the emotional tendency of user reviews and is of great significance to movie recommendation systems. However, existing methods often face challenges in practical applications due to complex model structures, low computational efficiency, or difficulties in balancing local details with global contextual features. To address these issues, this paper proposes a Multi-Scale Feature Fusion Linear Attention model (MSFFLA). The model consists of three core modules: the BERT Encoder module for extracting basic semantic features; the Parallel Multi-scale Feature Extraction module (PMFE) , which employs multi-branch dilated convolutions to accurately capture local fine-grained features; and the Global Multi-scale Linear Feature Extraction module (MGLFE) , which introduces a Multi-Scale Linear Attention mechanism (MSLA) to efficiently model global contextual dependencies with approximately linear computational complexity. Extensive experiments were conducted on three public datasets: SST-2, Amazon Reviews, and MR. The results show that compared to the state-of-the-art BERT-CondConv model, our model achieves improvements in accuracy and F1-Score by 1.8% and 0.4%, respectively, on the SST-2 dataset, and by 1.5% and 0.3% on the Amazon Reviews dataset. This study not only validates the effectiveness of the proposed model but also provides an efficient and lightweight solution for sentiment classification tasks in movie recommendation systems, demonstrating promising practical application prospects. Full article

Understanding outdoor thermal environments at fine spatial scales is essential for developing climate-responsive urban and building design strategies. This study investigates the determinants of local air temperature deviations in Seoul, Korea, using high-resolution in situ sensor data integrated with multi-source urban and building information. Hourly temperature records from 436 road-embedded sensors (March 2024–February 2025) were transformed into relative metrics representing deviations from the network-wide mean and were combined with semantic indicators derived from street-view imagery—Green View Index (GVI), Road View Index (RVI), Building View Index (BVI), Sky View Index (SVI), and Street Enclosure Index (SEI)—along with land-cover and building attributes such as impervious surface area (ISA), gross floor area (GFA), building coverage ratio (BCR), and floor area ratio (FAR). Employing an eXtreme Gradient Boosting (XGBoost)–Shapley Additive exPlanations (SHAP) framework, the study quantifies nonlinear and interactive relationships among morphological, environmental, and visual factors. SEI, BVI, and ISA emerged as dominant contributors to localized heating, while RVI, GVI, and SVI enhanced cooling potential. Seasonal contrasts reveal that built enclosure and vegetation visibility jointly shape micro-scale heat dynamics. The findings demonstrate how high-resolution, observation-based data can guide climate-responsive design strategies and support thermally adaptive urban planning. Full article

Understanding the forces imparted onto an object during manipulation can shed light on the quality of daily manual behaviors. We have developed an objective measure of the quality of hand function in children, the Bead Maze Hand Function Test, which quantifies how well the individual performs the activity by integrating measures of time and force control. Our main objectives were to examine associations between performance (total force output) on the Bead Maze Hand Function Test (BMHFT) and (1) performance on a sensitive measure of force scaling obtained on a laboratory-based dexterous manipulation task, and (2) general sensory and motor parameters important for fine motor skills. A total of 39 typically developing participants ranging in age from 5 to 10 years old (n = 28) and 15 to 17 years (n = 11). We found that the anticipatory coordination of digit forces was the best predictor of performance on the Bead Maze Hand Function test. We also found that factors such as age, gender, and pinch strength were associated with the BMHFT performance. These findings support the integration of more sensitive sensorimotor metrics, such as the total applied force, into clinical assessments. Linking the development of sensorimotor capabilities to functional task performance may facilitate more targeted and effective intervention strategies, ultimately improving a child’s participation in daily activities. Full article

Early identification of motor difficulties is essential in infancy and early childhood, and current American Academy of Pediatrics recommendations emphasize that motor surveillance should accompany routine clinical visits. One standardized tool widely used for evaluating motor development is the Peabody Developmental Motor Scales–Second Edition (PDMS-2). This review summarizes the theoretical foundations and psychometric properties of the PDMS-2, the principles of administering and scoring the assessment, and evidence from validation and standardization studies conducted in different countries. A non-systematic literature search was conducted in PubMed, Scopus, and Google Scholar (2000–February 2025) using the terms “PDMS-2” OR “Peabody Developmental Motor Scales Second Edition” combined with “reliability”, “validity”, “norms”, “reference”, or “standardization”. Original and review articles published in English were included without geographical restrictions. The PDMS-2 is widely applied in both clinical and research contexts. It has been used as an outcome measure in randomized controlled trials, interventional, and observational studies involving preterm infants, children with genetic syndromes, metabolic disorders, cerebral palsy, congenital heart defects, HIV, oncological conditions, and typically developing children. Key strengths of the PDMS-2 include its broad age range, the ability to assess both gross and fine motor skills, and its quantitative scoring system, which supports diagnosis, therapeutic planning, and monitoring of developmental change. Although the tool has been validated and standardized in multiple countries, additional work is still needed to establish normative data for underrepresented populations. Full article

This work establishes a large language model (LLM) specialized in the domain of thermoelectric generators (TEGs), for deployment on local hardware. Starting with the generalist JanV1-4B model and Qwen3-4B-Thinking-2507 models, an efficient fine-tuning (FT) methodology using quantized low-rank adaptation (QLoRA) was employed, modifying only 3.18% of the total parameters of thee base models. The key to the process is the use of a custom-designed dataset, which merges deep theoretical knowledge with rigorous instruction tuning to refine behavior and mitigate catastrophic forgetting. The dataset employed for FT contains 202 curated questions and answers (QAs), strategically balanced between domain-specific knowledge (48.5%) and instruction-tuning for response behavior (51.5%). Performance of the models was evaluated using two complementary benchmarks: a 16-question multilevel cognitive benchmark (94% accuracy) and a specialized 42-question TEG benchmark (81% accuracy), scoring responses as excellent, correct with difficulties, or incorrect, based on technical accuracy and reasoning quality. The model’s utility is demonstrated through experimental TEG design guidance, providing expert-level reasoning on thermal management strategies. This study validates the specialization of LLMs using QLoRA as an effective and accessible strategy for developing highly competent engineering support tools, eliminating dependence on large-scale computing infrastructures, achieving specialization on a consumer-grade NVIDIA RTX 2070 SUPER GPU (8 GB VRAM) in 263 s. Full article

Industrial predictive maintenance at the edge faces persistent challenges such as extreme class imbalance, limited labeled failure data, and the need for efficient yet scalable AI models. This paper proposes a transfer learning-based edge AI framework that addresses these challenges through a signal-to-image transformation and fine-tuning of deep residual networks (ResNet). One-dimensional sensor signals are converted into two-dimensional RGB images, enabling the use of powerful convolutional architectures originally trained on large-scale datasets. The approach emulates an edge–cloud synergy, where knowledge distilled from large pre-trained models is efficiently adapted and executed on resource-constrained edge environments. Trained on less than 5% of the original dataset, the model achieves a negative predictive value of 96.53%, significantly reducing classification cost and outperforming both conventional deep learning and traditional machine learning methods. The results demonstrate that transfer learning-driven edge intelligence offers a cost-effective, scalable, and generalizable solution for predictive maintenance and industrial automation under data scarcity. Full article

Low-resolution (LR) imagery poses significant challenges to object detection systems, particularly in autonomous and resource-constrained environments where bandwidth and sensor quality are limited. To address this issue, this paper presents an integrated framework that enhances object detection performance by incorporating a Super-Resolution (SR) preprocessing stage prior to detection. Specifically, a Dense Residual Connected Transformer (DRCT) is employed to reconstruct high-resolution (HR) images from LR inputs, effectively restoring fine-grained structural and textural information essential for accurate detection. The reconstructed HR images are subsequently processed by a YOLOv11 detector without requiring architectural modifications. Experimental evaluations demonstrate consistent improvements across multiple scaling factors, with an average increase of 13.4% in Mean Average Precision (mAP)@50 at ×2 upscaling and 9.7% at ×4 compared with direct LR detection. These results validate the effectiveness of the proposed SR-based preprocessing approach in mitigating the adverse effects of image degradation. The proposed method provides an improved yet computationally challenging solution for object detection. Full article

The SDG 15.3.1 framework provides a standardized approach using land use/land cover (LULC) change, land productivity, and soil organic carbon (SOC) dynamics to assess land degradation. However, SDG 15.3.1. faces limitations like coarse resolutions of Landsat-8 and Sentinel-2, particularly for fine-scale studies. Accordingly, this paper integrates Very Deep Super-Resolution (VDSR) for downscaling Landsat-8 imagery to 1 m resolution and the Vegetation Health Index (VHI) into SDG 15.3.1 to enhance detection in the heterogeneous Loiret region, France—a temperate agricultural hub featuring mixed croplands and peri-urban interfaces—using 2017 as baseline and 2024 as target. Results demonstrated that 1 m resolution detected more degraded LULC areas than coarser scales. SOC degradation was minimal (0.15%), concentrated in transitioned zones. VHI reduced overestimation of productivity declines compared to the Normalized Difference Vegetation Index by identifying more stable areas and 2.69 times less degradation in integrated assessments. The “One Out, All Out” rule classified 2.6% (using VHI) and 7.1% (using NDVI) of the region as degraded, mainly in peri-urban and cropland hotspots. This approach enables metre-scale land degradation mapping that remains effective in heterogeneous landscapes where fine-scale LULC changes drive degradation and would be missed at lower resolutions. However, future ground validation and longer timelines are essential to enhance the presented methodology. Full article

Many deep learning-based spatiotemporal fusion (STF) methods have been proven to achieve high accuracy and robustness. Due to the variable shapes and sizes of objects in remote sensing images, pyramid networks are generally introduced to extract multi-scale features. However, the down-sampling operation in the pyramid structure may lead to the loss of image detail information, affecting the model’s ability to reconstruct fine-grained targets. To address this issue, we propose a novel Dual-Stream Enhanced Pyramid based on Generative Adversarial Network (DSEPGAN) for the spatiotemporal fusion of remote sensing images. The network adopts a dual-stream architecture to separately process coarse and fine images, tailoring feature extraction to their respective characteristics: coarse images provide temporal dynamics, while fine images contain rich spatial details. A reversible feature transformation is embedded in the pyramid feature extraction stage to preserve high-frequency information, and a fusion module employing large-kernel and depthwise separable convolutions captures long-range dependencies across inputs. To further enhance realism and detail fidelity, adversarial training encourages the network to generate sharper and more visually convincing fusion results. The proposed DSEPGAN is compared with widely used and state-of-the-art STF models in three publicly available datasets. The results illustrate that DSEPGAN achieves superior performance across various evaluation metrics, highlighting its notable advantages for predicting seasonal variations in highly heterogeneous regions and abrupt changes in land use. Full article

Urban stormwater runoff, particularly during first-flush events, carries high loads of fine suspended solids and phosphorus that are difficult to remove with conventional best management practices (BMPs). This study developed and evaluated a laboratory-scale high-efficiency up-flow filtration system with Internet of Things (IoT)-based autonomous control. The system employed 20 mm fiber-ball media in a modular dual-stage up-flow configuration with optimized coagulant dosing to target fine particles (<3 μm) and total phosphorus (TP). Real-time turbidity and pressure monitoring via sensor networks connected to a microcontroller enabled wireless data logging and automated backwash initiation when thresholds were exceeded. Under manual operation, the two-stage filter achieved removals of 96.6% turbidity, 98.8% suspended solids (SS), and 85.6% TP while maintaining head loss below 10 cm. In IoT-controlled single-stage runs with highly polluted influent (turbidity ~400 NTU, SS > 1000 mg/L, TP ~1.6 mg/L), the system maintained >90% SS and ~58% TP removal with stable head loss (~8 cm) and no manual intervention. Turbidity correlated strongly with SS (R² ≈ 0.94) and TP (R² ≈ 0.87), validating its use as a surrogate control parameter. Compared with conventional BMPs, the developed filter demonstrated superior solids capture, competitive phosphorus removal, and the novel capability of real-time autonomous operation, providing proof-of-concept for next-generation smart BMPs capable of meeting regulatory standards while reducing maintenance. Full article

An accurate and reliable extraction of building structures from high-resolution (HR) remote sensing images is an important research topic in 3D cartography and smart city construction. However, despite the strong overall performance of recent deep learning models, limitations remain in handling significant variations in building scales and complex architectural forms, which may lead to inaccurate boundaries or difficulties in extracting small or irregular structures. Therefore, the present study proposes MSA-UNet, a reliable semantic segmentation framework that leverages multiscale feature aggregation and attentive skip connections for an accurate extraction of building footprints. This framework is constructed based on the U-Net architecture, incorporating VGG16 as a replacement for the original encoder structure, which enhances its ability to capture low-discriminative features. To further improve the representation of image buildings with different scales and shapes, a serial coarse-to-fine feature aggregation mechanism was used. Additionally, a novel skip connection was built between the encoder and decoder layers to enable adaptive weights. Furthermore, a dual-attention mechanism, implemented through the convolutional block attention module, was integrated to enhance the focus of the network on building regions. Extensive experiments conducted on the WHU and Inria building datasets validated the effectiveness of MSA-UNet. On the WHU dataset, the model demonstrated a state-of-the-art performance with a mean Intersection over Union (mIoU) of 94.26%, accuracy of 98.32%, F1-score of 96.57%, and mean Pixel accuracy (mPA) of 96.85%, corresponding to gains of 1.41% in mIoU over the baseline U-Net. On the more challenging Inria dataset, MSA-UNet achieved an mIoU of 85.92%, indicating a consistent improvement of up to 1.9% over the baseline U-Net. These results confirmed that MSA-UNet markedly improved the accuracy and boundary integrity of building extraction from HR data, outperforming existing classic models in terms of segmentation quality and robustness. Full article

Understanding the deformation–failure process of sandstone is essential for energy extraction and stability assessment. Here, laboratory mechanical tests and discrete element simulations are combined to resolve how grain size controls deformation, cracking, and failure. Under uniaxial compression, fine-grained sandstone shows the highest strength (60.85–65.37 MPa) yet undergoes an abrupt brittle transition to axial splitting at a small peak axial strain of 0.41–0.42%; coarse-grained sandstone exhibits lower strength (26.94–28.67 MPa) but fails at peak axial strains of 0.44–0.53%, on average about 17% higher than those of FGS, indicating enhanced ductility; medium-grained sandstone lies in between in both strength (41.15–43.79 MPa) and peak axial strain (0.42–0.45%). With confining pressure, fine- and medium-grained sandstones display pronounced process evolution toward ductility, whereas coarse-grained sandstone shows limited pressure sensitivity. DEM results link microcrack evolution with the macroscopic response: under uniaxial loading, fine-grained sandstone is dominated by intergranular tensile cracking, while coarse-grained sandstone includes more intragranular cracks. Increasing confinement controls the cracking process, shifting fine- and medium-grained rocks from intergranular tension to mixed intragranular tension–shear, thereby enhancing ductility; in contrast, coarse-grained sandstone at high confinement localizes shear bands and remains relatively brittle. Normalized microcrack aperture distributions and fragment identification capture a continuous damage accumulation process from micro to macro scales. These process-based insights clarify the controllability of failure modes via grain size and confinement and offer optimization-oriented guidance for design parameters that mitigate splitting and promote stable deformation in deep sandstone reservoirs and underground excavations. Full article