Search Results (372)

Methane fluxes in disturbed peatlands can exhibit significant heterogeneity with regard to land cover composition on abandoned peat extraction areas. The temporal and spatial variability of CH₄ fluxes is considered in this paper in the context of a detailed vegetation classification on a typical milled peatland in the Baltic region of Russia (Kaliningrad oblast, Rossyanka Carbon Supersite). The findings are derived from the analysis of 12,000 air samples obtained by the opaque emission chamber method at 10 peatland sites with different environmental characteristics during regular measurement campaigns of 2022–2024. The emission data have been mapped using a multilevel B-spline interpolation procedure. The mean cumulative methane flux was found to be 18.7–28.8 kg ha⁻¹yr⁻¹, which is close to the IPCC conventional value of 32.9 kg ha⁻¹yr⁻¹ estimated for boreal and temperate zones. However, environmental distinctions across the peatland sites result in considerable emission heterogeneity ranging from −0.02 to 11.5 kg ha⁻¹month⁻¹. Temperature is considered a principal factor responsible for the baseline CH₄ emission level in seasonal scale, while hydrology defines emission rate during the warm period of the year and in the inter-annual scales. Five peatland site types have been defined according to a level of methane emissions. Full article

Diabetic Retinopathy (DR) is a leading cause of vision impairment globally, underscoring the need for accurate and early diagnosis to prevent disease progression. Although fundus imaging serves as a cornerstone of Computer-Aided Diagnosis (CAD) systems, several challenges persist, including lesion scale variability, blurry morphological patterns, inter-class imbalance, limited labeled datasets, and computational inefficiencies. To address these issues, this study proposes an end-to-end diagnostic framework that integrates an enhanced preprocessing pipeline with a novel deep learning architecture, Hierarchical-Inception-Residual-Dense Network (HIRD-Net). The preprocessing stage combines Contrast Limited Adaptive Histogram Equalization (CLAHE) with Dilated Difference of Gaussian (D-DoG) filtering to improve image contrast and highlight fine-grained retinal structures. HIRD-Net features a hierarchical feature fusion stem alongside multiscale, multilevel inception-residual-dense blocks for robust representation learning. The Squeeze-and-Excitation Channel Attention (SECA) is introduced before each Global Average Pooling (GAP) layer to refine the Feature Maps (FMs). It further incorporates four GAP layers for multi-scale semantic aggregation, employs the Hard-Swish activation to enhance gradient flow, and utilizes the Focal Loss function to mitigate class imbalance issues. Experimental results on the IDRiD-APTOS2019, DDR, and EyePACS datasets demonstrate that the proposed framework achieves 93.46%, 82.45% and 79.94% overall classification accuracy using only 4.8 million parameters, highlighting its strong generalization capability and computational efficiency. Furthermore, to ensure transparent predictions, an Explainable AI (XAI) approach known as Gradient-weighted Class Activation Mapping (Grad-CAM) is employed to visualize HIRD-Net’s decision-making process. Full article

Lithology identification is a critical technology for geological resource exploration and engineering safety assessment. However, traditional methods suffer from insufficient feature representation and low classification accuracy due to challenges such as weathering, vegetation cover, and spectral overlap in complex sedimentary rock regions. This study proposes a hierarchical multi-feature random forest algorithm based on Feature-Preserved Compressive Sampling (FPCS). Using 3D laser point cloud data from the Manas River outcrop in the southern margin of the Junggar Basin as the test area, we integrate graph signal processing and multi-scale feature fusion to construct a high-precision lithology identification model. The FPCS method establishes a geologically adaptive graph model constrained by geodesic distance and gradient-sensitive weighting, employing a three-tier graph filter bank (low-pass, band-pass, and high-pass) to extract macroscopic morphology, interface gradients, and microscopic fracture features of rock layers. A dynamic gated fusion mechanism optimizes multi-level feature weights, significantly improving identification accuracy in lithological transition zones. Experimental results on five million test samples demonstrate an overall accuracy (OA) of 95.6% and a mean accuracy (mAcc) of 94.3%, representing improvements of 36.1% and 20.5%, respectively, over the PointNet model. These findings confirm the robust engineering applicability of the FPCS-based hierarchical multi-feature approach for point cloud lithology identification. Full article

Targeting the classification and evaluation of chemical flooding well groups in medium-thick sandstone reservoirs (single-layer thickness: 5–15 m), this study proposes a multi-level classification model based on decision trees. Through the comprehensive analysis of key static factors influencing chemical flooding efficiency, a four-tier classification index system was established, comprising: interlayer/baffle development frequency (Level 1), thickness-weighted permeability rush coefficient (Level 2), reservoir rhythm characteristics (Level 3), and pore-throat radius-based reservoir connectivity quality (Level 4) as its core components. The model innovatively transforms common reservoir physical parameters (porosity and permeability) into pore-throat radius parameters to enhance guidance for polymer molecular weight design, while employing a thickness-weighted permeability rush coefficient to simultaneously characterize heterogeneity impacts from both permeability and thickness variations. Unlike existing classification methods primarily designed for thin-interbedded reservoirs—which consider only connectivity or apply fuzzy mathematics-based normalization—this model specifically addresses medium-thick reservoirs’ unique challenges of interlayer development and intra-layer heterogeneity. Furthermore, its decision tree architecture clarifies classification logic and significantly reduces data preprocessing complexity. In terms of engineering practicality, the classification results are directly linked to well-group development bottlenecks, as validated in the J16 field application. By implementing customized chemical flooding formulations tailored to the study area, the production performance in the expansion zone achieved comprehensive improvement: daily oil output dropped from 332 tons to 243 tons, then recovered to 316 tons with sustained stabilization. Concurrently, recognizing that interlayer barriers were underdeveloped in certain well groups during production layer realignment, coupled with strong vertical heterogeneity posing polymer channeling risks, targeted profile modification and zonal injection were implemented prior to flooding conversion. This intervention elevated industrial replacement flooding production in the study area from 69 tons to 145 tons daily post-conversion. This framework provides a theoretical foundation for optimizing chemical flooding pilot well-group selection, scheme design, and dynamic adjustments, offering significant implications for enhancing oil recovery in medium-thick sandstone reservoirs through chemical flooding. Full article

Remote sensing objects often exhibit significant scale variations, high aspect ratios, and diverse orientations. The anisotropic spatial distribution of such objects’ features leads to the conflict between feature representation and boundary regression caused by the coupling of different attribute parameters: previous detection methods based on square-kernel convolution lack the overall perception of large-scale or slender objects due to the limited receptive field; if the receptive field is simply expanded, although more context information can be captured to help object perception, a large amount of background noise will be introduced, resulting in inaccurate feature extraction of remote sensing objects. Additionally, the extracted features face issues of feature conflict and discontinuous loss during parameter regression. Existing methods often neglect the holistic optimization of these aspects. To address these challenges, this paper proposes SODE-Net as a systematic solution. Specifically, we first design a multi-scale fusion and spatially orthogonal convolution (MSSO) module in the backbone network. Its multiple shapes of receptive fields can naturally capture the long-range dependence of the object without introducing too much background noise, thereby extracting more accurate target features. Secondly, we design a multi-level decoupled detection head, which decouples target classification, bounding-box position regression and bounding-box angle regression into three subtasks, effectively avoiding the coupling problem in parameter regression. At the same time, the phase-continuous encoding module is used in the angle regression branch, which converts the periodic angle value into a continuous cosine value, thus ensuring the stability of the loss value. Extensive experiments demonstrate that, compared to existing detection networks, our method achieves superior performance on four widely used remote sensing object datasets: DOTAv1.0, HRSC2016, UCAS-AOD, and DIOR-R. Full article

Cryo-electron microscopy single-particle analysis (cryo-EM SPA) has advanced three-dimensional protein structure determination, yet resolving intrinsically disordered proteins and regions (IDPs/IDRs) remains challenging due to conformational heterogeneity. This research evaluates cryo-EM’s capacity to map dynamic regions, assesses the adaptability of disorder prediction tools, and explores optimization strategies for dynamic structure prediction. Cryo-EM SPA datasets from 2000 to 2024 were categorized into different periods, forming a database integrating sequence data and disorder indices. Established prediction tools—AlphaFold2 (pLDDT), flDPnn, and IUPred—were evaluated for transferability, while a multi-level CLTC hybrid model (combining CNN, LSTM, Transformer, and CRF architectures) was developed to link local conformational fluctuations with global sequence contexts. Analyses revealed consistent advancements in average resolution and model counts over the past decade, although mapping disordered regions remained technically demanding. Both the adapted AlphaFold pLDDT and the CLTC model demonstrated efficacy in predicting structurally variable and poorly resolved regions. A subset of the cryo-EM missing residues exhibited intermediate conformational features, suggesting classification ambiguities potentially influenced by experimental conditions. These findings systematically outline the evolving capabilities of cryo-EM in resolving dynamic regions, benchmark the adaptability of computational tools, and introduce a hybrid model to enhance prediction accuracy. This study provides a framework for addressing conformational heterogeneity, contributing to methodological advancements in structural biology. Full article

With the growing adoption of electric vehicles, optimizing the user experience of charging infrastructure has become critical. However, extracting actionable insights from the vast number of user reviews remains a significant challenge, impeding demand-driven operational planning for charging stations and degrading the user experience. This study leverages three pre-trained language models to perform sentiment classification and multi-level topic identification on 168,129 user reviews from Beijing, facilitating a comprehensive understanding of user feedback. The experimental results reveal significant task-model specialization: RoBERTa-WWM excels in sentiment analysis (accuracy = 0.917) and fine-grained topic identification (Micro-F1 = 0.844), making it ideal for deep semantic extraction. Conversely, ELECTRA, after sufficient training, demonstrates a strong aptitude for coarse-grained topic summarization, highlighting its strength in high-level semantic generalization. Notably, the models offer capabilities beyond simple classification, including autonomous label normalization and the extraction of valuable information from comments with low information density. Furthermore, integrating textual and spatial analyses revealed striking patterns. We identified an urban–rural emotional gap—suburban users are more satisfied despite fewer facilities—and used geographically weighted regression (GWR) to quantify the spatial differences in the factors affecting user satisfaction in Beijing’s districts. We identified three types of areas requiring differentiated strategies, as follows: the northwestern region is highly sensitive to equipment quality, the central urban area has a complex relationship between supporting facilities and satisfaction, and the emerging adoption area is more sensitive to accessibility and price factors. These findings offer a data-driven framework for charging infrastructure planning, enabling operators to base decisions on real-world user feedback and tailor solutions to specific local contexts. Full article

As a national treasure, architectural heritage carries multiple value dimensions such as history, technology, art, and culture. With the increasing demand for architectural heritage protection and utilization, the traditional static digital model of architectural heritage based on geometric expression can no longer meet the practical application of multi-stage and multi-level scenarios. To this end, this paper proposes a value-chain-driven multi-level digital twin model of architectural heritage. Based on the three-stage logic of protection, management, and dissemination of value-chain classification, it integrates four types of models: geometry, physics, rules, and behavior. Combined with different hierarchical application levels, the digital model of architectural heritage is refined into a VCLOD (Value-Chain-Driven Level of Detail) detail hierarchy system to achieve a unified expression from spatial form restoration to intelligent response. Through the empirical application of three typical scenarios: the full-area guided tour of the Forbidden City, the exhibition curation of the central axis and the preventive protection of the Meridian Gate, the model shows the following specific results: (1) the efficiency of tourist guidance is improved through real-time personalized path planning; (2) the exhibition planning and visitor experience are improved through dynamic monitoring and interactive management of the exhibition environment; (3) the predictive analysis and preventive protection measures of structural safety are realized, effectively ensuring the structural safety of the Meridian Gate. The research results provide a theoretical basis and practical support for the systematic expression and intelligent evolution of digital twins of architectural heritage. Full article

This study aims to develop an enhanced YOLO algorithm to improve the ship detection performance of synthetic aperture radar (SAR) in complex marine environments. Current SAR ship detection methods face numerous challenges in complex sea conditions, including environmental interference, false detection, and multi-scale changes in detection targets. To address these issues, this study adopts a technical solution that combines multi-level feature fusion with a dynamic detection mechanism. First, a cross-stage partial dynamic channel transformer module (CSP_DTB) was designed, which combines the transformer architecture with a convolutional neural network to replace the last two C3k2 layers in the YOLOv11n main network, thereby enhancing the model’s feature extraction capabilities. Second, a general dynamic feature pyramid network (RepGFPN) was introduced to reconstruct the neck network architecture, enabling more efficient multi-scale feature fusion and information propagation. Additionally, a lightweight dynamic decoupled dual-alignment head (DYDDH) was constructed to enhance the collaborative performance of localization and classification tasks through task-specific feature decoupling. Experimental results show that the proposed DRGD-YOLO algorithm achieves significant performance improvements. On the HRSID dataset, the algorithm achieves an average precision (mAP50) of 93.1% at an IoU threshold of 0.50 and an mAP50–95 of 69.2% over the IoU threshold range of 0.50–0.95. Compared to the baseline YOLOv11n algorithm, the proposed method improves mAP50 and mAP50–95 by 3.3% and 4.6%, respectively. The proposed DRGD-YOLO algorithm not only significantly improves the accuracy and robustness of synthetic aperture radar (SAR) ship detection but also demonstrates broad application potential in fields such as maritime surveillance, fisheries management, and maritime safety monitoring, providing technical support for the development of intelligent marine monitoring technology. Full article

Mobile traffic classification serves as a fundamental component in network security systems. In recent years, pre-training methods have significantly advanced this field. However, as mobile traffic is typically mixed with third-party services, the deep integration of such shared services results in highly similar TCP flow characteristics across different applications. This makes it challenging for existing traffic classification methods to effectively identify mobile traffic. To address the challenge, we propose MS-PreTE, a two-phase pre-training framework for mobile traffic classification. MS-PreTE introduces a novel multi-level representation model to preserve traffic information from diverse perspectives and hierarchical levels. Furthermore, MS-PreTE incorporates a focal-attention mechanism to enhance the model’s capability in discerning subtle differences among similar traffic flows. Evaluations demonstrate that MS-PreTE achieves state-of-the-art performance on three mobile application datasets, boosting the F1 score for Cross-platform (iOS) to 99.34% (up by 2.1%), Cross-platform (Android) to 98.61% (up by 1.6%), and NUDT-Mobile-Traffic to 87.70% (up by 2.47%). Moreover, MS-PreTE exhibits strong generalization capabilities across four real-world traffic datasets. Full article

The accurate classification of staple crops is of great significance for scientifically promoting food production. Crop classification methods based on deep learning models or medium/low-resolution images have been applied in plain areas. However, existing methods perform poorly in complex mountainous scenes with rugged terrain, diverse planting structures, and fragmented farmland. This study introduces the Complex Scene Crop Classification U-Net+ (CSCCU+), designed to improve staple crop classification accuracy in intricate landscapes by integrating supplementary spectral information through an additional branch input. CSCCU+ employs a multi-branch architecture comprising three distinct pathways: the primary branch, auxiliary branch, and supplementary branch. The model utilizes a multi-level feature fusion architecture, including layered integration via the Shallow Feature Fusion (SFF) and Deep Feature Fusion (DFF) modules, alongside a balance parameter for adaptive feature importance calibration. This design optimizes feature learning and enhances model performance. Experimental validation using GaoFen-2 (GF-2) imagery in Xifeng County, Guizhou Province, China, involved a dataset of 2000 image patches (256 × 256 pixels) spanning seven categories. The method achieved corn and rice classification accuracies of 89.16% and 88.32%, respectively, with a mean intersection over union (mIoU) of 87.04%, outperforming comparative models (U-Net, DeeplabV3+, and CSCCU). This research paves the way for staple crop classification in complex land surfaces using high-resolution imagery, enabling accurate crop mapping and providing robust data support for smart agricultural applications. Full article

Background: The U.S. Medicare Annual Wellness Visit (AWV) offers a structured opportunity for cognitive screening and personalized prevention planning among older adults. Yet, implementation of AWVs, particularly for individuals with cognitive impairment, remains inconsistent across primary care or other diverse care settings. Methods: We conducted a scoping review using the Consolidated Framework for Implementation Research (CFIR) to explore multilevel factors influencing the implementation of the Medicare AWV’s cognitive screening component, with a focus on how these processes support the detection and management of cognitive impairment among older adults. We searched four databases and screened peer-reviewed studies published between 2011 and March 2025. Searches were conducted in Ovid MEDLINE, PubMed, EBSCOhost, and CINAHL databases. The initial search was completed on 3 January 2024 and updated monthly through 30 March 2025. All retrieved citations were imported into EndNote 21, where duplicates were removed. We screened titles and abstracts for relevance using the predefined inclusion criteria. Full-text articles were then reviewed and scored as either relevant (1) or not relevant (0). Discrepancies were resolved through consensus discussions. To assess the methodological quality of the included studies, we used the Joanna Briggs Institute critical appraisal tools appropriate to each study design. These tools evaluate rigor, trustworthiness, relevance, and risk of bias. We extracted the following data from each included study: Author(s), year, title, and journal; Study type and design; Data collection methods and setting; Sample size and population characteristics; Outcome measures; Intervention details (AWV delivery context); and Reported facilitators, barriers, and outcomes related to AWV implementation. The first two authors independently coded and synthesized all relevant data using a table created in Microsoft Excel. The CFIR guided our data analysis, thematizing our findings into facilitators and barriers across its five domains, viz: (1) Intervention Characteristics, (2) Outer Setting, (3) Inner Setting, (4) Characteristics of Individuals, and (5) Implementation Process. Results: Among 19 included studies, most used quantitative designs and secondary data. Our CFIR-based synthesis revealed that AWV implementation is shaped by interdependent factors across five domains. Key facilitators included AWV adaptability, Electronic Health Record (EHR) integration, team-based workflows, policy alignment (e.g., Accountable Care Organization participation), and provider confidence. Barriers included vague Centers for Medicare and Medicaid Services (CMS) guidance, limited reimbursement, staffing shortages, workflow misalignment, and provider discomfort with cognitive screening. Implementation strategies were often poorly defined or inconsistently applied. Conclusions: Effective AWV delivery for older adults with cognitive impairment requires more than sound policy and intervention design; it demands organizational readiness, structured implementation, and engaged providers. Tailored training, leadership support, and integrated infrastructure are essential. These insights are relevant not only for U.S. Medicare but also for global efforts to integrate dementia-sensitive care into primary health systems. Our study has a few limitations that should be acknowledged. First, our scoping review synthesized findings predominantly from quantitative studies, with only two mixed-method studies and no studies using strictly qualitative methodologies. Second, few studies disaggregated findings by race, ethnicity, or geography, reducing our ability to assess equity-related outcomes. Moreover, few studies provided sufficient detail on the specific cognitive screening instruments used or on the scope and delivery of educational materials for patients and caregivers, limiting generalizability and implementation insights. Third, grey literature and non-peer-reviewed sources were not included. Fourth, although CFIR provided a comprehensive analytic structure, some studies did not explicitly fit in with our implementation frameworks, which required subjective mapping of findings to CFIR domains and may have introduced classification bias. Additionally, although our review did not quantitatively stratify findings by year, we observed that studies from more recent years were more likely to emphasize implementation facilitators (e.g., use of templates, workflow integration), whereas earlier studies often highlighted systemic barriers such as time constraints and provider unfamiliarity with AWV components. Finally, while our review focused specifically on AWV implementation in the United States, we recognize the value of comparative analysis with international contexts. This work was supported by a grant from the National Institute on Aging, National Institutes of Health (Grant No. 1R01AG083102-01; PIs: Tzeng, Kuo, & Raji). Full article

Loess landslide disasters are influenced by a multitude of factors, including slope conditions, triggering mechanisms, and spatial attributes. Extracting these factors from unstructured geological texts is challenging due to nested entities, semantic ambiguity, and rare domain-specific terms. This study proposes a joint extraction framework guided by a domain ontology that categorizes six types of loess landslide influencing factors, including spatial relationships. The ontology facilitates conceptual classification and semi-automatic nested entity annotation, enabling the construction of a high-quality corpus with eight tag types. The model integrates a Soft-Lexicon mechanism that enhances character-level GloVe embeddings with explicit lexical features, including domain terms, part-of-speech tags, and word boundary indicators derived from a domain-specific lexicon. The resulting hybrid character-level representations are then fed into a BiLSTM-CRF architecture to jointly extract entities, attributes, and multi-level spatial and causal relationships. Extracted results are structured using a content-knowledge model to build a spatially enriched knowledge graph, supporting semantic queries and intelligent reasoning. Experimental results demonstrate improved performance over baseline methods, showcasing the framework’s effectiveness in geohazard information extraction and disaster risk analysis. Full article

This study presents the scientific and methodological foundation of Spain’s first national framework for the ethical management of community cat populations: the Action Plan for the Management of Community Cat Colonies (PACF), launched in 2025 under the mandate of Law 7/2023. This pioneering legislation introduces a standardized, nationwide obligation for trap–neuter–return (TNR)-based management of free-roaming cats, defined as animals living freely, territorially attached, and with limited socialization toward humans. The PACF aims to support municipalities in implementing this mandate through evidence-based strategies that integrate animal welfare, biodiversity protection, and public health objectives. Using standardized data submitted by 1128 municipalities (13.9% of Spain’s total), we estimated a baseline population of 1.81 million community cats distributed across 125,000 colonies. These data were stratified by municipal population size and applied to national census figures to generate a model-ready demographic structure. We then implemented a stochastic simulation using Vortex software to project long-term population dynamics over a 25-year horizon. The model integrated eight demographic–environmental scenarios defined by a combination of urban–rural classification and ecological reproductive potential based on photoperiod and winter temperature. Parameters included reproductive output, mortality, sterilization coverage, abandonment and adoption rates, stochastic catastrophic events, and territorial carrying capacity. Under current sterilization rates (~20%), our projections indicate that Spain’s community cat population could surpass 5 million individuals by 2050, saturating ecological and social thresholds within a decade. In contrast, a differentiated sterilization strategy aligned with territorial reproductive intensity (50% in most areas, 60–70% in high-pressure zones) achieves population stabilization by 2030 at approximately 1.5 million cats, followed by a gradual long-term decline. This scenario prioritizes feasibility while substantially reducing reproductive output, particularly in rural and high-intensity contexts. The PACF combines stratified demographic modeling with spatial sensitivity, offering a flexible framework adaptable to local conditions. It incorporates One Health principles and introduces tools for adaptive management, including digital monitoring platforms and standardized welfare protocols. While ecological impacts were not directly assessed, the proposed demographic stabilization is designed to mitigate population-driven risks to biodiversity and public health without relying on lethal control. By integrating legal mandates, stratified modeling, and realistic intervention goals, this study outlines a replicable and scalable framework for coordinated action across administrative levels. It exemplifies how national policy can be operationalized through data-driven, territorially sensitive planning tools. The findings support the strategic deployment of TNR-based programs across diverse municipal contexts, providing a model for other countries seeking to align animal welfare policy with ecological planning under a multi-level governance perspective. Full article

When emergency incidents, such as bridge damage, abruptly occur on highways and lead to traffic disruptions, the multidimensionality and complexity of driver behaviors present significant challenges to the design of effective emergency response mechanisms. This study introduces a multi-level collaborative emergency mechanism grounded in driver behavior characteristics, aiming to enhance both traffic safety and emergency response efficiency through hierarchical collaboration and dynamic optimization strategies. By capitalizing on human drivers’ perception and decision-making attributes, a driver behavior classification model is developed to quantitatively assess the risk response capabilities of distinct behavioral patterns (conservative, risk-taking, and conformist) under emergency scenarios. A multi-tiered collaborative framework, comprising an early warning layer, a guidance layer, and an interception layer, is devised to implement tailored emergency strategies. Additionally, a rear-end collision risk propagation model is constructed by integrating the risk field model with probabilistic risk assessment, enabling dynamic adjustments to interception range thresholds for precise and real-time emergency management. The efficacy of this mechanism is substantiated through empirical case studies, which underscore its capacity to substantially reduce the occurrence of secondary accidents and furnish scientific evidence and technical underpinnings for emergency management pertaining to highway bridge damage. Full article