Search Results (118)

The rapid development of generative artificial intelligence has paved the way for a new approach to reproduce and intelligently generate traditional patterns digitally. This paper focuses on the traditional Chinese phoenix pattern and constructs a “Phoenix Pattern Multidimensional Cultural Gene Model” based on the grounded theory. It summarises seven semantic dimensions covering composition pattern, pixel configuration, colour system, media technology, semantic implication, theme context, and application scenario and divides them into explicit and implicit cultural genes. The study further proposes a control mechanism of “semantic label–prompt–image generation”, constructs a cross-platform prompt structure system suitable for Midjourney and Dreamina AI, and completes 28 groups of prompt combinations and six rounds of iterative experiments. The analysis of the results from 64 user questionnaires and 10 expert ratings reveals that Dreamina AI excels in cultural semantic restoration and context recognition. In contrast, Midjourney has an advantage in composition coordination and aesthetic consistency. Overall, the study verified the effectiveness of the cultural gene model in generating AIGC control. It proposed a framework for generating innovative traditional patterns, providing a theoretical basis and practical support for the intelligent expression of cultural heritage. Full article

This review explores optical technologies for non-invasive glucose monitoring (NIGM) using aqueous humor (AH) as media, addressing the limitations of traditional invasive methods in diabetes management. It analyzes key techniques such as Raman spectroscopy, polarimetry, and mid- and near-infrared spectral methods, highlighting their respective challenges, alongside emerging hybrid approaches like photoacoustic spectroscopy and optical coherence tomography. Crucially, the practical realization of these optical methods for portable NIGM hinges on advanced instrumentation. Therefore, this review also details progress in compact NIR spectrometers. While conventional systems often lack suitability, significant advancements in on-chip technologies—including miniaturized dispersive spectrometers and various on-chip Fourier transform systems (e.g., spatial heterodyne, stationary wave integral, and temporally modulated FT systems)—utilizing integration platforms like SOI and SiN are promising. Such innovations offer the potential for high spectral resolution, large bandwidth, and miniaturization, which are essential for developing practical AH-based NIGM systems to improve diabetes care. Full article

Background: Exploring new strategies to improve type 2 diabetes mellitus (T2DM) is one of the frontier hotspots in the field of healthy food. Flavonoid–metal complexes have become one of the research hotspots in the field of health foods due to their unique structural and functional properties. Methods: In this study, the effect of hesperetin–copper(II) complex [Hsp–Cu(II)] on the gut microbiota of mice with T2DM was investigated by the 16S rRNA high-throughput sequencing. Results: The analyses of α and β diversity indicated that the richness and diversity of gut microbiota in the T2DM mice decreased and the community structure was significantly different from the normal mice. Hsp–Cu(II) increased the abundances of the beneficial bacteria (Lactobacillus, Ligilactobacillus, Romboutsia, Faecalibaculum, and Dubosiella), and decreased the amounts of the harmful bacteria (Desulfobacterota, Corynebacterium, and Desulfovibrio) and the ratio of Firmicutes/Bacteroidetes (from 44.5 to 5.8) in the T2DM mice, which was beneficial for regulating the composition of intestinal microbiota. The linear discriminant analysis effect size analysis showed that the intervention of Hsp–Cu(II) made the short-chain fatty acid (SCFA) producers (o_Lachnospirales, f_Lachnospiraceae, g_Faecalibaculum, g_Romboutsia, and g_Turicibacter) and the lactic acid bacteria producers (f_Lactobacillaceae and o_Lactobacillales) highly enriched, and the production of its metabolite SCFAs (acetic acid, propionic acid, butyric acid, and valeric acid) were increased in a dose-dependent manner, promoting the SCFA metabolism. Conclusions: Hsp–Cu(II) may improve glucose metabolic disorders and alleviate T2DM by modulating gut microbiota composition, promoting probiotics proliferation and SCFAs production, restoring intestinal barrier integrity, and suppressing local inflammation. These research findings may provide a theoretical basis for developing Hsp–Cu(II) as a new hypoglycemic nutritional supplement, and offer new ideas for the dietary food nutritional regulation to alleviate T2DM. Full article

To accurately monitor the operational state of Internet Data Centers (IDCs) and fulfill integrated management objectives, this paper introduces a bidirectional cross-attention LSTM–Informer with uncertainty-aware multi-task learning framework (BiCA-LI) for time series analysis. The architecture employs dual-branch temporal encoders—long short-term memory (LSTM) and Informer—to extract local transient dynamics and global long-term dependencies, respectively. A bidirectional cross-attention module is subsequently designed to synergistically fuse multi-scale temporal representations. Finally, task-specific regression and classification heads generate predictive outputs and anomaly detection results, while an uncertainty-aware dynamic loss weighting strategy adaptively balances task-specific gradients during training. Experimental results validate BiCA-LI’s superior performance across dual objectives. In regression tasks, it achieves an MAE of 0.086, MSE of 0.014, and RMSE of 0.117. For classification, the model attains 99.5% accuracy, 100% precision, and an AUC score of 0.950, demonstrating substantial improvements over standalone LSTM and Informer baselines. The dual-encoder design, coupled with cross-modal attention fusion and gradient-aware loss optimization, enables robust joint modeling of heterogeneous temporal patterns. This methodology establishes a scalable paradigm for intelligent IDC operations, enabling real-time anomaly mitigation and resource orchestration in energy-intensive infrastructures. Full article

Accurate real-time icing grid fields are critical for preventing ice-related disasters during winter and protecting property. These fields are essential for both mapping ice distribution and predicting icing using physical models combined with numerical weather prediction systems. However, developing precise real-time icing grids is challenging due to the uneven distribution of monitoring stations, data confidentiality restrictions, and the limitations of existing interpolation methods. In this study, we propose a new approach for constructing real-time icing grid fields using 1339 online terminal monitoring datasets provided by the China Southern Power Grid Research Institute Co., Ltd. (CSPGRI) during the winter of 2023. Our method integrates static geographic information, dynamic meteorological factors, and ice_kriging values derived from parameter-optimized Empirical Bayesian Kriging Interpolation (EBKI) to create a spatiotemporally matched, multi-source fused icing thickness grid dataset. We applied five machine learning algorithms—Random Forest, XGBoost, LightGBM, Stacking, and Convolutional Neural Network Transformers (CNNT)—and evaluated their performance using six metrics: R, RMSE, CSI, MAR, FAR, and fbias, on both validation and testing sets. The stacking model performed best, achieving an R-value of 0.634 (0.893), RMSE of 3.424 mm (2.834 mm), CSI of 0.514 (0.774), MAR of 0.309 (0.091), FAR of 0.332 (0.161), and fbias of 1.034 (1.084), respectively, when comparing predicted icing values with actual measurements on pylons. Additionally, we employed the SHAP model to provide a physical interpretation of the stacking model, confirming the independence of selected features. Meteorological factors such as relative humidity (RH), 10 m wind speed (WS₁₀), 2 m temperature (T₂), and precipitation (PRE) demonstrated a range of positive and negative contributions consistent with the observed growth of icing. Thus, our multi-source remote-sensing data-fusion approach, combined with the stacking model, offers a highly accurate and interpretable solution for generating real-time icing grid fields. Full article

Previously, we developed a porcine bronchial epithelial cell line designated as PBE cells and demonstrated that this cell line possesses functional Toll-like receptor 3 (TLR3), triggering the expressions of interferons (IFNs), antiviral factors, and inflammatory cytokines after its stimulation with the synthetic double-stranded ARN poly(I:C). In this work, we aimed to further characterize the PBE cell line as a reliable in vitro model for investigating swine influenza virus (SIV) infection and immunity. We evaluated the capacity of two SIV subtypes, H1N1 and H3N2, to replicate and induce cytopathic effects in PBE cells and to modulate the expressions of IFNs, antiviral factors, inflammatory cytokines, and negative regulators of the TLR signaling. We demonstrated that PBE cells are susceptible to both H1N1 and H3N2. SIV infected PBE cells inducing notable cytopathic effects as shown by the alteration of transepithelial electrical resistance (TEER) and cilia. Both SIV subtypes replicated in PBE cells in similar proportion and altered TEER values in comparable magnitudes. However, SIV H3N2 induced higher alterations of cilia than H1N1. SIV infection induced changes in all the immune factors evaluated in PBE cells. We detected quantitative differences when the subtypes H1N1 and H3N2 were compared. The fold expression changes of IFN-β, Mx1, Mx2, IFITM1, OAS1, OAS2, and OASL were higher in PBE cells infected with H3N2 than in cells challenged with H1N1. In addition, although both subtypes stimulated IL-8 expression, only the H3N2 induced IL-6 in infected PBE cells. SIV H1N1 and H3N2 also upregulated the expressions of the negative regulators A20, BCL-3, and MKP-1, while only H1N1 increased SIGIRR and Tollip. Immortalized respiratory cell lines from pigs can be useful in vitro systems for the study of viral infections and immune responses. These studies are of importance in the context of influenza infections not only for the agricultural field because pigs are natural hosts of these viruses but also because these animals serve as intermediate reservoirs of viruses that can threaten humans’ health. We demonstrated here that the PBE cell line can be a useful in vitro model to study SIV infection and immunity. Full article

Continental shale oil fracturing dynamics are governed by interactions between hydraulic fractures and pre-existing natural fractures. This study establishes a fluid–solid coupling model using globally embedded cohesive elements to simulate fracture propagation in naturally fractured reservoirs. Key factors affecting fracture network complexity were quantified: (1) Weakly cemented natural fractures (bond strength coefficient <0.5) promote 23% higher fracture tortuosity compared to strongly cemented formations. (2) Optimal horizontal stress differentials (Δσ = 8–10 MPa) balance fracture length (increased by 35–40%) and branching complexity. (3) Injection rate elevation from 0.06 to 0.132 m³/min enhances the stimulated volume by 90% through improved fracture dimensions. The findings provide mechanistic insights for optimizing fracture network complexity in shale reservoirs. Full article

Hopanoids are a series of important lipid biomarkers in the bacterial cellular membranes that are found ubiquitously in different spatial and temporal environments. Squalene-hopane cyclase, a key and prerequisite molecular component of the hopanoid biosynthesis pathway, is encoded by the sqhC gene. To investigate the composition, niche, and distribution of microbial sqhC-containing communities, we analyzed hopanoid producer data and environmental parameters across different ecosystems on the basis of sequencing reads of peat samples from increasing gradient depths across peatland profile C in the Dajiuhu Peatland, as well as data collected from available published papers. The results indicated that the acidic Dajiuhu Peatland harbored mainly Acidobacteria (59.16%) among its sqhC-containing groups. The main composition of hopanoid producers in the peatland was different from that in other ecosystems, with Alphaproteobacteria found in soil (37.78%), cave (48.21%), hypersaline lagoon (34.04%), and marine (32.59%) ecosystems; Betaproteobacteria, Gammaproteobacteria, and Deltaproteobacteria found in reef (100%), acid mine drainage (55.00%), and estuary, mangrove, and harbor (39.66%) ecosystems; and an unknown cluster found in freshwater (29.43%) and hot spring (89.58%) ecosystems. Compared with other phyla or sub-phyla, Alphaproteobacteria, Betaproteobacteria, and Gammaproteobacteria were the most widespread, occurring in eight ecosystems. Peatland was significantly separated from the other nine ecosystem modules in the occurrence network, and the marine ecosystem had the greatest impact on the eco-network of sqhC microbes. An RDA indicated that pH, DO, salinity, and TOC had significant impacts on sqhC-containing microbial communities across the different ecosystems. Our results will be helpful to understanding the diversity, composition, and distribution of the sqhC community and its response to multiple environmental factors across different ecosystems. Full article

Understanding the characteristics of groundwater chemistry is essential for water resource development and utilization. However, few studies have focused on the chemical evolution processes of shallow groundwater in typical areas of the Huaibei Plain. We analyzed 28 water samples from the study area using hydrogeochemical mapping, multivariate statistical analysis, and other approaches. The study found that the hydrogeochemical facies of groundwater are mainly HCO₃-Ca·Mg (64.3%), mixed SO₄·Cl-Ca·Mg, and SO₄·Cl-Na. The hydrochemical composition is primarily controlled by natural water–rock interactions, including carbonate weathering and cation exchange processes. Correlation analysis and principal component analysis (PCA) revealed that mineral dissolution was the predominant source of Na⁺, Mg²⁺, Cl⁻, and SO₄²⁻ in shallow groundwater, significantly contributing to total dissolved solids (TDS) accumulation. Hierarchical cluster analysis (HCA) identified three characteristic zones: (1) agricultural/urban-influenced areas, (2) high-F⁻/low-hardness zones, and (3) nitrate-contaminated regions. These findings provide critical insights for assessing the geochemical status of groundwater in the Huaibei Plain and formulating targeted resource management strategies. Full article

The Qinghai–Tibet Plateau (QTP), one of China’s most snow-rich regions, has an extremely fragile ecosystem, with drought being the primary driver of ecological degradation. Given that the water resources in this region predominantly exist in the form of snow, high-spatiotemporal-resolution snow mapping is essential for understanding snow distribution and managing snow water resources effectively. However, although FY-4A/AGRI is capable of obtaining wide-area remote sensing data, only the first to third bands have a resolution of 1 km, which greatly limits its ability to produce high-resolution snow maps. This study proposes U-SwinFusionNet (USFNet), a deep learning-based snow cover retrieval algorithm that leverages the multi-scale advantages of FY-4A/AGRI remote sensing data in the shortwave infrared and visible bands. By integrating 1 km and 2 km resolution remote sensing imagery with auxiliary terrain information, USFNet effectively enhances snow cover mapping accuracy. The proposed model innovatively combines Swin Transformer and convolutional neural networks (CNNs) to capture both global contextual information and local spatial details. Additionally, an Attention Feature Fusion Module (AFFM) is introduced to align and integrate features from different modalities through an efficient attention mechanism, while the Feature Complementation Module (FCM) facilitates interactions between the encoded and decoded features. As a result, USFNet produces snow cover maps with a spatial resolution of 1 km. Experimental comparisons with Artificial Neural Networks (ANNs), Random Forest (RF), U-Net, and ResNet-FSC demonstrate that USFNet exhibits superior robustness, enhanced snow cover continuity, and lower error rates. The model achieves a correlation coefficient of 0.9126 and an R² of 0.7072. Compared to the MOD10A1 snow product, USFNet demonstrates an improved sensitivity to fragmented and low-snow-cover areas while ensuring more natural snow boundary transitions. Full article

In situ vertical load field tests were carried out on two bored piles used in the Chizhou Highway Bridge across the Yangtze River, both of which were rooted piles. Based on the test results, such as those on the relationship between the load and settlement, axial force distribution, and the relationship between shaft friction and pile–soil relative displacement, the vertical load transfer mechanics of the rooted piles were analyzed. The results showed that the load-carrying curves of the rooted piles vary gradually and also that the rooted piles exhibit the bearing characteristics of friction piles because the loads at the pile tips are less than 15% of the total bearing capacity of the piles. The slope of the axial force distribution curve of the rooted piles first increased at the upper interface and then decreased at the lower interface of the root-reinforced zone. The axial force of the rooted piles decreased faster in soil layers where the piles had roots, which can be explained by the fact that roots share the vertical load with piles and that roots improve the bearing properties of piles. Considering the shaft and end resistance of the roots on the piles, the relationship between load and settlement of the rooted piles was calculated by a three-line model based on the load transfer method. The results calculated from the model were in good agreement with the results from the tests. The results from the tests could inform the design and analysis of rooted piles. Full article

Piles in marine environments are subjected to various loads of differing magnitudes and directions, and their long-term stability has attracted much attention. Most research focuses on lateral cyclic loading; there are few full-scale tests that consider the effects of cyclic loading at different inclined angles. A long-term inclined cyclic loading strategy was used to carry out laboratory tests to study different inclined angles on the pile. The results show that a smaller inclined angle (θ_L) or a larger pile–soil relative stiffness (T/L) results in wider and deeper sediment subsidence after 10,000 cycles. As θ_L increases from 0° to 80°, the peak displacement at the pile head during the first load decreases, while the accumulated displacement initially decreases and then increases. For slender piles, the normalized inclined cyclic loading stiffness (k_lN/k_l1) and unloading stiffness (k_uN/k_u1) first decrease and then increase. For semi-rigid piles, both k_lN/k_l1 and k_uN/k_u1 gradually decrease. On the other hand, as θ_L increases, k_lN/k_l1 and k_uN/k_u1 increased more sharply in the initial stage, with a quicker transition from rapid growth to stability. At θ_L = 80°, peak values are reached early during the initial loading phase. Based on this, prediction formulas for inclined cyclic cumulative displacement, loading stiffness, and unloading stiffness were established and verified. Full article

In this paper, 316 L stainless steel deposited samples were fabricated by direct layer deposition (DED) using both continuous-wave (CW) and pulsed-wave (PW) laser modes. Effects of laser modes on residual stress of deposited samples were investigated. On this basis, a mathematical model of thermal stress evolution during DED was established for the first time based on the energy equation. The variation law of thermal stress on the top of the substrate under multi-material and multi-process conditions was qualitatively predicted and the corresponding residual stress reduction mechanism has been studied using this model. Meanwhile, in situ thermal strain evolution is used to prove the correctness of the mathematical model. This model lays the foundation for predicting the thermal stress evolution and the magnitude of the residual stress of deposited samples under multi-material and process conditions during DED. Full article

Anterior cruciate ligament (ACL) rupture and patellar dislocation (PD) are common knee injuries. Dual-energy X-ray absorptiometry (DXA) and computed tomography (CT) are widely used clinical diagnostic tools. The aim was to investigate the characteristics of knee bone mineral density (BMD) in patients with ACL rupture and PD and to explore the relationship between BMD and Hounsfield unit (HU) values. This prospective cross-sectional study included 32 ACL rupture and 32 PD patients assessed via DXA and CT. BMD and CT measurements were taken from regions of interest in the femoral and tibial condyles. Statistical analyses included t-tests and mixed-effects models. The results showed that BMD in the PD group was significantly lower than in the ACL group (p < 0.05). The HU values of cortical bone in the femur and tibia differed significantly between the ACL group and the PD group (p < 0.05). The BMD of the femur and tibia showed significant correlations with the HU values of cancellous bone and cortical bone (p < 0.05). The conclusion was that PD patients had lower BMD and HU values in the femur and tibia compared to patients with ACL ruptures, particularly in the cortical bone of the tibia, and there was a strong correlation between HU value and BMD. Full article

Ecological integrity is fundamental to human survival and development. However, rapid urbanization and population growth have significantly disrupted ecosystems. Despite the focus of the International Geosphere-Biosphere Programme (IGBP) on terrestrial ecosystems and land use/cover changes, existing ecological indices, such as the Remote Sensing Ecological Index (RSEI), have limitations, including an overreliance on single indicators and inability to fully encapsulate the ecological conditions of urban areas. This study addresses these gaps by proposing a Deep-learning-based Remote Sensing Ecological Index (DRSEI) that integrates human economic activities and leverages an autoencoder neural network with long short-term memory (LSTM) modules to account for nonlinearity in ecological quality assessments. The DRSEI model utilizes multi-temporal remote sensing data from the Landsat series, WorldPop, and NPP-VIIRS and was applied to evaluate the ecological conditions of Hubei Province, China, over the past two decades. The key findings indicate that ecological environmental quality gradually improved, particularly from 2000 to 2010, with the rate of improvement subsequently slowing. The DRSEI outperformed the traditional RSEI and had a significantly higher Pearson correlation coefficient than the Ecological Index (EI), thus demonstrating enhanced accuracy and predictive performance. This study presents an innovative approach to ecological assessment that offers a more comprehensive, accurate, and nuanced understanding of ecological changes over time. Integrating socioeconomic factors with deep learning techniques contributes significantly to the field and has implications for ecological risk control and sustainable development. Full article