Search Results (2,758)

The green-making stage is crucial for forming the distinctive aroma and flavor of Tieguanyin tea. Current green-making stage recognition relies on tea makers’ sensory experience, which is labor-intensive and time-consuming. To address these issues, this paper proposes a lightweight automatic recognition model named T-GSR for the accurate and objective identification of Tieguanyin tea green-making stages. First, an extensive set of Tieguanyin tea images at different green-making stages was collected. Subsequently, preprocessing techniques, i.e., multi-color-space fusion and morphological filtering, were applied to enhance the representation of target tea features. Furthermore, three targeted improvements were implemented based on the MobileNet V3 backbone network: (1) an adaptive residual branch was introduced to strengthen feature propagation; (2) the Rectified Linear Unit (ReLU) activation function was replaced with the Gaussian Error Linear Unit (GELU) to improve gradient propagation efficiency; and (3) an Improved Coordinate Attention (ICA) mechanism was adopted to replace the original Squeeze-and-Excitation (SE) module, enabling more accurate capture of complex tea features. Experimental results demonstrate that the T-GSR model outperforms the original MobileNet V3 in both classification performance and model complexity, achieving a recognition accuracy of 93.38%, an F1-score of 93.33%, with only 3.025 M parameters and 0.242 G FLOPs. The proposed model offers an effective solution for the intelligent recognition of Tieguanyin tea green-making stages, facilitating online monitoring and supporting automated tea production. Full article

The Soil and Plant Analyzer Development (SPAD) value is a reliable proxy for chlorophyll, yet conventional field measurement remains labor-intensive and spatially limited. Current remote sensing inversion models typically depend on costly multispectral sensors and rarely account for phenological changes, restricting their applicability across orchards and seasons. To overcome these limitations, this study introduces a stage-aware and low-cost SPAD inversion framework for jujube trees, integrating multi-source data fusion and an optimized ensemble model. A two-year experiment (2023–2024) combined UAV multispectral vegetation indices (VI) with RGB-derived color indices (CI) across leaf expansion, flowering, and fruit-setting stages. Rather than using static features, stage-specific predictors were systematically identified through a hybrid selection mechanism combining Random Forest Cumulative Feature Importance (RF-CFI), Recursive Feature Elimination (RFE), and F-tests. Building on these tailored features, XGBoost, decision tree (DT), CatBoost, and an Optimized Integrated Architecture (OIA) were developed, with all hyperparameters globally tuned using a genetic algorithm (GA). The RFI-CFI-OIA-GA model delivered superior accuracy (R² = 0.758–0.828; MSE = 0.214–2.593; MAPE = 0.01–0.045 in 2024) in the training dataset, and robust cross-year transferability (R² = 0.541–0.608; MSE = 0.698–5.139; MAPE = 0.015–0.058 in 2023). These results demonstrate that incorporating phenological perception into multi-source data fusion substantially reduces interference and enhances generalizability, providing a scalable and reusable strategy for precision orchard management and spatiotemporal SPAD mapping. Full article

Cassava remains a critical food-security crop across Africa and Southeast Asia but is highly vulnerable to diseases such as cassava mosaic disease (CMD) and cassava brown streak disease (CBSD). Traditional diagnostic approaches are slow, labor-intensive, and inconsistent under field conditions. This review synthesizes current advances in combining unmanned aerial vehicles (UAVs) with deep learning (DL) to enable scalable, data-driven cassava disease detection. It examines UAV platforms, sensor technologies, flight protocols, image preprocessing pipelines, DL architectures, and existing datasets, and it evaluates how these components interact within UAV–DL disease-monitoring frameworks. The review also compares model performance across convolutional neural network-based and Transformer-based architectures, highlighting metrics such as accuracy, recall, F1-score, inference speed, and deployment feasibility. Persistent challenges—such as limited UAV-acquired datasets, annotation inconsistencies, geographic model bias, and inadequate real-time deployment—are identified and discussed. Finally, the paper proposes a structured research agenda including lightweight edge-deployable models, UAV-ready benchmarking protocols, and multimodal data fusion. This review provides a consolidated reference for researchers and practitioners seeking to develop practical and scalable cassava-disease detection systems. Full article

With the rapid development of the power Internet of Things (IoT), the traditional manual inspection mode can no longer meet the growing demand for power equipment inspection. Unmanned aerial vehicle (UAV) intelligent inspection technology, with its efficient and flexible features, has become the mainstream solution. The rapid development of computer vision and deep learning (DL) has significantly improved the accuracy and efficiency of UAV intelligent inspection systems for power equipment. However, mainstream deep learning models have complex structures, and manual design is time-consuming and labor-intensive. In addition, the images collected during the power inspection process by UAVs have problems such as complex backgrounds, uneven lighting, and significant differences in object sizes, which require expert DL domain knowledge and many trial-and-error experiments to design models suitable for application scenarios involving power inspection with UAVs. In response to these difficult problems, evolutionary computation (EC) technology has demonstrated unique advantages in simulating the natural evolutionary process. This technology can independently design lightweight and high-precision deep learning models by automatically optimizing the network structure and hyperparameters. Therefore, this review summarizes the development of evolutionary deep learning (EDL) technology and provides a reference for applying EDL in object detection models used in UAV intelligent power inspection systems. First, the application status of DL-based object detection models in power inspection is reviewed. Then, how EDL technology improves the performance of the models in challenging scenarios such as complex terrain and extreme weather is analyzed by optimizing the network architecture. Finally, the challenges and future research directions of EDL technology in the field of UAV power inspection are discussed, including key issues such as improving the environmental adaptability of the model and reducing computing energy consumption, providing theoretical references for promoting the development of UAV power inspection technology to a higher level. Full article

The closure of rural maternity units in hospitals across the United States contributes to health inequities; however, little is known about the effects of reopening maternity services in this context. We conducted a mixed-methods study to characterize labor and delivery outcomes and patient experiences associated with the reopening of a rural Level 1 Maternity Care Center (MCC) at a critical access hospital. We compared clinical outcomes and distance to care for patients who gave birth at the rural MCC in the three years after its opening with outcomes from a similar low-risk and geographically located sample who gave birth at a large suburban academic medical center in the same hospital system in the three years before the MCC reopened. We also conducted in-depth interviews with patients who gave birth at the MCC. Labor and delivery outcomes were similar across both groups, with significantly more care provided by family physicians and midwives and lower neonatal intensive care unit use at the MCC. The opening of the MCC halved the distance patients traveled to give birth, and patients reported high rates of satisfaction. Rural maternity care centers can improve access to quality care closer to home using a resource-appropriate model. Full article

Cerebrovascular diseases are a leading cause of global mortality, underscoring the need for objective and quantitative 3D visualization of cerebral vasculature from dynamic imaging modalities. Conventional analysis is often labor-intensive, subjective, and prone to errors due to image noise and subtraction artifacts. This study tackles the challenge of achieving fast and accurate volumetric reconstruction from angiography sequences. We propose a multi-stage pipeline that begins with image restoration to enhance input quality, followed by neural segmentation to extract vascular structures. Camera poses and sparse geometry are estimated through Structure-from-Motion, and these reconstructions are refined by leveraging the segmentation maps to isolate vessel-specific features. The resulting data are then used to initialize and optimize a 3D Gaussian Splatting model, enabling anatomically precise representation of cerebral vasculature. The integration of deep neural segmentation priors with explicit geometric initialization yields highly detailed 3D reconstructions of cerebral angiography. The resulting models leverage the computational efficiency of 3D Gaussian Splatting, achieving near-real-time rendering performance competitive with state-of-the-art reconstruction methods. The segmentation of brain vessels using nnU-Net and our trained model achieved an accuracy of 84.21%, highlighting the improvement in the performance of the proposed approach. Overall, our pipeline significantly improves both the efficiency and accuracy of volumetric cerebral vasculature reconstruction, providing a robust foundation for quantitative clinical analysis and enhanced guidance during endovascular procedures. Full article

To address the inefficiency and high labor intensity associated with traditional manual mussel seedling unloading, this study proposes an automated traction-rope mussel unloading machine. This study focuses on the thick-shelled mussel (Mytilus coruscus) as the research subject. Furthermore, the key mussel unloading processes were simulated using the EDEM software to analyze mechanical interactions during detachment. A breakable mussel discrete element model was developed, and its Bonding V2 model parameters were systematically calibrated. Using the ultimate crushing displacement (2.25 mm) and ultimate crushing load (552 N) as response variables, the model was optimized through a sequential experimental design comprising Plackett–Burman screening, the steepest ascent method, and the Box–Behnken response surface methodology. The results demonstrate that the optimal parameter combination consists of unit area normal stiffness (2.48 × 10¹¹ N/m³), unit area tangential stiffness (3.80 × 10⁸ N/m³), critical normal stress (3.15 × 10⁶ Pa), critical tangential stress (2.90 × 10⁷ Pa), and the contact radius (1.60 mm). The model’s accuracy was validated through integrated discrete element simulations and prototype testing. The equipment achieves an exceptionally low mussel damage rate of only 1.2%, effectively meeting the operational requirements for mussel unloading. This study provides both theoretical foundations and practical insights for the design of mechanized mussel unloading systems in China. Full article

Pollen identification is critical for melissopalynology (honey authentication), ecological monitoring, and allergen tracking, yet manual microscopic analysis remains labor-intensive, subjective, and error-prone when multiple grains overlap in realistic samples. Existing automated approaches often fail to address multi-grain scenarios or lack systematic comparison across classical and deep learning paradigms, limiting their practical deployment. This study proposes a subinstance-based classification framework combining YOLOv12n object detection for grain isolation, independent classification via classical machine learning (ML), convolutional neural networks (CNNs), or Vision Transformers (ViTs), and majority voting aggregation. Five classical classifiers with systematic feature selection, three CNN architectures (ResNet50, EfficientNet-B0, ConvNeXt-Tiny), and three ViT variants (ViT-B/16, ViT-B/32, ViT-L/16) are evaluated on four datasets (full images vs. isolated grains; raw vs. CLAHE-preprocessed) for four berry pollen species (Ribes nigrum, Ribes uva-crispa, Lonicera caerulea, and Amelanchier alnifolia). Stratified image-level splits ensure no data leakage, and explainable AI techniques (SHAP, Grad-CAM++, and gradient saliency) validate biological interpretability across all paradigms. Results demonstrate that grain isolation substantially improves classical ML performance (F1 from 0.83 to 0.91 on full images to 0.96–0.99 on isolated grains, +8–13 percentage points), while deep learning excels on both levels (CNNs: F1 = 1.000 on full images with CLAHE; ViTs: F1 = 0.99). At the instance level, all paradigms converge to near-perfect discrimination (F1 ≥ 0.96), indicating sufficient capture of morphological information. Majority voting aggregation provides +3–5% gains for classical methods but only +0.3–4.8% for deep models already near saturation. Explainable AI analysis confirms that models rely on biologically meaningful cues: blue channel moments and texture features for classical ML (SHAP), grain boundaries and exine ornamentation for CNNs (Grad-CAM++), and distributed attention across grain structures for ViTs (gradient saliency). Qualitative validation on 211 mixed-pollen images confirms robust generalization to realistic multi-species samples. The proposed framework (YOLOv12n + SVC/ResNet50 + majority voting) is practical for deployment in honey authentication, ecological surveys, and fine-grained biological image analysis. Full article

Glaucoma is a leading cause of irreversible blindness worldwide, with its asymptomatic progression highlighting the urgent need for early, minimally invasive biomarkers. Exosomes derived from the aqueous humor (AH) have emerged as promising candidates, as they carry proteins, nucleic acids, and lipids that reflect the physiological and pathological state of ocular tissues such as the trabecular meshwork and ciliary body. However, their low abundance, nanoscale size, and the limited volume of AH complicate detection and characterization. Conventional methods, including Western blotting, PCR or mass spectrometry, are labor-intensive, time-consuming, and often incompatible with microliter-scale samples. Electrochemical biosensors offer a highly sensitive, rapid, and low-volume alternative, enabling the detection of exosomal surface markers and internal cargos such as microRNAs, proteins, and lipids. Recent advances in nanomaterial-enhanced electrodes, microfluidic integration, enzyme- and nanozyme-mediated signal amplification, and ratiometric detection strategies have significantly improved sensitivity, selectivity, and multiplexing capabilities. While most studies focus on blood or serum, these platforms hold great potential for AH-derived exosome analysis, supporting early-stage glaucoma diagnosis, monitoring of disease progression, and evaluation of therapeutic responses. Continued development of miniaturized, point-of-care electrochemical biosensors could facilitate clinically viable, noninvasive exosome-based diagnostics for glaucoma. Full article

Artificial intelligence (AI) is transforming numerous sectors, and its integration into the legal domain holds significant potential for automating labor-intensive tasks, enhancing judicial decision-making, and improving overall system efficiency. This study introduces an AI-powered model, named the Legal Judgment Prediction Ensemble (LJPE), which is designed to predict legal case outcomes by leveraging historical judicial data. By using natural language processing (NLP) techniques, feature engineering, and a complex two-level stacking ensemble, the LJPE model has better predictive accuracy at 94.68% compared to modern legal language and conventional machine learning models. Moreover, the findings underline the predictive strength of textual features obtained from case facts, vote margins, and legal-specific features. This study offers a solid technical solution for predicting legal judgments for the responsible use of the model, helping to create a more efficient, transparent, and fair legal system. Full article

At present, conventional cold storage facilities in China are poorly suited to on-farm storage demands for agricultural produce, mainly due to their large spatial requirements, complex and labor-intensive installation procedures, limited portability, and insufficient coverage in rural areas. These limitations significantly contribute to post-harvest losses of perishable crops such as cherry tomatoes. To address this challenge, the present study proposes a compact and temporary cold storage system—gas-inflated membrane cold storage (GIMCS)—which exploits the inherent safety, cost-effectiveness, ease of deployment, and adaptability of inflatable membrane structures. A series of mechanical performance tests, including tensile strength, pressure resistance, and burst tests, were conducted on PA/PE (Polyamide/Polyethylene) composite membranes. The optimal configuration was identified as a membrane thickness of 70 μm, a gas column width of 2 cm, and a PA/PE composition ratio of 35%/65%. Thermal performance evaluations further revealed that filling the inflatable structure with 100% CO₂ yielded the most effective insulation. Through structural optimization, a cotton-filled gas-inflated membrane cold storage system (CF-GIMCS) incorporating a dual insulation strategy—combining intra-membrane and extra-membrane insulation—was developed. This multilayer configuration significantly reduced conductive and convective heat transfer, resulting in enhanced thermal performance. A comparative evaluation between GIMCS and a conventional cold storage system of equivalent capacity was conducted over a 15-day storage period, considering construction cost, temperature uniformity, and fruit preservation quality. The results showed that the construction cost of GIMCS was only 38% of that of conventional cold storage. The internal temperature distribution of GIMCS was highly uniform, with a maximum horizontal temperature difference of 1.4 °C, demonstrating thermal stability comparable to conventional systems. No statistically significant differences were observed between the two systems in key post-harvest quality indicators, including weight loss and respiration rate. Notably, GIMCS exhibited superior performance in maintaining fruit firmness, with a hardness of 1.30 kg·cm⁻² compared to 1.26 kg·cm⁻² in conventional storage, indicating a potential advantage in shelf-life extension. Overall, these findings demonstrate that GIMCS represents an affordable, technically robust, and portable cold storage solution capable of delivering preservation performance comparable to—or exceeding—that of conventional cold storage. Its modularity, mobility, and ease of relocation make it particularly well suited to the operational and economic constraints of smallholder farming systems, offering a practical and scalable pathway for improving on-farm cold chain infrastructure. Full article

Minimum wage is an important tool for reducing income inequality and supporting social welfare. Consequently, governments around the world have established minimum wage systems. As such, minimum wage policies connect distributive justice with the economy’s capacity to sustain broad-based welfare over time, placing the equity–efficiency trade-off at the center of societal sustainability. However, the micro-level impact of the minimum wage system on firms has always been an important topic for scholars. This study uses panel data from listed Chinese manufacturing firms over a period from 2005 to 2021 to construct an indicator of the minimum wage standards implemented in the firm locations. Employing the multiple linear regression model, this paper empirically examines the effects of minimum wage on labor productivity. The empirical findings demonstrate that minimum wage significantly reduced the sample firms’ labor productivity. Moreover, the negative impact of the minimum wage was primarily concentrated among non-state-owned firms, labor-intensive firms, firms operating in industries characterized by intense product market competition, firms situated in regions with strong legal protections, firms with comparatively low average employee wages, and export-oriented firms. Subsequently, this study delves into the mechanism through which minimum wage negatively affects labor productivity. We find that implementation of minimum wage leads to a reduction in corporate investment, indicating that there is no significant substitution relationship between capital and labor. These adjustment margins provide microfoundations through which statutory wage floors can influence the resilience and inclusiveness of development, indicating that the pace and design of wage increases should balance income protection with the preservation of productive capacity to support sustainable human development—grounded in steady productivity growth, equitable income distribution, and stable firm investment. Our findings contribute to a better understanding of the mechanism through which minimum wage affects labor productivity in theory, while concurrently furnishing policy insights for the optimization of the minimum wage system and maintaining sustainable societal development in practice. Full article

Grain aging during storage leads to quality deterioration and significant economic losses. Traditional analytical approaches are often labor-intensive, slow, and inadequate for modern intelligent grain storage management. This review summarizes recent advances in the intelligent discrimination of grain aging using volatile organic compound (VOC) fingerprints combined with machine learning (ML) techniques. It first outlines the biochemical mechanisms underlying grain aging and identifies VOCs as early and sensitive biomarkers for timely determination. The review then examines VOC determination methodologies, with a focus on headspace solid-phase microextraction coupled with gas chromatography-mass spectrometry (HS-SPME-GC-MS), for constructing volatile fingerprinting profiles, and discusses related method standardization. A central theme is the application of ML algorithms, including Partial Least Squares Discriminant Analysis (PLS-DA), Support Vector Machines (SVM), Random Forest (RF), and Convolutional Neural Networks (CNN)) for feature extraction and pattern recognition in high-dimensional datasets, enabling effective discrimination of aging stages, spoilage types, and grain varieties. Despite these advances, key challenges remain, such as limited model generalizability, the lack of large-scale multi-source databases, and insufficient validation under real storage conditions. Finally, future directions are proposed that emphasize methodological standardization, algorithmic innovation, and system-level integration to support intelligent, non-destructive, real-time grain quality monitoring. This emerging framework provides a promising powerful pathway for enhancing global food security. Full article

Few-Shot Text Classification (FSTC) aims to classify text accurately into predefined categories using minimal training samples. Recently, prompt-tuning-based methods have achieved promising results by constructing verbalizers that map input data to the label space, thereby maximizing the utilization of pre-trained model features. However, existing verbalizer construction methods often rely on external knowledge bases, which require complex noise filtering and manual refinement, making the process time-consuming and labor-intensive, while approaches based on pre-trained language models (PLMs) frequently overlook inherent prediction biases. Furthermore, conventional data augmentation methods focus on modifying input instances while overlooking the integral role of label semantics in prompt tuning. This disconnection often leads to a trade-off where increased sample diversity comes at the cost of semantic consistency, resulting in marginal improvements. To address these limitations, this paper first proposes a novel Bayesian Mutual Information-based method that optimizes label mapping to retain general PLM features while reducing reliance on irrelevant or unfair attributes to mitigate latent biases. Based on this method, we propose two synergistic generators that synthesize semantically consistent samples by integrating label word information from the verbalizer to effectively enrich data distribution and alleviate sparsity. To guarantee the reliability of the augmented set, we propose a Low-Entropy Selector that serves as a semantic filter, retaining only high-confidence samples to safeguard the model against ambiguous supervision signals. Furthermore, we propose a Difficulty-Aware Adversarial Training framework that fosters generalized feature learning, enabling the model to withstand subtle input perturbations. Extensive experiments demonstrate that our approach outperforms state-of-the-art methods on most few-shot and full-data splits, with F1 score improvements of up to +2.8% on the standard AG’s News benchmark and +1.0% on the challenging DBPedia benchmark. Full article

This study proposes a novel framework for quantifying curved grain boundaries that overcomes key limitations of existing methods. Unlike Fourier-based approaches that require labor-intensive sequential analysis of individual boundaries and selectively represent only high-amplitude regions, or spline-based methods that demand complex parameter selection for interpolation points, the proposed framework integrates curvature variance filtering with U-chord curvature calculation to enable automated, comprehensive, and noise-resistant characterization of grain boundary morphology. The curvature variance filtering adaptively determines smoothing parameters based on local curve properties, while the U-chord curvature method ensures rotational invariance and robustness against digitization errors. Four heat treatment processes were applied to GH4169 alloy, producing distinct grain boundary morphologies with mean curvature (MC) values ranging from 0.0625 to 0.1252. Controlled cooling alone (Process A) yielded predominantly straight boundaries (91.06% straight, 0.12% serrated), while re-dissolution treatment (Process D) produced the highest serration degree (58.81% straight, 3.53% serrated). The quantitative analysis reveals that dispersed δ-phase precipitation creates discrete pinning points, forming serrated boundaries with sharp curvature peaks, whereas dense, parallel δ-phase arrays at specific angles produce coordinated wavy undulations. This framework provides a reliable quantitative tool for optimizing heat treatment protocols to achieve target grain boundary configurations in nickel-based superalloys. Full article