This paper introduces an interactive calibration system for color vision assessment utilizing augmented reality (AR) glasses, and investigates its potential relevance as a perceptual calibration stage for human-robot systems. The system re-imagines the traditional Farnsworth-Munsell 100 Hue (FM 100 Hue) test-a benchmark in the textile industry for evaluating the color discrimination of quality control personnel-as an interactive AR-based task, enabling portable color discrimination assessment under ordinary ambient lighting conditions. We quantitatively evaluate the correlation between this AR-implemented assessment and the standard physical test through a comparative user study. Results demonstrate a high degree of diagnostic consistency, indicating that the AR platform can reproduce key characteristics of the conventional assessment while reducing part of the variability associated with display and ambient conditions. Beyond clinical assessment, this work positions AR glasses as a bidirectional perceptual interface, enabling individualized modeling of human color perception for robotic systems. The proposed method facilitates robust, color-aware perception, decision-making, and task execution in human-centered robotics, with potential applications in automated textile quality assurance and assistive robotics. This calibration framework is a promising step toward achieving seamless and adaptive visual perception alignment between humans and humanoid or other robotic platforms.
Natural fiber fabrics are increasingly used in sustainable fashion apparel because of their renewability and biodegradability, but hemp- and linen-containing fabrics often suffer from excessive stiffness, poor drape, and dimensional instability after laundering. This study developed an integrated framework for process optimization and apparel-oriented evaluation of natural-fiber fabrics. Three woven fabrics, namely organic cotton, hemp/organic cotton, and linen/organic cotton, were selected. An enzyme-assisted finishing process for hemp/organic cotton fabric was optimized using a Box-Behnken design with cellulase concentration, treatment temperature, and treatment time as variables. Flexural rigidity, tensile strength retention, air permeability, and residual shrinkage were used as response indicators. The optimum condition was 1.18% owf cellulase, 54 °C, and 42 min. Under this condition, flexural rigidity decreased by 28.6%, tensile strength retention remained 93.6%, air permeability improved, and residual shrinkage was reduced to 2.5%. Comparative evaluation showed that the optimized hemp/cotton fabric achieved the highest overall apparel suitability index among the tested materials. However, the improvement in softness and dimensional behavior was accompanied by a moderate loss in tensile strength retention, indicating a practical compromise rather than simultaneous optimization of all properties.
The mechanical properties of mine roof strata affect the safety of fully mechanized mining face. The roof is easy to collapse, under the action of cutting vibration and gravity. In ultra-thin coal seams with a thickness of less than 0.8 meters and relatively high coal hardness, the cutting of coal mining machines will cause serious interference to the weak roof, thereby reducing coal quality and endangering production safety. To address this issue, this study established a cutting dynamics model of the shearer drum for ultra-thin coal seams under soft roof conditions using the discrete element method in EDEM. Based on the single factor method and using the top plate damage rate as the evaluation index, the influence of drum structure and operating parameters on soft top plate damage was systematically studied. The results show that the drum diameter is 650 mm, the hub diameter is 325 mm, the traction speed is 2 m/min, and the cutting depth is 650 mm under the single factor condition, and the shearer drum in ultra-thin coal seam has achieved the best cutting effect respectively.
Traditional deep learning-based image classification methods rely on large-scale labeled data. In practical scenarios such as medical imaging and agricultural monitoring, obtaining labeled data requires substantial manpower and time, becoming a bottleneck for deployment. Similarly, in the textile industry, the automated inspection of fabric defects and high-precision fiber texture identification often face similar challenges due to the scarcity of highquality annotated data. Self-supervised learning offers a solution. This study evaluates the effectiveness of contrastive learning-based selfsupervised feature representation in image classification, focusing on feature extraction from unlabeled data. An efficient contrastive learning framework was constructed and evaluated on CIFAR-10, ImageNet-1K, and CUB-200-2011. Based on the ResNet architecture, combined with the InfoNCE loss and data augmentation, a two-stage training strategy of self-supervised pre-training followed by supervised fine-tuning was adopted. Model performance was assessed using classification accuracy and F1 score. Results show that the proposed method outperforms traditional supervised learning, especially in low-label regimes and as a robust initialization strategy across datasets. The model demonstrates strong generalization in low-sample settings and adaptability to different data distributions. This study clarifies the role of contrastive learning in feature representation for image classification and provides support for applying self-supervised learning in domains with limited annotations, such as medical image analysis and agricultural monitoring. It also offers a transferable framework for related computer vision tasks.
To address national total medal count prediction and performance evaluation, within the context of increasing techno logical competition in sports equipment, a Bivariate-Hurdle-Tobit composite model was constructed. This model is suitable for predicting zero-inflated, count-type, and highly correlated dependent variables. The model recognizes that a nation's athletic success is increasingly underpinned by its industrial and technological prowess, including advancements in textile engineering for. Through factor analysis and correlation tests, six significant regression factors were identified: total athlete count, host status, participation rate in dominant events. The model underwent comprehensive evaluation via rolling cross-validation and five metrics, demonstrating low prediction error and high accuracy with AUC values reaching 0.98 and 0.96 respectively, and Pseudo R2 consistently exceeding 75%. Predictions indicate the United States, China, and the United Kingdom will occupy the top three positions on the medal tally. The study reveals a significant positive impact of the "host nation effect" on national performance. For instance, the U.S. as the next host nation is projected to increase its gold medal share by 0.75%. Additionally, countries like Andorra, Benin, and Belize are predicted to have a 95% probability of winning their first-ever medals. Analyzing the influence of great coaches, the study quantifies the "great coach effect" using Chow tests and Difference-in-Differences models. Findings indicate this effect has limited impact in non-host nations but exhibits significant synergistic effects in host countries. For additional insights into the Olympic medal standings, the model incorporates gender ratio analysis. Results show male athlete participation rates have a significant negative impact on medal distribution, suggesting an increase in female athlete representation is advisable.
Addressing the challenges of opaque audience voting data and variable elimination rules in competitive reality shows, this study constructs a quantitative framework integrating latent variable reconstruction, rule auditing, and dynamic optimization. First, a Bayesian state space model is employed to perform latent state reconstruction on audience support rates across thirty-four seasons. Using Laplace approximation techniques to quantify estimation uncertainty, the analysis reveals significantly heightened sensitivity of voting fluctuations to late-stage decision-making. Subsequently, counterfactual simulations under different scoring rules were conducted using the reconstructed data, deeply analyzing the systemic differences between percentage-based and ranking-based systems in addressing expert bias versus public popularity bias. Furthermore, a panel regression model was applied to deconstruct the influence mechanisms of factors like professional dancer background, contestant occupation, and age on advancement probabilities, identifying significant divergences in the driving logic between expert judging and public voting. Finally, addressing the vulnerability of existing mechanisms to noisy data, we propose an uncertainty-aware weighted moving average mechanism. By dynamically adjusting voting weights, this approach simultaneously enhances the robustness and fairness of elimination decisions. Experiments demonstrate that this framework effectively restores latent competitive dynamics, providing a robust algorithmic foundation for optimizing dual-track judging systems.
"The Romance of the Western Chamber" exerted a profound influence on the romantic legend creations of Ming dynasty playwright Wang Tingne, manifested in his creative approach through "quotation and imitation" -namely, replicating plotlines, adopting literary expressions, and adapting textual elements. Furthermore, through active transformation of personal literary philosophy, Wang Tingne's romantic legends enriched the original "Cui-Zhang" love paradigm. His revisions of the plot and illustrations demonstrated new interpretations: he critiqued Zhang Sheng's accurate interpretation of the poem and criticized his method of jumping over the wall, while crafting character portrayals more aligned with the social context of the Ming dynasty, thereby infusing fresh meaning into the play's dissemination during that era. Wang Tingne's practice of "quotation and imitation" also influenced other forms of popular literature, further facilitating the spread of "The Romance of the Western Chamber".
Martial arts training requires garments capable of accommodating large-amplitude, multi-planar movements, yet quantitative evaluation of garment structural effects remains limited in textile engineering research. This study proposes a biomechanics-informed design and evaluation approach for functional martial arts training clothing and reports a pilot experimental assessment. A functional garment incorporating targeted panel segmentation, directional elasticity, and localized structural features was developed based on movement demand analysis. Sixteen experienced practitioners performed representative training tasks under functional and conventional clothing conditions. Lower-limb joint range of motion (ROM) was quantified using three-dimensional inertial motion capture. The results revealed movement-and plane-specific differences in joint ROM between clothing conditions, indicating that garment structural design can influence clothing-movement interaction and joint kinematic responses in high-mobility training applications.
Apigenin, as a natural flavonoid compound, is widely present in various fruits, vegetables, and herbal medicines. In recent years, numerous studies have shown that apigenin exhibits significant potential in multiple liver diseases through various pathways such as anti-inflammation, antioxidation, anti-fibrosis, regulation of lipid metabolism, induction of autophagy and apoptosis. This article elaborates on the pharmacological mechanisms and clinical application progress of apigenin in major liver diseases including non-alcoholic fatty liver disease (NAFLD), alcoholic liver disease (ALD), liver fibrosis, liver cancer, and drug-induced liver injury, and also looks forward to future research directions, aiming to provide a basis for the in-depth exploration of apigenin as a therapeutic drug for liver diseases.
As a core of power electronic systems, three-phase voltage source inverters' power devices are prone to open-circuit faults, causing system waveform distortion and load abnormalities. Traditional diagnostic methods suffer from low accuracy and poor real-time performance. To address this, this paper proposes a real-time diagnosis and location method based on a wavelet-Transformer fusion network. It uses wavelet transform to extract multi-scale time-frequency features of fault signals and combines Transformer's self-attention mechanism to mine temporal correlation. Data preprocessing, feature selection, and comprehensive location strategies optimize model performance. Experimental results show that the proposed wavelet-Transformer fusion network achieves 95.6% overall diagnostic accuracy, 96.2% fault location accuracy, and an average end-to-end response latency of 8.4 ms under the main test setting. Compared with single-module base-lines, the proposed framework provides a better balance between diagnostic accuracy and online deployment efficiency, offering a practical solution for intelligent operation and maintenance of power electronic equipment.
Health state assessment and remaining useful life prediction are two key tasks for predictive maintenance of rotating machinery. To address limitations of existing studies-such as the separation of health assessment and lifetime prediction, insufficient modeling of temporal degradation features, and limited interpretability-this study proposes a deep temporal learning framework that jointly implements both tasks. The method takes raw monitoring signals as input, constructs temporal samples using a sliding window, and employs a shared temporal feature encoder to extract degradation representations. Based on these shared features, a health state assessment branch generates a continuous health indicator, while an RUL prediction branch estimates the remaining useful life, enabling collaborative modeling of the degradation process within a unified feature space. A multi-task joint loss is introduced to jointly optimize health state modeling and lifetime regression, enhancing the representation of local degradation patterns, long-term trends, and stage-wise characteristics. Experimental results on public run-to-failure datasets show that the proposed method produces health indicators with strong monotonicity, trendability, and robustness, and achieves superior prediction accuracy compared to several baselines. These findings verify the effectiveness of the proposed framework and highlight its potential for intelligent and predictive maintenance applications.
The global textile and apparel industry faces critical challenges in supply chain transparency and substantiating sustainability claims. Conventional textile labeling, while mandated to disclose fiber composition and origin, offers only static data that is inadequate for tracking complex manufacturing processes or verifying environmental performance. The European Union’s forthcoming Digital Product Passport (DPP) is set to create a dynamic, data-rich digital record for textile products. This paper addresses the crucial challenge of ensuring data integrity and validity for textile product attributes such as fiber traceability, chemical usage, and circularity data. We propose and evaluate a sustainable product tracking system that integrates labeling regulations with the DPP, underpinned by a permissioned blockchain architecture. To validate the framework’s feasibility and efficacy, we developed a prototype system based on Hyperledger Fabric and conducted performance evaluations and a case study. The system’s architecture is designed to ensure immutable tracking throughout the textile supply chain. Through an end-to-end tracking experiment of an organic cotton garment from raw fiber to finished apparel, the prototype demonstrated a transaction latency of approximately 3.2 seconds and a stable throughput of over 45 TPS under realistic network constraints, proving its technical viability. The analysis confirms that leveraging blockchain as the data backbone for the DPP provides a verifiable and immutable record, effectively combating greenwashing and fostering a circular economy for textiles and apparel.
Operational performance evaluation is essential for improving competitiveness and sustainability in textile enterprises, which are characterized by complex production processes, high resource consumption, and stringent quality require ments. Traditional evaluation approaches relying mainly on financial indicators cannot comprehensively reflect multidi mensional operational characteristics such as production efficiency, quality stability, cost control, and environmental performance. To address this limitation, this study proposes a textile enterprise operational performance evaluation model based on multi criteria decision making (MCDM) methods. First, a hierarchical evaluation index system is constructed from four dimensions: production efficiency, quality management, cost control, and sustainability capability. Second, the analytic hierarchy process (AHP) is applied to determine subjective indicator weights, while the entropy method is used to derive objective weights from enterprise operational data. A combined weighting approach is then developed to integrate expert knowledge and data information. Finally, the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) is employed to evaluate and rank operational performance. A case study involving five medium sized textile manufacturing enterprises from the same industrial region was selected to demonstrate the feasibility and reliability of the proposed model. While this selection ensures comparability, the lack of variation in technological conditions could lead to a small divergence degree in the entropy method, which might undermine the objective weight calculation. Future research could expand the case study to include enterprises from different regions or with varying technological conditions to enhance the representativeness and robustness of the findings Results indicate that the model effectively differentiates performance levels and identifies key influencing factors, particularly equipment utilization, defect rate, and energy consumption intensity. The proposed approach provides a scientific and practical decision support tool for textile enterprise performance assessment and operational improvement.
Silicon-based ultraviolet sensors enable direct and non-contact monitoring of true discharge by capturing ultraviolet emission patterns. This monitoring is of significant value for the electrical safety of industrial facilities, such as textile production lines, where discharge events can lead to equipment failure or safety hazards. However, discharge targets in silicon-based UV images are often weak and small, and they are easily confused with background noise. To address these challenges, this paper proposes PDDA, a discharge detection method built on YOLOv8. PDDA introduces an EMA attention module into the bottleneck to enhance pixel-level feature weighting, replaces the original neck with a Cross-level Context Fusion Module to strengthen multi-scale fusion, and adopts the Swish activation function to improve training stability. Experimental results demonstrate that PDDA achieves 86.9% mAP@0.5 with a throughput of 12 fps, outperforming several mainstream detectors while maintaining practical efficiency for direct, continuous monitoring.
To address the limited capability of conventional distributed secondary control methods in exploiting multi-node coupling and dynamic information during frequency and voltage restoration in islanded microgrids, this paper proposes a spatiotemporal convolutional network-based distributed secondary restoration control method. First, a microgrid model incorporating droop control, communication topology, and restoration objectives is established, and the restoration problem is formulated as a distributed control task based on local measurements and neighborhood information. Then, spatiotemporal input features integrating frequency and voltage deviations and power information are constructed. A spatiotemporal convolutional network is employed to extract spatial correlation among distributed generators and temporal characteristics during disturbances, thereby generating secondary compensation signals. Furthermore, a hybrid control framework combining a distributed secondary controller with a spatiotemporal compensator is developed to enhance regulation capability while preserving the control structure. Simulation results show that the proposed approach achieves smaller deviations, shorter recovery times, and stronger disturbance rejection under load variations, parameter perturbations, and communication delays. The results indicate that introducing spatiotemporal feature modeling into distributed secondary control reduces transient deviations, shortens recovery time, and improves disturbance rejection capability in the tested microgrid scenarios.
The evolution of solid-state electrolyte interfaces decisively affects all-solid-state lithium batteries' cycling stability and reliability. However, most studies rely on static analysis or limited aggregated descriptors, failing to capture dynamic interfacial degradation during continuous evolution. To address this, we propose an interfacial stability detection framework integrating LAMMPS-based molecular dynamics (MD) simulations with temporal deep learning, aiming to identify state transitions and provide early instability warnings from atomic trajectories. First, representative solid-state electrolyte interface models are built, and continuous evolution trajectories under varying conditions are generated via MD simulations. Multi-scale temporal features are then extracted to characterize interfacial migration, local structural reconstruction, and anomalous dynamics. Using these sequential representations, a temporal deep learning model detects stability states and anticipates degradation trends. Results show that our method effectively distinguishes stable from unstable interfaces by their evolutionary pathways, outperforming conventional static-descriptor-based methods in detection accuracy, temporal sensitivity, and early-warning capability. Interpretability analysis further reveals stagedependent roles of different features during degradation, offering new insights into key drivers of interface instability. This study provides a temporally aware computational framework for dynamic evaluation, risk diagnosis, and data-driven optimization of solid-state electrolyte interfaces.
In the context of computational textile design and intelligent manufacturing, the key challenge lies in collaboratively achieving both the functionality and artistic expression of fabrics. Existing methods often focus on a single dimension and lack a unified framework that integrates multi-physics constraints with artistic semantics. To address this, this paper proposes a generative artificial intelligence-based method, with the core innovations being: 1) a dual-path decoupled representation of yarn arrangement and weave patterns; 2) the design of a function-semantic dual encoding and joint generation framework; 3) the introduction of curriculum learning strategies to enhance training stability. Experiments are based on a self-built dataset, TextileArt-Weave v1.0 (8, 642 samples), covering three types of functional fabrics and five artistic styles. Quantitative results show that the proposed method significantly outperforms visual generation and engineering CAD baselines in functional indicators (porosity error: 3.2 ± 0.4% vs. 9.8 ± 0.6% / 12.4 ± 0.8%), and outperforms rule-driven methods in artistic semantic alignment (CLIP Score: 0.78 ± 0.02 vs. 0.32 ± 0.02), while maintaining excellent manufacturability (interweaving cycle compliance: 96.8 ± 0.7%, whereas pure visual models yield 0%). Ablation studies and robustness validation show that each module makes a clear contribution to functionality, artistry, and manufacturability, and the method remains stable under noise, material changes, and process disturbances. To the best of our knowledge, this study presents one of the first end-to-end frameworks for mapping multimodal prompts to manufacturable fabric structures under joint functional and artistic constraints and has open-sourced the textile multimodal benchmark dataset, providing feasible paths for smart wearables, cultural digitization, and sustainable fashion.
This study aims to construct a high-precision, real-time UAV flight attitude stability prediction and evaluation system to solve the core problems of insufficient accuracy and poor real-time performance of traditional prediction methods. This is particularly significant for UAVs integrated with flexible textile sensors or those performing structural health monitoring of high-strength fiber-reinforced composite airframes. First, key attitude parameters such as pitch angle, roll angle, and yaw angle, as well as environmental data, are collected through multi-sensor fusion technology. After cleaning, feature extraction, and standardization preprocessing, a multi-layer LSTM prediction model combined with an attention mechanism is constructed. A multi-step rolling prediction strategy and a grid search-cross-validation hyperparameter optimization method are adopted. At the same time, a multi-dimensional stability evaluation framework covering indicators such as angular velocity variance, attitude deviation, and prediction error is established. Experimental results show that the proposed LSTM model significantly outperforms traditional ARIMA and other models in attitude prediction, reducing RMSE and MAE by 8.82% and 10.22%, respectively. It exhibits optimal prediction accuracy, particularly within the 25°-40°dive angle range, and effectively captures the temporal dependencies of attitude data. This study provides a scientific basis for optimizing UAV attitude control. Meanwhile, it holds significant theoretical and engineering value for improving fault early warning mechanisms.
Manual inspection of landscape pathology and textile fiber defects suffers from inherent subjective bias and suboptimal throughput. To bypass these bottlenecks, we propose the Ghost-Convolution Enlightened Vision Transformer (GeT). We constructed a novel hybrid neural network architecture, the Ghost-Convolution Enlightened Vision Transformer (GeT), which synergistically integrates the lightweight local feature extraction proficiency of Convolutional Neural Networks (CNN) with the global semantic modeling capabilities of Vision Transformers (ViT). Utilizing a newly established standard dataset (GLDP15k) comprising 15, 000 heterogeneous field images, the model was subjected to rigorous hyperparameter optimization and ablation studies. Optimization on the GLDP15k dataset yielded a peak accuracy of 96.8% across 12 target classes, maintaining a Kappa-coefficient of 0.941. Constrained to 1.16 M parameters, the architecture executes at 5.5 ms per image (180 FPS) on edge hardware. A 6-month application at Yuexiu Park demonstrated a 3.2-fold improvement in detection efficiency and a 35% reduction in pesticide usage compared to manual inspections. This study not only elucidates the interpretability of hybrid attention mechanisms in phytopathology but also adapts these vision-based paradigms to the detection of microscopic anomalies in textile weaving patterns, providing a scalable and computationally efficient solution for both precision plant protection and industrial fabric defect inspection.
This study constructs a series of models for water flow propagation in mine tunnel water inrushes and dynamic escape path planning for miners, aiming to provide reference escape strategies for mine water disasters. For single-point water outburst scenarios, the study abstracts the roadway network as a graph structure. Employing a breadth-first search algorithm, it calculates the time of first water arrival at endpoints and the time of roadway flooding, incorporating spatial parameters and initial outburst flow rates. For optimal miner escape route planning, the study introduces dynamic velocity constraints based on water depth: 4 m/s in dry conditions, 2 m/s downstream and 1 m/s upstream in low water, with high water levels prohibiting movement. The defined velocity constraints under different water depths provide essential ergonomic performance metrics for the development of miners'protective clothing in hydraulic environments. The model employs a time-extended Dijkstra algorithm to search for the shortest total escape time path. Building upon this foundation, the model is extended to predict flow superposition in delayed dual-source water inrushes. By applying time axis shifts, selecting minimum arrival times, and superimposing flow rates, it calculates flow states under dual-source interactions, enabling dynamic simulation of multi-source inrush systems. Finally, integrating dualsource flow results with the time-varying Dijkstra algorithm, the model plans optimal miner escape routes during delayed dual-source water inrushes. This study offers a theoretical framework for integrating wearable safety systems into intelligent mine disaster mitigation.
