Search Results (567)

Small and medium-sized enterprises (SMEs) in the manufacturing sector often struggle to make effective use of production data due to fragmented systems and limited digital infrastructure. This paper presents a case study of implementing an AI-enhanced digital twin in a custom sportswear manufacturing SME developed under the AI and Robotics Estonia (AIRE) initiative. The solution integrates real-time production data collection using the Digital Manufacturing Support Application (DIMUSA); data processing and control; clustering-based data analysis; and virtual simulation for evaluating improvement scenarios. The framework was applied in a live production environment to analyze workstation-level performance, identify recurring bottlenecks, and provide interpretable visual insights for decision-makers. K-means clustering and DBSCAN were used to group operational states and detect process anomalies, while simulation was employed to model production flow and assess potential interventions. The results demonstrate how even a lightweight AI-driven system can support human-centered decision-making, improve process transparency, and serve as a scalable foundation for Industry 5.0-aligned digital transformation in SMEs. Full article

Understanding the biodiversity of aquatic communities and the underlying mechanisms that shape biodiversity patterns and community dynamics is crucial for the effective conservation and management of freshwater ecosystems. However, traditional survey methods often fail to comprehensively capture species diversity, particularly for low-abundance taxa. Moreover, studies integrating both metazoan and fish communities at fine spatial scales remain limited. To address these gaps, we employed a multi-marker eDNA metabarcoding approach, targeting both the 12S and 18S rRNA gene regions, to comprehensively investigate the composition of metazoan and fish communities in the Yujiang River. A total of 12 metazoan orders were detected, encompassing 15 families, 21 genera, and 19 species. For the fish community, 32 species were identified, belonging to 25 genera, 10 families, and 7 orders. Among these, Adula falcatoides and Coptodon zillii were identified as the most prevalent and abundant metazoan and fish species, respectively. Notably, the most prevalent fish species, C. zillii and Oreochromis niloticus, are both recognized as invasive species. The Bray–Curtis distance of metazoa (average: 0.464) was significantly lower than that of fish communities (average: 0.797), suggesting higher community heterogeneity among fish assemblages. Beta-diversity decomposition indicated that variations in the metazoan and fish communities were predominantly driven by species replacement (turnover) (65.4% and 70.9% for metazoa and fish, respectively) rather than nestedness. Mantel tests further revealed that species turnover in metazoan communities was most strongly influenced by water temperature, while fish community turnover was primarily affected by water transparency, likely reflecting the physiological sensitivity of metazoans to thermal gradients and the dependence of fish on visual cues for foraging and habitat selection. In addition, a co-occurrence network of metazoan and fish species was constructed, highlighting potential predator-prey interactions between native species and Corbicula fluminea, which emerged as a potential keystone species. Overall, this study demonstrates the utility of multi-marker eDNA metabarcoding in characterizing aquatic community structures and provides new insights into the spatial dynamics and species interactions within river ecosystems. Full article

Bats, as key ecological players, interact with a diverse array of organisms and perform essential roles in ecosystems, including pollination, pest control, and seed dispersal. However, their populations face significant threats from habitat contamination, particularly from microplastics (MPs). This study introduces a novel, efficient, and cost-effective method for visualizing transparent microplastics using ultraviolet (UV) light. By employing handheld UV flashlights with a wavelength range of 312 to 400 nm, we enhance the detection of MPs that may otherwise go unnoticed due to color overlap with filtration membranes. All necessary precautions were taken during sampling and analysis to minimize the risk of contamination and ensure the reliability of the results. Our findings demonstrate that the application of UV light significantly improves the visualization and identification of MPs, particularly transparent fibers. This innovative approach contributes to our understanding of plastic contamination in bat habitats and underscores the importance of monitoring environmental pollutants to protect bat populations and maintain ecosystem health. Full article

Beef fat injection technology, used to enhance the perceived quality of lower-grade meat, often results in artificially marbled beef that mimics the visual traits of Wagyu, characterized by dense fat distribution. This practice, driven by the high cost of Wagyu and the affordability of fat-injected beef, has led to the proliferation of mislabeled “Wagyu-grade” products sold at premium prices, posing potential food safety risks such as allergen exposure or consumption of unverified additives, which can adversely affect consumer health. Addressing this, this study introduces a smart sensing system integrated with handheld mobile devices, enabling consumers to capture beef images during purchase for real-time health-focused assessment. The system analyzes surface texture and color, transmitting data to a server for classification to determine if the beef is artificially marbled, thus supporting informed dietary choices and reducing health risks. Images are processed by applying a region of interest (ROI) mask to remove background noise, followed by partitioning into grid blocks. Local binary pattern (LBP) texture features and RGB color features are extracted from these blocks to characterize surface properties of three beef types (Wagyu, regular, and fat-injected). A support vector machine (SVM) model classifies the blocks, with the final image classification determined via majority voting. Experimental results reveal that the system achieves a recall rate of 95.00% for fat-injected beef, a misjudgment rate of 1.67% for non-fat-injected beef, a correct classification rate (CR) of 93.89%, and an F1-score of 95.80%, demonstrating its potential as a human-centered healthcare tool for ensuring food safety and transparency. Full article

Urban green spaces are increasingly recognized as important habitats for pollinators, and wildflower seed mixes marketed as pollinator-friendly are gaining popularity, though their actual conservation value remains poorly understood. This study provides the first systematic screening of commercially available seed mixes in Portugal, evaluating their taxonomic composition, origin, life cycle traits, and potential to support pollinator communities. A total of 229 seed mixes were identified. Although these have a predominance of native species (median 86%), the taxonomic diversity was limited, with 91% of mixes comprising species from only one or two families, predominantly Poaceae and Fabaceae, potentially restricting the range of floral resources available to pollinators. Only 21 seed mixes met the criteria for being pollinator-friendly, based on a three-step decision tree prioritizing native species, extended flowering periods, and visual diversity. These showed the highest percentage of native species (median 87%) and a greater representation of flowering plants. However, 76% of all mixes still included at least one non-native species, although none is considered invasive. Perennial species dominated all seed mix types, indicating the potential for the long-term persistence of wildflower meadows in urban spaces. Despite their promise, the ecological quality and transparency of the seed mix composition remain inconsistent, with limited certification or information on species origin. This highlights the need for clearer labeling, regulatory guidance, and ecologically informed formulations. Seed mixes, if properly designed and implemented, represent a largely untapped yet cost-effective tool for enhancing the pollinator habitats and biodiversity within urban landscapes. Full article

This research investigates the metric accuracy of 3D transparent object reconstruction, a task where conventional photogrammetry often fails. The topic is especially relevant in cultural heritage (CH), where accurate digital documentation of glass and transparent artifacts is important. The work proposes a practical methodology using existing tools to verify metric accuracy standards. The study compares three methods, conventional photogrammetry, 3D Gaussian splatting (3DGS), and 2D Gaussian splatting (2DGS), to assess their ability to produce complete and metrically reliable 3D models suitable for measurement and geometric analysis. A transparent glass artifact serves as the case study. Results show that 2DGS captures fine surface and internal details with better geometric consistency than 3DGS and photogrammetry. Although 3DGS offers high visual quality, it introduces surface artifacts that affect metric reliability. Photogrammetry fails to reconstruct the object entirely. The study highlights that visual quality does not ensure geometric accuracy, which is critical for measurement applications. In this work, ground truth comparisons confirm that 2DGS offers the best trade-off between accuracy and appearance, despite higher computational demands. These findings suggest extending the experimentation to other sets of images featuring transparent objects, and possibly also reflective ones. Full article

Grouting, as a widely applicable and versatile foundation treatment technology, plays a crucial role in addressing seepage control problems in cover layers due to its flexibility and convenience. The effectiveness of grouting largely depends on slurry diffusion; however, due to the opaque nature of geotechnical media, the diffusion mechanism of slurry in the cover layers remains insufficiently understood. To investigate this, a visual grouting model device was designed and fabricated, and grouting tests were conducted using transparent soil materials to simulate the cover layers. The slurry diffusion patterns and the velocity field within the transparent soil were analyzed. The results show that, based on refractive-index matching, fused quartz sand of specific gradation and white mineral oil were selected as simulation materials for the cover layers. A stable slurry suitable for transparent grouting was also chosen to satisfy visualization requirements. The transparent soil grouting model, integrated with a Digital Image Correlation (DIC) monitoring system, has the advantages of demonstrating simple operation, real-time monitoring, and high precision. These tests verify the feasibility of visualizing slurry diffusion in cover layers. Furthermore, step-pressure grouting tests preliminarily reveal the dynamic mechanism of slurry diffusion. The results suggest that, in the cover layer, the cover layer in this grouting test is mainly splitting grouting, accompanied by compaction grouting. These methods offer new insights and methods for model testing of cover layer grouting mechanisms. Full article

The Industrial Internet of Things (IIoT) has revolutionized industry through interconnected devices and intelligent applications. Leveraging the advancements in sixth-generation cellular networks (6G), the 6G-aided IIoT has demonstrated a superior performance across applications requiring low latency and high reliability, with image recognition being among the most pivotal. However, the existing algorithms often neglect the explainability of image recognition processes and fail to address the collaborative potential between edge computing servers. This paper proposes a novel method, IRCE (Intelligent Recognition with Collaborative Edges), designed to enhance the explainability and transferability in 6G-aided IIoT image recognition. By incorporating an explainable layer into the feature extraction network, IRCE provides visual prototypes that elucidate decision-making processes, fostering greater transparency and trust in the system. Furthermore, the integration of the local maximum mean discrepancy (LMMD) loss facilitates seamless transfer learning across geographically distributed edge servers, enabling effective domain adaptation and collaborative intelligence. IRCE leverages edge intelligence to optimize real-time performance while reducing computational costs and enhancing scalability. Extensive simulations demonstrate the superior accuracy, explainability, and adaptability of IRCE compared to those of the traditional methods. Moreover, its ability to operate efficiently in diverse environments highlights its potential for critical industrial applications such as smart manufacturing, remote diagnostics, and intelligent transportation systems. The proposed approach represents a significant step forward in achieving scalable, explainable, and transferable AI solutions for IIoT ecosystems. Full article

Effective management of urban development requires robust decision-support tools, including land suitability analysis and its visual communication. This study introduces and evaluates seven 3D geovisualization methods—Horizontal Planes, Point Cloud, 3D Surface, Vertical Planes, 3D Graduated Symbols, Prism Map, and Voxels—for visualizing land suitability for residential development in Jihlava, Czechia. Using five raster-based data layers derived from a multi-criteria evaluation (Urban Planner methodology) across three time horizons (2023, 2028, 2033), the visualizations were implemented in ArcGIS Online and assessed by 19 domain experts via a structured questionnaire. The evaluation focused on clarity, usability, and accuracy in interpreting land suitability values, with the methods being rated on a five-point scale. Results show that the Horizontal Planes method was rated highest in terms of interpretability and user satisfaction, while 3D Surface and Vertical Planes were considered the least effective. The study demonstrates that visualization methods employing visual variables (e.g., color and transparency) are better suited for land suitability communication. The methodological contribution lies in systematically comparing 3D visualization techniques for thematic spatial data, providing guidance for their application in planning practice. The results are primarily intended for urban planners, designers, and local government representatives as supportive tools for efficient planning of future built-up area development. Full article

Background/Objectives: This paper introduces a pedagogically grounded framework for transforming canonical literary texts in primary education through generative AI. Guided by multiliteracies theory, Vygotskian pedagogy, and epistemic justice, the system aims to enhance interpretive literacy, developmental alignment, and cultural responsiveness among learners aged 7–12. Methods: The proposed system enables educators to perform age-specific text simplification, visual re-narration, lexical reinvention, and multilingual augmentation through a suite of modular tools. Central to the design is the Ethical–Pedagogical Validation Layer (EPVL), a GPT-powered auditing module that evaluates AI-generated content across four normative dimensions: developmental appropriateness, cultural sensitivity, semantic fidelity, and ethical transparency. Results: The framework was fully implemented and piloted with primary educators (N = 8). The pilot demonstrated high usability, curricular alignment, and perceived value for classroom application. Unlike commercial Large Language Models (LLMs), the system requires no prompt engineering and supports editable, policy-aligned controls for normative localization. Conclusions: By embedding ethical evaluation within the generative loop, the framework fosters calibrated trust in human–AI collaboration and mitigates cultural stereotyping and ideological distortion. It advances a scalable, inclusive model for educator-centered AI integration, offering a new pathway for explainable and developmentally appropriate AI use in literary education. Full article

Biometric authentication has become one of the most widely used tools in the current technological era to authenticate users and to distinguish between genuine users and impostors. The face is the most common form of biometric modality that has proven effective. Deep learning-based face recognition systems are now commonly used across different domains. However, these systems usually operate like black-box models that do not provide necessary explanations or justifications for their decisions. This is a major disadvantage because users cannot trust such artificial intelligence-based biometric systems and may not feel comfortable using them when clear explanations or justifications are not provided. This paper addresses this problem by applying an efficient method for explainable face recognition systems. We use a Class Activation Mapping (CAM)-based discriminative localization (very narrow/specific localization) technique called Scaled Directed Divergence (SDD) to visually explain the results of deep learning-based face recognition systems. We perform fine localization of the face features relevant to the deep learning model for its prediction/decision. Our experiments show that the SDD Class Activation Map (CAM) highlights the relevant face features very specifically and accurately compared to the traditional CAM. The provided visual explanations with narrow localization of relevant features can ensure much-needed transparency and trust for deep learning-based face recognition systems. We also demonstrate the adaptability of the SDD method by applying it to two different techniques: CAM and Score-CAM. Full article

Generative artificial intelligence (AI) is rapidly transforming healthcare systems since the advent of OpenAI in 2022. It encompasses a class of machine learning techniques designed to create new content and is classified into large language models (LLMs) for text generation and image-generating models for creating or enhancing visual data. These generative AI models have shown widespread applications in clinical practice and research. Such applications range from medical documentation and diagnostics to patient communication and drug discovery. These models are capable of generating text messages, answering clinical questions, interpreting CT scan and MRI images, assisting in rare diagnoses, discovering new molecules, and providing medical education and training. Early studies have indicated that generative AI models can improve efficiency, reduce administrative burdens, and enhance patient engagement, although most findings are preliminary and require rigorous validation. However, the technology also raises serious concerns around accuracy, bias, privacy, ethical use, and clinical safety. Regulatory bodies, including the FDA and EMA, are beginning to define governance frameworks, while academic institutions and healthcare organizations emphasize the need for transparency, supervision, and evidence-based implementation. Generative AI is not a replacement for medical professionals but a potential partner—augmenting decision-making, streamlining communication, and supporting personalized care. Its responsible integration into healthcare could mark a paradigm shift toward more proactive, precise, and patient-centered systems. Full article

Organizations face significant challenges in measuring and enhancing sustainability performance across complex operational processes. Current assessment methods frequently lack granularity, real-time capability, and integration with operational data. This study addresses these gaps by developing a conceptual framework that integrates business process mining with Global Reporting Initiative (GRI) metrics. The methodology incorporates environmental, social, and economic sustainability indicators into process mining techniques through systematic metric mapping and event log enrichment. The framework enables the extraction and analysis of sustainability performance data at the process level, creating detailed heat maps that visualize resource utilization, emissions, and waste generation. An application to a Purchase-to-Pay process case study demonstrates how process variants impact sustainability metrics differently. Delays increase emissions by 16.7%, while rework increases waste generation by 41.7%. The results identify specific process bottlenecks with high environmental impact and reveal critical misalignments between economic and environmental sustainability goals. This framework provides organizations with a standardized yet flexible approach to measuring sustainability performance, bridging the gap between high-level sustainability reporting and operational processes. It enables continuous monitoring, targeted interventions, and transparent reporting across diverse industry contexts. Full article

Background: Previous proteomic studies provided valuable information about cataracts, but unclarified issues, such as sex and ethnicity-associated differences, remain. This study aimed to provide additional data on cataract-related proteins regarding age, sex, and cataract type. Methods: Twenty-six female and seven male Turkish cataract patients were screened for visual acuity and dysfunctional lens index. A nano-LC-MS/MS system and Progenesis QI software v3.0 were used for protein identification and quantification. The remaining data were evaluated with SPSS Version 29.0 software. Results: Proteins that showed age-associated changes were mainly involved in cytoskeletal organization. A glyoxalase enzyme, caveolin 1, and HS90B were lower, and RAB8B and ATP6V1B1 were higher in lenses in women. Proteins with lower levels in cataractous lenses than in transparent lenses included filensin and phakinin, concurrent with previous publications, and LCTL, GDI, HSPB1, and EIF4A2, not reported before. Corticonuclear cataracts constituted the only group showing depletions in putatively protective proteins, while the cortical type was the least influenced. ANXA1 and DNHD1 positively, and TCPD, SEC14L2, and PRPS1 proteins negatively correlated with visual acuity. Conclusions: This study revealed cataract-related proteins concurrent with earlier studies and new ones hitherto unreported. Despite the low number of patients investigated, the results merit further research, as these new proteins are highly likely to be involved in cataractogenesis. Full article

Timely and accurate detection of forest fires is crucial for protecting forest ecosystems. However, traditional monitoring methods face significant challenges in effectively detecting forest fires, primarily due to the dynamic spread of flames and smoke, irregular morphologies, and the semi-transparent nature of smoke, which make it extremely difficult to extract key visual features. Additionally, deploying these detection systems to edge devices with limited computational resources remains challenging. To address these issues, this paper proposes a lightweight hybrid receptive field model (LHRF-YOLO), which leverages deep learning to overcome the shortcomings of traditional monitoring methods for fire detection on edge devices. Firstly, a hybrid receptive field extraction module is designed by integrating the 2D selective scan mechanism with a residual multi-branch structure. This significantly enhances the model’s contextual understanding of the entire image scene while maintaining low computational complexity. Second, a dynamic enhanced downsampling module is proposed, which employs feature reorganization and channel-wise dynamic weighting strategies to minimize the loss of critical details, such as fine smoke textures, while reducing image resolution. Furthermore, a scale weighted Fusion module is introduced to optimize multi-scale feature fusion through adaptive weight allocation, addressing the issues of information dilution and imbalance caused by traditional fusion methods. Finally, the Mish activation function replaces the SiLU activation function to improve the model’s ability to capture flame edges and faint smoke textures. Experimental results on the self-constructed Fire-SmokeDataset demonstrate that LHRF-YOLO achieves significant model compression while further improving accuracy compared to the baseline model YOLOv11. The parameter count is reduced to only 2.25M (a 12.8% reduction), computational complexity to 5.4 GFLOPs (a 14.3% decrease), and mAP50 is increased to 87.6%, surpassing the baseline model. Additionally, LHRF-YOLO exhibits leading generalization performance on the cross-scenario M4SFWD dataset. The proposed method balances performance and resource efficiency, providing a feasible solution for real-time and efficient fire detection on resource-constrained edge devices with significant research value. Full article