Search Results (2,310)

Quantifying microbial growth with high temporal resolution remains essential yet challenging due to limitations of optical, manual, and biochemical methods. Here, we introduce an AI-enhanced electrochemical impedance spectroscopy platform for real-time, label-free monitoring of Saccharomyces cerevisiae growth. Broadband impedance measurements (1 Hz–100 kHz) were collected from yeast cultures across log-phase development. Engineered features—derived from impedance magnitude and phase—captured dielectric and conductive shifts associated with cell proliferation, membrane polarization, and ionic redistribution. A Gaussian Process Regression model trained on these features predicted optical density (OD600) with high precision (RMSE = 0.79 min; R² = 0.9996; r = 0.9998), and achieved 100% classification accuracy when discretized into 15-min growth intervals. The system operated with sub-millisecond latency and minimal memory footprint, enabling embedded deployment. Benchmarking against conventional methods revealed superior throughput, automation potential, and independence from labeling or turbidity-based optics. This AI-driven platform forms the core of a real-time digital twin for yeast culture monitoring, capable of predictive tracking and adaptive control. By fusing electrochemical biosensing with machine learning, our method offers a scalable and robust solution for intelligent fermentation and bioprocess optimization. Full article

Accurate landslide mapping near critical infrastructure requires not only data on landslide characteristics but also clear definitions of the spatial extent of surveyed areas. While national projects like Italian Landslide Inventory (IFFI) and Italian Guidelines for the classification and management of risk, safety assessment and monitoring of existing bridges (LLG 2022) provide a list of data to collect during a field visit survey, they lack clear specifications for buffer zones, limiting data comparability and risk assessment reliability. This study refines a hierarchical framework developed by the FABRE Geo Working Group, in alignment with LLG 2022, introducing five key zones—Landslide Inventory Reference Area, Diagnostic Area, Geomorphological Significant Area, Relevant Area and the Approach Zone, plus a newly defined Geomorphological Significant Area—Close Zone. By explicitly quantifying buffer zones and their hierarchical roles, the framework ensures consistent data collection across varied terrains and reduces ambiguity in landslide risk evaluation. Applied to 95 bridges in Tuscany and Basilicata, the framework offers standardized definitions and dimensions for Diagnostic Area, Geomorphological Significant Area and Relevant Area, based on detailed field surveys. Approach Zone and Geomorphological Significant Area—Close Zone are quantified as percentages of Relevant Area and Geomorphological Significant Area, supporting efficient, reproducible inspections using both manual and UAV-assisted methods. The Geomorphological Significant Area—Close Zone distinguishes core data, which requires direct surveys, from supplementary data that can be analyzed remotely or in the office. This distinction ensures that essential hazards are observed directly, while supplementary insights are efficiently integrated, enhancing field reliability and desk-based analysis. This integrated approach enhances the accuracy of landslide susceptibility assessment and the classification of attention levels, supporting the maintenance of the national IFFI. Ultimately, the comparison of IFFI catalog data, available in the Diagnostic Area, Geomorphological Significant Area, and Relevant Area, revealed previously unrecorded landslides in Matera and confirmed the reliability of the catalog in Lucca, highlighting that inventories can be systematically integrated only by using standardized areas with field verification to improve risk and infrastructure management. The structured framework bridges gaps between national inventory standards and localized survey needs, ensuring that both previously recorded and new landslide events are systematically captured. Full article

This study conducts a systematic bibliometric review of artificial intelligence (AI)-based approaches to tacit knowledge extraction and management. Drawing on data retrieved from Scopus and Web of Science, this study analyzes 126 publications published between 1985 and 2025 using VOSviewer and Biblioshiny to map citation networks, keyword co-occurrence patterns, and thematic evolution. The results identify nine major clusters spanning machine learning, natural language processing, semantic modeling, expert systems, knowledge-based decision support, and emerging hybrid techniques. Collectively, these findings indicate a field-wide shift from manual codification toward scalable, context-aware, and semantically enriched approaches that better support tacit knowing in organizational practice. Building on these insights, the paper introduces the AI–Tacit Knowledge Co-Evolution Model, which situates AI as an epistemic partner—augmenting human interpretive processes rather than merely codifying experience. The framework integrates Polanyi’s concept of tacit knowing, Nonaka’s SECI model, and sociotechnical learning theories to elucidate how human–AI interaction transforms the dynamics of knowledge creation. The review consolidates fragmented research streams and provides a conceptual foundation for guiding future methodological development in AI-enabled tacit knowledge management. Full article

The species, quantity, and tree diameter at breast height (DBH) are important indicators for assessing species distribution, individual growth status, and overall health in the forest. The existing tree information collection mainly relies on manual labor, which results in low efficiency and high labor intensity. To address these issues, we propose a method for tree species identification and diameter estimation by combining deep learning algorithms with binocular vision. First, an image acquisition platform is designed and integrated with a weeding machine to capture images during weeding operation. Images of seven types of trees are captured to develop a dataset. Second, a tree species identification model is established based on the YOLOv8n network, achieving 98.5% accuracy, 99.0% recall, and 99.2% mAP. Then, an improved YOLOv8n-seg model is proposed. It simplifies the network by introducing VanillaBlock in the backbone. FasterNet with a CCFM structure is added at the neck to enhance the model’s multi-scale expression capability. The mIoU of the improved model is 93.7%. Finally, the improved YOLOv8n-seg model is combined with binocular vision. After obtaining the segmentation mask of the tree, the spatial position of the two measurement points is calculated, allowing for the measurement of tree diameter. Verification experiments show that the average error for tree diameter ranges from 4.40~6.40 mm, and the proposed error compensation method can reduce diameter errors. This study provides a theoretical foundation and technical support for intelligent collection of tree information. Full article

Background: The clinical research laboratory plays a pivotal role in the execution of clinical studies. The accurate and consistent registration of patients is dependent on the competent use of laboratory equipment and manual techniques by technicians, ensuring the reliability of the data collected. To support these activities, the Groupe de Recherche Action en Santé (GRAS) has been registered with the College of American Pathologists (CAP) and the Clinical Laboratories Services (CLS) in Johannesburg, South Africa, for external proficiency testing (EPT) of its laboratory, as part of our commitment to quality assurance. The following report details the performance achievements over the past five years. Methods: Proficiency testing (PT) samples are dispatched to GRAS Lab three times a year (quarterly) and the results are generally returned within two to three weeks. In the field of parasitology, challenge specimens were prepared as follows: thick and thin blood films were stained with Giemsa and mounted with strips to protect them for multiple uses. Photographs, also known as whole slide images (WSIs), were also taken. For the biochemistry and haematology tests, a set of five samples were received for processing. All evaluations were carried out in accordance with the GRAS laboratory’s internal procedures. Results: The CAP laboratory’s performance in terms of the diagnosis of malaria and other blood parasites from 2020 to 2024 was 97.3% accurate (ranging from 93.33% to 100%), with 93.33%, 100%, 100%, 93.33% and 100% achieved in 2020, 2021, 2022, 2023 and 2024, respectively. The number of microscopists evaluated annually has been subject to variation according to operational staff at the time of evaluation. A total of 31 microscopists were enrolled in the CLS PT scheme, of which 73.9% were classified as ‘experts’ and 19.2% as ‘reference’ microscopists. In the field of haematology, the PT demonstrated 100% accuracy over the four-year study period. This outcome is indicative of the high-performance levels exhibited by the automated systems under scrutiny and the comparable nature of the data produced by these systems. The same trend was observed in the biochemistry PT results, with an overall score of 92.12%, ranging from 78% to 100%. Conclusions: Proficiency testing has been shown to be an effective tool for quality assurance in laboratories, helping to ensure the accuracy of malaria and other blood parasite diagnoses made by microscopists, as well as the results generated by automated systems. It has been instrumental in assisting laboratories in identifying issues related to test design and performance. Full article

The integrity of certification processes in the agrifood and fishing industries is essential for combating fraud, ensuring food safety, and meeting rising consumer expectations for transparency and sustainability. Yet, current certification systems remain fragmented, and they are vulnerable to tampering and highly dependent on manual or centralized procedures. This study addresses these gaps by providing a comprehensive survey that systematically classifies blockchain-based certification technologies and methodologies applied to the fisheries sector. The survey examines how the blockchain enhances trust through immutable record-keeping, smart contracts, and decentralized verification mechanisms, ensuring authenticity and accountability across the supply chain. Special attention is given to case studies and implementations that focus on ensuring food safety, verifying sustainability claims, and fostering consumer trust through transparent labeling. Furthermore, the paper identifies technological barriers, such as scalability and interoperability, and puts forward a collection of functional and non-functional requirements for holistic blockchain implementation. By providing a detailed overview of current trends and gaps, this study aims to guide researchers, industry stakeholders, and policymakers in adopting and optimizing blockchain technologies for certification. The findings highlight the potential of blockchain to innovate certification systems, easing the way for more resilient, sustainable, and consumer-centric agrifood and fishing industries. Full article

Quantifying tissue, molecular, and structural integrity is essential for biobank development. However, current assessment methods either involve destructive testing that depletes valuable biospecimens or rely on manual evaluations, which are not scalable and lead to interindividual variation. To overcome these challenges, we present PathQC, a deep-learning framework that directly predicts the tissue RNA Integrity Number (RIN) and the extent of autolysis from hematoxylin and eosin (H & E)-stained whole-slide images of normal tissue biopsies. Advancing over prior QC methods focused on imaging quality control, PathQC provides sample-quality control through the direct quantification of molecular integrity (RIN) and structural degradation (autolysis). PathQC first extracts morphological features from the slide using a recently developed digital pathology foundation model (UNI), followed by a supervised model that learns to predict RNA Integrity Number and autolysis scores from these morphological features. PathQC is trained on and applied to the Genotype-Tissue Expression (GTEx) cohort, which comprises 25,306 non-diseased post-mortem samples across 29 tissues from 970 donors, when paired ground-truth RIN and autolysis scores were available. Here, PathQC predicted RIN with an average Pearson correlation of 0.47 and an autolysis score of 0.45, with notably high performance using adrenal gland tissue (R = 0.82) for RIN and colon tissue (R = 0.83) for autolysis. We provide a pan-tissue model for predicting RIN and autolysis scores for new slides from any tissue type (GitHub). Overall, PathQC enables a scalable assessment of tissue molecular and structural integrity from routine H & E images, enhancing biobank quality control and retrospective analyses across 29 tissues and multiple collection sites. Full article

Accurate grading of corn kernels is critical for seed certification, directional seeding, and breeding, yet it is still predominantly performed by manual inspection. This work introduces CornViT, a three-stage Convolutional Vision Transformer (CvT) framework that emulates the hierarchical reasoning of human seed analysts for single-kernel evaluation. Three sequential CvT-13 classifiers operate on $384\times 384$ RGB images: Stage 1 distinguishes pure from impure kernels; Stage 2 categorizes pure kernels into flat and round morphologies; and Stage 3 determines the embryo orientation (up vs. down) for pure, flat kernels. Starting from a public corn seed image collection, we manually relabeled and filtered images to construct three stage-specific datasets: 7265 kernels for purity, 3859 pure kernels for morphology, and 1960 pure–flat kernels for embryo orientation, all released as benchmarks. Head-only fine-tuning of ImageNet-22k pretrained CvT-13 backbones yields test accuracies of 93.76% for purity, 94.11% for shape, and 91.12% for embryo-orientation detection. Under identical training conditions, ResNet-50 reaches only 76.56 to 81.02 percent, whereas DenseNet-121 attains 86.56 to 89.38 percent accuracy. These results highlight the advantages of convolution-augmented self-attention for kernel analysis. To facilitate adoption, we deploy CornViT in a Flask-based web application that performs stage-wise inference and exposes interpretable outputs through a browser interface. Together, the CornViT framework, curated datasets, and web application provide a deployable solution for automated corn kernel quality assessment in seed quality workflows. Source code and data are publicly available. Full article

Traditional segmentation methods are slow and rely on manual annotations, which are labor-intensive. To address these limitations, we propose YOLO-SAM AgriScan, a unified framework that combines the fast object detection capabilities of YOLOv11 with the zero-shot segmentation power of the Segment Anything Model 2 (SAM2). Our approach adopts a hybrid paradigm for on-plant ripe strawberry segmentation, wherein YOLOv11 is fine-tuned using a few-shot learning strategy with minimal annotated samples, and SAM2 performs mask generation without additional supervision. This architecture eliminates the bottleneck of pixel-wise manual annotation and enables the scalable and efficient segmentation of strawberries in both controlled and natural farm environments. Experimental evaluations on two datasets, a custom-collected dataset and a publicly available benchmark, demonstrate strong detection and segmentation performance in both full-data and data-constrained scenarios. The proposed framework achieved a mean Dice score of 0.95 and an IoU of 0.93 on our collected dataset and maintained competitive performance on public data (Dice: 0.95, IoU: 0.92), demonstrating its robustness, generalizability, and practical relevance in real-world agricultural settings. Our results highlight the potential of combining few-shot detection and zero-shot segmentation to accelerate the development of annotation-light, intelligent phenotyping systems. Full article

Visual inspection is a crucial quality assurance process across many manufacturing industries. While many companies now employ machine learning-based systems, they face a significant challenge, particularly in safety-critical domains. The outcomes of these systems are often complex and difficult to comprehend, making them less reliable and trustworthy. To address this challenge, we build on our previously proposed R4VR-framework and provide practical, step-by-step guidelines that enable the safe and efficient implementation of machine learning in visual inspection tasks, even when starting from scratch. The framework leverages three complementary safety mechanisms—uncertainty detection, explainability, and model diversity—to enhance both accuracy and system safety while minimizing manual effort. Using the example of steel surface inspection, we demonstrate how a self-accelerating process of data collection where model performance improves while manual effort decreases progressively can arise. Based on that, we create a system with various safety mechanisms where less than 0.1% of images are classified wrongly and remain undetected. We provide concrete recommendations and an open-source code base to facilitate reproducibility and adaptation to diverse industrial contexts. Full article

Background: Preventing postoperative infections in hand surgery is an important factor for achieving sustainable results of surgical procedures. To prevent infections, especially when implants are used, preoperative prophylactic antibiotics are applied in adherence to predominantly national guidelines, which are not specifically tailored to hand surgery. However, several studies related to elective soft tissue hand surgery indicate that the preoperative use of antibiotics does not reduce the incidence of postoperative infections. Evidence regarding their efficacy in osteosynthesis of the hand and wrist remains limited. Methods: In this retrospective study, we analyzed 542 adult patients who underwent hand or wrist osteosynthesis between 2016 and 2019 at our university center. They were enrolled in an antibiotic treatment group and a control group without antibiotic treatment. The prophylaxis group (P) underwent surgery in the main operating theater under intravenous anesthesia, whereas the non-prophylaxis group (NP) was treated under WALANT (Wide Awake Local Anesthetic No Tourniquet) in an outpatient operating theater without receiving preoperative antibiotics. Theater construction and installation were otherwise similar, and both were classified as grade 1 theaters. We applied propensity modeling and inverse probability of treatment weighting (IPTW) to achieve balanced treatment groups with respect to risk factors for infection, and we calculated the odds ratio of prophylaxis and infection. Inclusion factors for risk of infection were age, female sex, smoking, diabetes, metabolic disease, inflammatory disease, substance abuse, cardiovascular disease, hepatopathy, renal disease, polytrauma, open fracture, being a manual worker, and occupational accidents. To assess the severity of the cases, we considered whether the fractures were intraarticular, multi-fragmentary, or open, and we collected data on the types of surgical implants that were used. Results: No significant association was found between antibiotic prophylaxis and postoperative infection rate (infection rate P: 3.86%; NP: 3.27%; unadjusted OR: 1.19; adjusted OR after IPTW: 1.09). In terms of risk factors, there was an insignificant trend of higher infection rates in the subgroups smoking, cardiovascular disease, open fracture, occupational accident, and open fixations. Conclusions: In this cohort, routine use of preoperative antibiotics in hand osteosynthesis did not reduce infection rates. The effectiveness of the widespread standardized application of prophylactic antibiotics to reduce the risk of postoperative infections in osteosynthesis of the hand and wrist remains debatable. Our findings set the basis for further prospective studies aiming at clearer guidelines for evidence-based perioperative patient care. Full article

Background/Objectives: Demodex mites are a common yet underdiagnosed cause of ocular surface diseases, including blepharitis and meibomian gland dysfunction (MGD). Traditional diagnosis via microscopic examination is labor-intensive and time-consuming. This study aimed to develop a deep learning-based system for the automated detection and quantification of Demodex mites from microscopic eyelash images. Methods: We collected 1610 microscopic images of eyelashes from patients clinically suspected to have ocular demodicosis. After quality screening, 665 images with visible Demodex features were annotated and processed. Two deep learning models, YOLOv11 and RT-DETR, were trained and evaluated using standard metrics. Grad-CAM visualization was applied to confirm model attention and feature localization. Results: Both YOLO and RT-DETR models were able to detect Demodex mites in our microscopic images. The YOLOv11 boxing model revealed an average precision of 0.9441, sensitivity of 0.9478, and F1-score of 0.9459 in our detection system, while the RT-DETR model showed an average precision of 0.7513, sensitivity of 0.9389, and F1-score of 0.8322. Moreover, Grad-CAM visualization confirmed the models’ focus on relevant mite features. Quantitative analysis enabled consistent mite counting across overlapping regions, with a confidence level of 0.4–0.8, confirming stable enumeration performance. Conclusions: The proposed artificial intelligence (AI)-based detection system demonstrates strong potential for assisting ophthalmologists in diagnosing ocular demodicosis efficiently and accurately, reducing reliance on manual microscopy and enabling faster clinical decision making. Full article

Educational chatbots are increasingly deployed to scaffold student learning, yet educators lack scalable ways to assess the cognitive depth of these dialogues in situ. Bloom’s taxonomy provides a principled lens for characterizing reasoning, but manual tagging of conversational turns is costly and difficult to scale for learning analytics. We present a reproducible high-confidence pseudo-labeling pipeline for multi-label Bloom classification of Socratic student–chatbot exchanges. The dataset comprises 6716 utterances collected from conversations between a Socratic chatbot and 34 undergraduate statistics students at Nanyang Technological University. From three chronologically selected workbooks with expert Bloom annotations, we trained and compared two labeling tracks: (i) a calibrated classical approach using SentenceTransformer (all-MiniLM-L6-v2) embeddings with one-vs-rest Logistic Regression, Linear SVM, XGBoost, and MLP, followed by per-class precision–recall threshold tuning; and (ii) a lightweight LLM track using GPT-4o-mini after supervised fine-tuning. Class-specific thresholds tuned on 5-fold cross-validation were then applied in a single pass to assign high-confidence pseudo-labels to the remaining unlabeled exchanges, avoiding feedback-loop confirmation bias. Fine-tuned GPT-4o-mini achieved the highest prevalence-weighted performance (micro-F1 $=0.814$), whereas calibrated classical models yielded stronger balance across Bloom levels (best macro-F1 $=0.630$ with Linear SVM; best classical micro-F1 $=0.759$ with Logistic Regression). Both model families reflect the corpus skew toward lower-order cognition, with LLMs excelling on common patterns and linear models better preserving rarer higher-order labels, while results should be interpreted as a proof-of-concept given limited gold labeling, the approach substantially reduces annotation burden and provides a practical pathway for Bloom-aware learning analytics and future real-time adaptive chatbot support. Full article

Background: Endobronchial ultrasound-guided transbronchial biopsy (EBUS-TBB) is a valuable technique for diagnosing peripheral pulmonary lesions (PPLs). Although computer-aided diagnostic (CAD) systems have been explored for EBUS interpretation, most rely on manually selected 2D static frames and overlook temporal dynamics that may provide important cues for differentiating benign from malignant lesions. This study aimed to develop an artificial intelligence model that incorporates temporal modeling to analyze EBUS videos and improve lesion classification. Methods: We retrospectively collected EBUS videos from patients undergoing EBUS-TBB between November 2019 and January 2022. A dual-path 3D convolutional network (SlowFast) was employed for spatiotemporal feature extraction, and contrastive learning (SwAV) was integrated to enhance model generalizability on clinical data. Results: A total of 465 patients with corresponding EBUS videos were included. On the validation set, the SlowFast + SwAV_Frame model achieved an AUC of 0.857, accuracy of 82.26%, sensitivity of 93.18%, specificity of 55.56%, and F1-score of 88.17%, outperforming pulmonologists (accuracy 70.97%, sensitivity 77.27%, specificity 55.56%, F1-score 79.07%). On the test set, the model achieved an AUC of 0.823, accuracy of 76.92%, sensitivity of 84.85%, specificity of 63.16%, and F1-score of 82.35%. The proposed model also demonstrated superior performance compared with conventional 2D architectures. Conclusions: This study introduces the first CAD framework for real-time malignancy classification from full-length EBUS videos, which reduces reliance on manual image selection and improves diagnostic efficiency. In addition, given its higher accuracy compared with pulmonologists’ assessments, the framework shows strong potential for clinical applicability. Full article

Morphometric traits are closely linked to body condition, health, welfare, and productivity in livestock. In recent years, contactless 3D reconstruction technologies have been increasingly adopted to improve the accuracy and efficiency of morphometric evaluations. Conventional approaches for 3D reconstruction mainly employ Light Detection and Ranging (LiDAR) or photogrammetry. In contrast, emerging Artificial Intelligence (AI)-based methods, such as Neural Surface Reconstruction, 3D Gaussian Splatting, and Neural Radiance Fields, offer new opportunities for high-fidelity digital modeling. Smartphones’ affordability represents a cost-effective and portable platform for deploying these advanced tools, potentially supporting enhanced agricultural performance, accelerating sector digitalization, and thus reducing the urban–rural digital gap. This preliminary study assessed the viability of using smartphone-based LiDAR, photogrammetry, and AI models to obtain body measurements of Marchigiana cattle. Five morphometric traits manually collected on animals were compared with those extracted from smartphone-based 3D reconstructions. LiDAR measurements offer more consistent estimates, with relative error ranging from −1.55% to 4.28%, while photogrammetry demonstrated accuracy ranging from 0.75 to −14.56. AI-based models (NSR, 3DGS, NeRF) reported more variability between accuracy results, pointing to the need for further refinement. Overall, the results highlight the preliminary potential of portable 3D scanning technologies, particularly LiDAR-equipped smartphones, for non-invasive morphometric data collection in cattle. Full article