Search Results (21,274)

The retina offers a unique window into both ocular and systemic health, motivating the development of AI-based tools for disease screening and risk assessment. In this study, we present a comprehensive evaluation of six state-of-the-art deep neural networks, including convolutional neural networks and vision transformer architectures, on the Brazilian Multilabel Ophthalmological Dataset (BRSET), comprising 16,266 fundus images annotated for multiple clinical and demographic labels. We explored seven classification tasks: Diabetes, Diabetic Retinopathy (2-class), Diabetic Retinopathy (3-class), Hypertension, Hypertensive Retinopathy, Drusen, and Sex classification. Models were evaluated using precision, recall, F1-score, accuracy, and AUC. Among all models, the Swin-L generally delivered the best performance across scenarios for Diabetes (AUC = 0.88, weighted F1-score = 0.86), Diabetic Retinopathy (2-class) (AUC = 0.98, weighted F1-score = 0.95), Diabetic Retinopathy (3-class) (macro AUC = 0.98, weighted F1-score = 0.95), Hypertension (AUC = 0.85, weighted F1-score = 0.79), Hypertensive Retinopathy (AUC = 0.81, weighted F1-score = 0.97), Drusen detection (AUC = 0.93, weighted F1-score = 0.90), and Sex classification (AUC = 0.87, weighted F1-score = 0.80). These results reflect excellent to outstanding diagnostic performance. We also employed gradient-based saliency maps to enhance explainability and visualize decision-relevant retinal features. Our findings underscore the potential of deep learning, particularly vision transformer models, to deliver accurate, interpretable, and clinically meaningful screening tools for retinal and systemic disease detection. Full article

Background and Objectives: Lower limb defects often present significant reconstructive challenges due to limited soft tissue availability and exposure of critical structures. Perforator-based flaps offer reliable solutions, with minimal donor site morbidity. This study aimed to evaluate the efficacy of infrared thermography (IRT) in preoperative planning and postoperative monitoring of perforator-based flaps, assessing its accuracy in identifying perforators, predicting complications, and optimizing outcomes. Materials and Methods: A prospective observational study was conducted on 76 patients undergoing lower limb reconstruction with fascio-cutaneous perforator flaps between 2022 and 2024. Perforator mapping was performed concurrently with IRT and Doppler ultrasonography (D-US), with intraoperative confirmation. Flap design variables and systemic parameters were recorded. Postoperative monitoring employed thermal imaging on days 1 and 7. Outcomes were correlated with thermal, anatomical, and systemic factors using statistical analyses, including t-tests and Pearson correlation. Results: IRT showed high sensitivity (97.4%) and positive predictive value (96.8%) for perforator detection. A total of nine minor complications occurred, predominantly in patients with diabetes mellitus and/or elevated glycemia (p = 0.05). Larger flap-to-defect ratios (A/C and B/C) correlated with increased complications in propeller flaps, while smaller ratios posed risks for V-Y and Keystone flaps. Thermal analysis indicated significantly lower flap temperatures and greater temperature gradients in flaps with complications by postoperative day 7 (p < 0.05). CRP levels correlated with glycemia and white blood cell counts, highlighting systemic inflammation’s impact on outcomes. Conclusions: IRT proves to be a reliable, non-invasive method for perforator localization and flap monitoring, enhancing surgical planning and early complication detection. Combined with D-US, it improves perforator selection and perfusion assessment. Thermographic parameters, systemic factors, and flap design metrics collectively predict flap viability. Integration of IRT into surgical workflows offers a cost-effective tool for optimizing reconstructive outcomes in lower limb surgery. Full article

The classification of submarine topography and geomorphology is essential for marine resource exploitation and ocean engineering, with wide-ranging implications in marine geology, disaster assessment, resource exploration, and autonomous underwater navigation. Submarine landscapes are highly complex and diverse. Traditional visual interpretation methods are not only inefficient and subjective but also lack the precision required for high-accuracy classification. While many machine learning and deep learning models have achieved promising results in image classification, limited work has been performed on integrating backscatter and bathymetric data for multi-source processing. Existing approaches often suffer from high computational costs and excessive hyperparameter demands. In this study, we propose a novel approach that integrates pruning-enhanced ConDenseNet with label smoothing regularization to reduce misclassification, strengthen the cross-entropy loss function, and significantly lower model complexity. Our method improves classification accuracy by 2% to 10%, reduces the number of hyperparameters by 50% to 96%, and cuts computation time by 50% to 85.5% compared to state-of-the-art models, including AlexNet, VGG, ResNet, and Vision Transformer. These results demonstrate the effectiveness and efficiency of our model for multi-source submarine topography classification. Full article

Optical Coherence Tomography (OCT) has evolved from a breakthrough ophthalmologic imaging tool into a cornerstone technology in interventional cardiology. After its initial applications in retinal imaging in the early 1990s, OCT was subsequently envisioned for cardiovascular use. In 1995, its ability to visualize atherosclerotic plaques was demonstrated in an in vitro study, and the following year marked the acquisition of the first in vivo OCT image of a human coronary artery. A major milestone followed in 2000, with the first intracoronary imaging in a living patient using time-domain OCT. However, the real inflection point came in 2006 with the advent of frequency-domain OCT, which dramatically improved acquisition speed and image quality, enabling safe and routine imaging in the catheterization lab. With the advent of high-resolution, second-generation frequency-domain systems, OCT has become clinically practical and widely adopted in catheterization laboratories. OCT progressively entered interventional cardiology, first proving its safety and feasibility, then demonstrating superiority over angiography alone in guiding percutaneous coronary interventions and improving outcomes. Today, it plays a central role not only in clinical practice but also in cardiovascular research, enabling precise assessment of plaque biology and response to therapy. With the advent of artificial intelligence and hybrid imaging systems, OCT is now evolving into a true precision-medicine tool—one that not only guides today’s therapies but also opens new frontiers for discovery, with vast potential still waiting to be explored. Tracing its historical evolution from ophthalmology to cardiology, this narrative review highlights the key technological milestones, clinical insights, and future perspectives that position OCT as an indispensable modality in contemporary interventional cardiology. As a guiding thread, the myth of Prometheus is used to symbolize the evolution of OCT—from its illuminating beginnings in ophthalmology to its transformative role in cardiology—as a metaphor for how light, innovation, and knowledge can reveal what was once hidden and redefine clinical practice. Full article

Monitoring sediment concentration in water bodies is crucial for assessing water quality, ecosystems, and environmental health. However, physical sampling and sensor-based approaches are labor-intensive and unsuitable for large-scale, continuous monitoring. This study employs machine learning models to estimate suspended sediment concentration using images captured in natural light, named RGB, and near-infrared (NIR) conditions. A controlled dataset of approximately 1300 images with SSC values ranging from 1000 mg/L to 150,000 mg/L was developed, incorporating temperature, time of image capture, and solar irradiance as additional features. Random forest regression and gradient boosting regression were trained on mean RGB values, red reflectance, time of captured, and temperature for natural light images, achieving up to 72.96% accuracy within a 30% relative error. In contrast, NIR images leveraged gray-level co-occurrence matrix texture features and temperature, reaching 83.08% accuracy. Comparative analysis showed that ensemble models outperformed deep learning models like Convolutional Neural Networks and Multi-Layer Perceptrons, which struggled with high-dimensional feature extraction. These findings suggest that using machine learning models and RGB and NIR imagery offers a scalable, non-invasive, and cost-effective way of sediment monitoring in support of water quality assessment and environmental management. Full article

The ongoing secular changes in human movement development means that an assessment of primitive reflexes is now required not only in disabled but also in neurotypical children. This study had three aims: (1) presenting the TASHUN test battery as suitable for the assessment of primitive reflex activity in normal children and child athletes; (2) analyzing reflex characteristics of neurotypical children; (3) verifying validity and reliability of tests. Spearman’s rank correlation and ROC analysis were used for validation. Intraclass Correlation Coefficient and RM ANOVA analyzed reliability. The test on 242 schoolgirls has demonstrated that retained primitive reflexes are present in almost every individual (84.7–95.7%). Correlations showed strong positive association, with all values exceeding 0.8, and ROC analysis demonstrated excellent predictive strength (AUC values over 0.9). Interobserver reliability showed excellent agreement (ICC values above 0.9). No significant offset was present among the scoring by evaluators. Therefore, testing for primitive reflexes is necessary in neurotypical children in order to obtain a realistic image about the physiology of reflexes and their role in motor development. Our screening could be useful for practicing sport professionals, researchers and academics, to identify deficiencies, to further explore reflexes and to train future PE teachers and trainers. Full article

Museums play an essential role in preserving both cultural and natural heritage, safeguarding samples that offer invaluable insights into our history and scientific understanding. The integration of micro-computed tomography (micro-CT) has significantly advanced the study, restoration, and conservation of these priceless objects. This work explores the application of micro-CT across three critical areas of museum practice: sample virtualization, restoration assessment, and the analysis of fossil specimens. Specifically, micro-CT scanning was applied to fossils stored in the G.A.M.P.S. collection (Scandicci, Italy), enabling the creation of highly detailed non-invasive 3D models for digital archiving and virtual exhibitions. At the Opificio delle Pietre Dure in Florence, micro-CT was employed to evaluate fossil bone restoration treatments, focusing on the internal impact of menthol as a consolidant and its effects on the structural integrity of the material. Furthermore, micro-CT was utilized to investigate a sealed bee preserved in its cocoon within a paleosol in Costa Vicentina (Portugal), providing unprecedented insights into its internal anatomy and state of preservation, all while maintaining the integrity of the specimen. The results of this study underscore the versatility of micro-CT as a powerful non-destructive tool for advancing the fields of conservation, restoration, and scientific analysis of cultural and natural heritage. By integrating high-resolution imaging with both virtual and hands-on conservation strategies, micro-CT empowers museums to enhance research capabilities, improve preservation methodologies, and foster greater public engagement with their collections. Full article

Stress in pigs poses significant challenges to animal welfare and productivity in modern pig farming, contributing to increased antimicrobial use and the rise of antimicrobial resistance (AMR). This study involves stress classification in pregnant sows by exploring five deep learning models: ConvNeXt, EfficientNet_V2, MobileNet_V3, RegNet, and Vision Transformer (ViT). These models are used for stress detection from facial images, leveraging an expanded dataset. A facial image dataset of sows was collected at Scotland’s Rural College (SRUC) and the images were categorized into primiparous Low-Stressed (LS) and High-Stress (HS) groups based on expert behavioural assessments and cortisol level analysis. The selected deep learning models were then trained on this enriched dataset and their performance was evaluated using cross-validation on unseen data. The Vision Transformer (ViT) model outperformed the others across the dataset of annotated facial images, achieving an average accuracy of 0.75, an F1 score of 0.78 for high-stress detection, and consistent batch-level performance (up to 0.88 F1 score). These findings highlight the efficacy of transformer-based models for automated stress detection in sows, supporting early intervention strategies to enhance welfare, optimize productivity, and mitigate AMR risks in livestock production. Full article

pH is a critical parameter requiring precise monitoring across scientific, industrial, and biological domains. Fluorescent pH probes offer a powerful alternative to traditional methods (e.g., electrodes, indicators), overcoming limitations in miniaturization, long-term stability, and electromagnetic interference. By utilizing photophysical mechanisms—including intramolecular charge transfer (ICT), photoinduced electron transfer (PET), and fluorescence resonance energy transfer (FRET)—these probes enable high-sensitivity, reusable, and biocompatible sensing. This review systematically details recent advances, categorizing probes by operational pH range: strongly acidic (0–3), weakly acidic (3–7), strongly alkaline (>12), weakly alkaline (7–11), near-neutral (6–8), and wide-dynamic range. Innovations such as ratiometric detection, organelle-specific targeting (lysosomes, mitochondria), smartphone colorimetry, and dual-analyte response (e.g., pH + Al³⁺/CN⁻) are highlighted. Applications span real-time cellular imaging (HeLa cells, zebrafish, mice), food quality assessment, environmental monitoring, and industrial diagnostics (e.g., concrete pH). Persistent challenges include extreme-pH sensing (notably alkalinity), photobleaching, dye leakage, and environmental resilience. Future research should prioritize broadening functional pH ranges, enhancing probe stability, and developing wide-range sensing strategies to advance deployment in commercial and industrial online monitoring platforms. Full article

Background: Pancreatic cystic lesions (PCLs), including intraductal papillary mucinous neoplasms (IPMNs) and mucinous cystic neoplasms (MCNs), pose a diagnostic challenge due to their variable malignant potential. Current guidelines, such as Fukuoka and American Gastroenterological Association (AGA), have moderate predictive accuracy and may lead to overtreatment or missed malignancies. Artificial intelligence (AI), incorporating machine learning (ML) and deep learning (DL), offers the potential to improve risk stratification, diagnosis, and management of PCLs by integrating clinical, radiological, and molecular data. This is the first systematic review to evaluate the application, performance, and clinical utility of AI models in the diagnosis, classification, prognosis, and management of pancreatic cysts. Methods: A systematic review was conducted in accordance with PRISMA guidelines and registered on PROSPERO (CRD420251008593). Databases searched included PubMed, EMBASE, Scopus, and Cochrane Library up to March 2025. The inclusion criteria encompassed original studies employing AI, ML, or DL in human subjects with pancreatic cysts, evaluating diagnostic, classification, or prognostic outcomes. Data were extracted on the study design, imaging modality, model type, sample size, performance metrics (accuracy, sensitivity, specificity, and area under the curve (AUC)), and validation methods. Study quality and bias were assessed using the PROBAST and adherence to TRIPOD reporting guidelines. Results: From 847 records, 31 studies met the inclusion criteria. Most were retrospective observational (n = 27, 87%) and focused on preoperative diagnostic applications (n = 30, 97%), with only one addressing prognosis. Imaging modalities included Computed Tomography (CT) (48%), endoscopic ultrasound (EUS) (26%), and Magnetic Resonance Imaging (MRI) (9.7%). Neural networks, particularly convolutional neural networks (CNNs), were the most common AI models (n = 16), followed by logistic regression (n = 4) and support vector machines (n = 3). The median reported AUC across studies was 0.912, with 55% of models achieving AUC ≥ 0.80. The models outperformed clinicians or existing guidelines in 11 studies. IPMN stratification and subtype classification were common focuses, with CNN-based EUS models achieving accuracies of up to 99.6%. Only 10 studies (32%) performed external validation. The risk of bias was high in 93.5% of studies, and TRIPOD adherence averaged 48%. Conclusions: AI demonstrates strong potential in improving the diagnosis and risk stratification of pancreatic cysts, with several models outperforming current clinical guidelines and human readers. However, widespread clinical adoption is hindered by high risk of bias, lack of external validation, and limited interpretability of complex models. Future work should prioritise multicentre prospective studies, standardised model reporting, and development of interpretable, externally validated tools to support clinical integration. Full article

To address the inefficiency of traditional concrete crack detection methods and the heavy reliance of supervised learning on extensive labeled data, in this study, an intelligent assessment method of concrete damage based on pseudo-label semi-supervised learning and fractal geometry theory is proposed to solve two core tasks: one is binary classification of pixel-level cracks, and the other is multi-category assessment of damage state based on crack morphology. Using three-channel RGB images as input, a dual-path collaborative training framework based on U-Net encoder–decoder architecture is constructed, and a binary segmentation mask of the same size is output to achieve the accurate segmentation of cracks at the pixel level. By constructing a dual-path collaborative training framework and employing a dynamic pseudo-label refinement mechanism, the model achieves an F1-score of 0.883 using only 50% labeled data—a mere 1.3% decrease compared to the fully supervised benchmark DeepCrack (F1 = 0.896)—while reducing manual annotation costs by over 60%. Furthermore, a quantitative correlation model between crack fractal characteristics and structural damage severity is established by combining a U-Net segmentation network with the differential box-counting algorithm. The experimental results demonstrate that under a cyclic loading of 147.6–221.4 kN, the fractal dimension monotonically increases from 1.073 (moderate damage) to 1.189 (failure), with 100% accuracy in damage state identification, closely aligning with the degradation trend of macroscopic mechanical properties. In complex crack scenarios, the model attains a recall rate (Re = 0.882), surpassing U-Net by 13.9%, with significantly enhanced edge reconstruction precision. Compared with the mainstream models, this method effectively alleviates the problem of data annotation dependence through a semi-supervised strategy while maintaining high accuracy. It provides an efficient structural health monitoring solution for engineering practice, which is of great value to promote the application of intelligent detection technology in infrastructure operation and maintenance. Full article

Large soil pores critically influence water and solute transport in soils. The presence of preferential flow paths created by soil macropores can profoundly impact water quality, underscoring the necessity of accurately assessing the characteristics of these macropores. However, it remains unclear whether variations in macropore structure exist between different altitudes and positions of terraced paddy fields. The primary objective of this research was to utilize X-ray computed tomography (CT) and image analysis techniques to characterize the soil pore structure at both the inner field and ridge positions across different altitude levels (high, medium, and low altitude) within terraced paddy fields. The results indicate that there are significant differences in the distribution of large soil pores at different altitudes, with large pores concentrated in the surface layer (0–10 cm) in low-altitude areas, while in high-altitude areas, the distribution of large pores is more uniform. Additionally, as altitude increases, the porosity of large pores shows an increasing trend. The three-dimensional equivalent diameter and large pore volume are primarily characterized by large pores ranging from 1 to 2 mm and 0 to 5 mm³, respectively, with their morphology predominantly appearing spherical or ellipsoidal. The connectivity of large pores in the surface layer of paddy soil is stronger than that in the bunds. However, this connectivity gradually weakens with increasing soil depth. The findings from this study provide valuable quantitative insights into the unique characteristics of soil macropores that vary according to the altitude and position in terraced paddy fields. Moreover, this study emphasizes the necessity for future research that encompasses a broader range of soil types, altitudes, and terraced paddy locations to validate and further explore the identified relationships between altitude and macropore characteristics. Full article

This study applies Computational Fluid Dynamics (CFD) and mathematical modeling to examine uterine and umbilical arterial blood flow during pregnancy, providing a more detailed understanding of hemodynamic changes across gestation. Statistical analysis of Doppler ultrasound data from a large cohort of more than 200 pregnant women (in the second and third trimesters) reveals significant increases in the umbilical arterial peak systolic velocity ($PSV$) between the 22nd and 30th weeks, while uterine artery velocities remain relatively stable, suggesting adaptations in vascular resistance during pregnancy. By combining the Navier–Stokes equations with Doppler ultrasound-derived inlet velocity profiles, we quantify several key fluid dynamics parameters, including time-averaged wall shear stress ($TAWSS$), oscillatory shear index ($OSI$), relative residence time ($RRT$), Reynolds number ($Re$), and Dean number ($De$), evaluating laminar flow stability in the uterine artery and secondary flow patterns in the umbilical artery. Since blood exhibits shear-dependent viscosity and complex rheological behavior, modeling it as a non-Newtonian fluid is essential to accurately capture pulsatile flow dynamics and wall shear stresses in these vessels. Unlike conventional imaging techniques, CFD offers enhanced visualization of blood flow characteristics such as streamlines, velocity distributions, and instantaneous particle motion, providing insights that are not easily captured by Doppler ultrasound alone. Specifically, CFD reveals secondary flow patterns in the umbilical artery, which interact with the primary flow, a phenomenon that is challenging to observe with ultrasound. These findings refine existing hemodynamic models, provide population-specific reference values for clinical assessments, and improve our understanding of the relationship between umbilical arterial flow dynamics and fetal growth restriction, with important implications for maternal and fetal health monitoring. Full article

Water treadmill (WT) exercise is used for horses’ rehabilitation and training. Given that each training needs to be individualized for each horse, the goal is to assess whether infrared thermography (IRT) can serve as a non-invasive tool for daily monitoring of individual training and rehabilitation progress in horses undergoing WT exercise. Fifteen Polish Warmblood school horses were subjected to five WT sessions: dry treadmill, fetlock-depth water, fetlock-depth water with artificial river (AR), carpal-depth water, and carpal-depth water with AR. IRT images, collected pre- and post-exercise, were analyzed for the mean temperature (Tmean) and maximal temperature (Tmax) across 14 regions of interest (ROIs) representing the body surface overlying specific superficial muscles. While on a dry treadmill, Tmean and Tmax increased post-exercise in all ROIs; wetting of the hair coat limited surface temperature analysis in ROIs annotated on limbs. Tmax over the m. brachiocephalicus, m. trapezius pars cervicalis, m. triceps brachii, and m. semitendinosus increased during walking in carpal-depth water, which therefore may be suggested as an indirect indicator of increased activity related to forelimb protraction and flexion–extension of the limb joints. Tmax over the m. latissimus dorsi and m. longissimus increased during carpal-depth WT exercise with active AR mode, which may be suggested as an indicator of increased workload including vertical displacement of the trunk. Full article

Timely and accurate identification of surface damage in hydraulic structures is essential for maintaining structural integrity and ensuring operational safety. Traditional manual inspections are time-consuming, labor-intensive, and prone to subjectivity, especially for large-scale or inaccessible infrastructure. Leveraging advancements in aerial imaging, unmanned aerial vehicles (UAVs) enable efficient acquisition of high-resolution visual data across expansive hydraulic environments. However, existing deep learning (DL) models often lack architectural adaptations for the visual complexities of UAV imagery, including low-texture contrast, noise interference, and irregular crack patterns. To address these challenges, this study proposes a lightweight, robust, and high-precision segmentation framework, called LFPA-EAM-Fast-SCNN, specifically designed for pixel-level damage detection in UAV-captured images of hydraulic concrete surfaces. The developed DL-based model integrates an enhanced Fast-SCNN backbone for efficient feature extraction, a Lightweight Feature Pyramid Attention (LFPA) module for multi-scale context enhancement, and an Edge Attention Module (EAM) for refined boundary localization. The experimental results on a custom UAV-based dataset show that the proposed damage detection method achieves superior performance, with a precision of 0.949, a recall of 0.892, an F1 score of 0.906, and an IoU of 87.92%, outperforming U-Net, Attention U-Net, SegNet, DeepLab v3+, I-ST-UNet, and SegFormer. Additionally, it reaches a real-time inference speed of 56.31 FPS, significantly surpassing other models. The experimental results demonstrate the proposed framework’s strong generalization capability and robustness under varying noise levels and damage scenarios, underscoring its suitability for scalable, automated surface damage assessment in UAV-based remote sensing of civil infrastructure. Full article