Search Results (4,788)

Background/Objectives: Anterior segment optical coherence tomography (AS-OCT) and optical coherence tomography angiography (AS-OCTA) have enhanced the evaluation of the cornea, ocular surface, and ocular surface diseases (OSD), offering high-resolution structural and anterior segment vascular imaging. This review summarizes recent advances in these modalities and their clinical applications. Methods: A comprehensive literature search was conducted using PubMed, Web of Science, and Google Scholar with the terms OCT, OCTA, anterior segment, and ocular surface disease. Studies published in the past five years were included, emphasizing more recent developments such as ultra-high-resolution AS-OCT (UHR-AS-OCT) and swept-source AS-OCTA technologies. Results: UHR-AS-OCT provides non-invasive, sub-micron imaging of the cornea and the ocular surface, including tear film morphology and epithelial thickness to correlate with clinical tests such as tear break-up time, and fluorescein staining. Advances in AS-OCTA allow dye-free, depth-resolved imaging of corneal and conjunctival vasculature. These vascular biomarkers have shown promising utility in conditions such as limbal stem cell deficiency, chemical ocular injury, and ocular surface squamous neoplasia. Improvements in image acquisition, motion correction, and segmentation algorithms have enhanced accuracy and repeatability, supporting broader clinical translation. Conclusions: AS-OCT and AS-OCTA have become useful adjunctive imaging tools for the cornea and ocular surface evaluation. Their non-invasive, quantitative, and reproducible metrics may enable earlier diagnosis, objective staging, and longitudinal monitoring of OSD. Integration of OCT-based imaging with artificial intelligence and multimodal data, including tear proteomics and meibography, may optimize personalized treatment for ocular surface disorders. Full article

Soil salinization severely threatens agricultural sustainability in saline–alkali regions, and high-throughput, efficient screening of salt-tolerant rice varieties is critical to mitigating this threat. Traditional evaluation methods are constrained by low throughput, limited spatiotemporal resolution, and the lack of standardized indicators. To address these gaps, this study established a multi-scale spectral phenotyping framework integrating ground-based hyperspectral, UAV-borne multispectral, and Sentinel-2 satellite remote sensing data for high-throughput screening of salt-tolerant rice. Field experiments were conducted with 12 rice lines at five key growth stages in Ningxia, China, with synchronous ground spectral measurements and UAV image acquisition on the same day for each stage. Five feature selection methods were employed to screen salt stress-sensitive hyperspectral bands, with classification accuracy validated via a Support Vector Machine (SVM) model. The results showed that: (1) rice spectral characteristics varied dynamically across growth stages, and first-order differential transformation effectively amplified subtle spectral variations in stress-sensitive regions; (2) the Minimum Redundancy–Maximum Relevance (mRMR) method outperformed other methods, achieving 100% classification accuracy at key growth stages, with sensitive bands dominated by red edge bands (58.33%); (3) the constructed Salt Stress Index (SIR) showed strong correlations with classical vegetation indices and rice yield, and could clearly distinguish salt-tolerant and salt-sensitive rice varieties, with stable performance against field environmental noise; and (4) band matching between UAV and Sentinel-2 data enabled multi-scale data fusion and regional-scale salt stress monitoring. This framework realizes the transformation from qualitative spectral description to quantitative salt tolerance evaluation, providing standardized technical support for salt-tolerant rice breeding and precision management of saline–alkali lands. Full article

In the field of orthopedic surgery, particularly distraction osteogenesis (DO), the mechanical environment plays a decisive role in the quality of bone regeneration and the safety of the soft tissue envelope. The continuous monitoring and accurate prediction of distraction resisting forces (DRF) are critical for preventing soft tissue complications such as nerve ischemia, joint contractures, and mechanical failure of the lengthening device. However, current clinical practice relies heavily on subjective assessment or passive monitoring tools that lack predictive capabilities. To address this gap, this study proposes a comprehensive solution combining a custom mechanical acquisition system with a high-fidelity finite element (FE) prediction method. The system design features a novel “double-ring” sensor interface specifically engineered to decouple axial distraction forces from parasitic bending moments generated by asymmetric muscle tension. Furthermore, a patient-specific FE model utilizing the Ogden hyperelastic constitutive law was derived, explicitly based on the patient’s muscle volume from preoperative CT imaging, to predict the non-linear force evolution. The feasibility and accuracy of the system were validated in a pilot in vivo study using a single ovine model ($N=1$). To isolate the soft tissue resistance from callus formation, distraction was performed immediately postoperatively up to a total length of 4 cm. Experimental results demonstrated the system’s high linearity ($R^2>0.999$) and its ability to capture the characteristic viscoelastic relaxation of living tissues. The FE model successfully predicted the peak distraction forces, showing improved agreement with experimental data at larger distraction magnitudes. By integrating mechanical sensing with predictive modeling, this framework lays the foundation for future closed-loop, patient-specific control in distraction osteogenesis. Full article

Accurate and stable acquisition of the spatial distribution of soybean planting areas is essential for supporting precision agricultural monitoring and ensuring food security. However, crop remote-sensing mapping for specific regions still faces critical data bottlenecks: high-precision, large-scale pixel-level annotation is costly, resulting in scarce available labeled samples that make it difficult to construct large-scale training datasets. Although parameter-intensive models such as FCN and SegNet can achieve sufficient end-to-end training on large-scale public remote sensing datasets like LoveDA, when directly applied to the data-limited dataset in this study area, the models are prone to overfitting, leading to a significant decline in generalization ability. To address these issues, this study proposes a lightweight U-shaped semantic segmentation model, SimSDRU-Net. The model utilizes a pre-trained VGG-16 backbone to extract shallow texture and deep semantic features. The pre-trained weights mitigate the impact of overfitting in data-limited settings. In the decoding stage, a parameter-free lightweight SimAM attention module enhances effective soybean features and suppresses soil background redundancy, while an embedded S-DRU unit fuses multi-scale features for deep complementary reconstruction to improve edge detail capture. A label dataset was constructed using Sentinel-2 images as the data source and Menard County (USA) as the study area. The USDA CDL was used as a foundation for the dataset, with Google high-resolution images serving as visual interpretation aids. In the context of the experiment, Deeplabv3+ and U-Net++ were compared with U-Net under identical conditions. The results demonstrated that SimSDRU-Net exhibited optimal performance, with MIoU of 89.03%, MPA of 93.81%, and OA of 95.96%. Specifically, SimSDRU-Net uses the SimAM attention module to generate spatial attention weights by analyzing feature statistical differences through an energy function, so as to adaptively enhance soybean texture features. Meanwhile, the S-DRU unit groups, dynamically weights, and cross-branch reconstructs multi-scale convolutional features to preserve fine boundary details and achieve accurate segmentation of soybean plots. The present study demonstrates that SimSDRU-Net integrates lightweight design and high precision in data-limited scenarios, thereby providing effective technical support for the rapid extraction of soybean planting areas in North America. Full article

Magnetic resonance imaging (MRI), a widely adopted modality in clinical practice, enables the acquisition of multi-contrast images from the same anatomical structure, commonly referred to as multimodal images. Integrating these diverse modalities is crucial for enhancing model performance across a variety of medical image analysis tasks. However, in real-world clinical scenarios, it is often impractical to acquire all MRI modalities simultaneously due to factors such as patient discomfort, time constraints, and scanning costs. As a result, synthesizing missing modalities from available ones has emerged as an effective solution. To address these challenges, we propose HMF-MambaINR, a hierarchical multi-scale feature fusion network for cross-modality MRI synthesis. The model integrates Mamba-based Selective State Space Modeling (SSM) and implicit neural representation (INR) to capture long-range dependencies and enable continuous spatial reconstruction. A Multi-Feature Extraction Block (MFEB) captures local and global representations via multi-scale receptive fields, while a Modulation Fusion Module (MFM) adaptively fuses multi-modal features with dynamic weighting. Extensive experiments show that HMF-MambaINR surpasses state-of-the-art CNN-, Transformer-, and Mamba-based methods in synthesizing missing MRI modalities. Notably, the synthesized MRI images received positive feedback from radiologists in terms of image quality, contrast, and structural contour accuracy, highlighting the potential of the proposed method as a practical tool for clinical applications. Full article

Background/Objectives: Assessing structural damage in pediatric sacroiliitis is challenging, necessitating radiation-free alternatives to computed tomography (CT). This study evaluated the diagnostic performance of advanced MRI sequences—3D-MENSA (Multi-Echo in Steady-State Acquisition), 3D-MERGE (Multiple-Echo Recombined Gradient Echo), and Zero Echo Time (ZTE)—against conventional T1-weighted sequences for detecting structural lesions. Low-dose computed tomography (LDCT) served as the reference standard. A secondary objective was to qualitatively assess the visibility of active inflammatory lesions and fat metaplasia. Methods: In this cross-sectional study, 23 pediatric patients with enthesitis-related arthritis (ERA) were included. To adhere strictly to radiation safety principles, the study used pre-existing ldCT datasets from a clinical cohort as the reference standard. No new CT scans were performed for this study. Structural lesions (erosions, sclerosis, and joint-space changes) were independently scored by two blinded radiologists. Interobserver agreement was assessed using intraclass correlation coefficients (ICC). Results: Advanced sequences (ZTE, 3D-MENSA, 3D-MERGE) demonstrated high agreement with ldCT for erosion detection (ICC range: 0.924–0.998) and significantly outperformed conventional T1-weighted MRI (ICC: 0.707). 3D-MENSA provided distinct contrast, effectively differentiating the ligamentous component of the sacroiliac joint from both the synovial component and the adjacent bone cortex. Qualitatively, 3D-MENSA also identified bone marrow edema and fat metaplasia, which cannot be visualized by ZTE or ldCT. Conclusions: 3D-MENSA and 3D-MERGE enable comprehensive evaluation of structural sacroiliitis lesions in pediatric patients with diagnostic accuracy comparable to ldCT. Specifically, 3D-MENSA demonstrates the potential to detect both active and chronic lesions in a single, rapid, radiation-free acquisition. These findings suggest that it should be considered for routine pediatric imaging protocols. Full article

Reliable process monitoring and quality evaluation for resistance spot welding (RSW) have become more important now than ever. An ultrasonic probe embedded into welding electrodes has enabled the acquisition of data about molten pool formation throughout welding, but automation of high-performance ultrasonic data analyses is still necessary to fully realize a monitoring system. This work proposes a two-stage deep learning (DL) approach for automated ultrasonic data analysis for RSW processing monitoring. The first stage conducts semantic segmentation on ultrasonic M-scan welding process signatures, yielding masks for identified molten pool and stack regions from which weld penetration measurements can be directly extracted, as well as expulsion occurrences throughout welding. From input images and segmentation outputs, the second stage directly estimates resultant weld nugget diameters using an additional neural network. Both stages leveraged architectures based on TransUNet, mixing elements of both convolutional neural networks (CNN) and vision transformers, and the effect of cross-attention for stack-up sheet thickness data fusion was investigated via an ablation study. Additionally, in the diameter estimation stage, the ablation study included alternative feature extraction architectures in the network and investigated the provision of M-scans to the model alongside segmentation masks. In both cases, cross-attention was determined to improve performance, and in the case of diameter estimation, providing M-scans as input was found to be beneficial in general. With cross-attention, the segmentation approach yielded a mean intersection over union (IoU) of 0.942 on molten pool, stack, and expulsion regions in the M-scans with 13.4 ms inference time. With cross-attention, diameter estimates yielded a mean absolute error of 0.432 mm with 4.3 ms inference time, representing a significant improvement over algorithmic approaches based on ultrasonic time of flight. Additionally, the approach attained >90% probability of detection (POD) at 0.830 mm below the acceptable diameter threshold and <10% probability of false alarm (PFA) at 0.828 mm above the threshold. These results demonstrate a novel production-ready application of DL in ultrasonic nondestructive evaluation (NDE) and pave the way for zero-defect RSW manufacturing. Full article

Synthetic aperture radar (SAR), which appeared in the early 1990s, refers to a technology that creates a virtual large aperture by receiving/combining signals from various locations while moving with a fixed antenna. Using SAR-based image acquisition technology, a reconnaissance satellite can obtain high-quality images regardless of the weather and day/night conditions. In this study, the qualification tests of a space-grade 5m-class deployable reflector antenna for satellites, which is the primary payload of a SAR-based satellite, were conducted. In order to ensure the electrical performance of the reflector antenna, an alignment verification test was performed using a laser tracker system during the assembly and integration process. Generally, the satellite experiences a considerable amount of structural load under the launch condition and is exposed to extremely low- and high-temperature thermal environments under the orbital condition. For the space mission, environmental tests should be conducted to verify the structural/thermal stability for the launch and orbital conditions. A deployment repeatability test was conducted to ensure that the deployment mechanism operated properly before/after each test. The qualification process and philosophy proposed in this work could be applied to the development of the space-grade deployable reflector antenna. Full article

Crohn’s disease (CD) is a complex inflammatory bowel disease, defined by chronic transmural inflammation and marked heterogeneity in both anatomical distribution and disease behavior, with potential involvement of any segment of the gastrointestinal tract and multiple phenotypes. Advanced cross-sectional imaging nowadays plays a central role in CD management, reliably assessing both luminal and extraluminal inflammatory manifestations, supporting initial diagnosis, phenotypic characterization, and longitudinal monitoring of disease activity, complications and treatment response. Over the last two decades, Intestinal Ultrasound (IUS), MR Enterography (MRE), and Computed Tomography Enterography (CTE) have become central components of the diagnostic pathway. MRE has emerged as the most comprehensive, radiation-free modality for evaluating intestinal extent, inflammatory activity, and complications in Crohn’s disease. Multiparametric MRE, combining T2-weighted imaging, contrast-enhanced sequences, diffusion-weighted imaging, and cine acquisitions, enables a real “Crohn’s disease staging”, namely a thorough evaluation of the transmural inflammation, of fibrotic and fistulizing lesions in the small and large bowel, as well as in the perianal region. IUS provides a dynamic, widely accessible, safe and repeatable imaging technique that is particularly well suited for tight-monitoring strategies, early assessment of therapeutic response, and routine follow-up, especially in experienced centers. Notably CTE, despite concerns related to cumulative ionizing radiation exposure, remains indispensable in acute clinical settings owing to its rapid acquisition, broad availability, and high diagnostic accuracy for detecting abscesses, perforation, and bowel obstruction. Combined, these three modalities offer a complementary and patient-tailored framework for optimal CD management. This review outlines the pathological complexity of Crohn’s disease, traces the evolution of imaging approaches, and provides a comparative overview highlighting the specific strengths and limitations of each modality. Full article

The integration of outdoor camera images with three-dimensional (3D) geographic information on the observed scene is of interest for many video acquisition applications. To solve this data fusion problem, camera images have to be matched with the 3D geometry provided by a geographic information system (GIS). Considering a camera with a known geographical position, this paper proposes the use of a dense local azimuth–elevation map (LAEM) derived from a gridded digital elevation model (DEM) to represent the data and thus facilitate the matching of GIS and image data. To each regularly sampled azimuth and elevation angle pair, this map assigns the geographic point derived from the DEM viewed in this direction. The problem of computing the LAEM from the DEM is closely related to that of surface rendering, for which solutions exist in computer graphics. However, rendering software cannot be used directly in this case, since their view directions are constrained by the pinhole camera model and the apparent colour, rather than the position of the viewed point, is assigned to the viewing direction. Therefore, this paper also proposes a specific algorithm for the computation of the LAEM from the DEM. A MATLAB^® implementation of the algorithm is also provided, which is tailored to process the DEM dataset swissALTI3D from the Swiss Federal Office of Topography swisstopo. Full article

Background/Objectives: Virtual surgical planning (VSP) and CAD/CAM technologies have revolutionized complex maxillofacial reconstruction. While high-resolution imaging is critical for these workflows, the specific clinical impact of photon-counting computed tomography (PCCT) remains to be fully established. This prospective pilot study evaluates the feasibility and clinical utility of integrating PCCT into the preoperative planning and surgical workflow of complex maxillofacial reconstructive cases. Methods: This feasibility study included ten patients requiring complex maxillofacial reconstruction with microvascular free flaps. All underwent preoperative imaging with photon-counting CT. Primary endpoints included clinical assessment of osseous invasion, reliability of donor-site vascular mapping from a single acquisition, and compatibility of PCCT datasets with VSP/CAD-CAM platforms. Secondary endpoints included resection margin status, flap survival, and short-term oncologic outcomes. Results: PCCT provided high-resolution visualization of cortical and medullary bone, enabling detailed assessment of tumor-related osseous involvement. In selected cases, findings supported refinement of resection planning when prior imaging had been inconclusive. Spectral reconstructions reduced metal artifacts and facilitated precise segmentation for multi-segment osteotomies. Donor-site vascular anatomy was successfully evaluated within the same scan, supporting operative planning without additional imaging. PCCT datasets were fully compatible with the virtual surgical planning (VSP) software used in this study (CMX Portal, version 2.6.1158, Medartis AG, Basel, Switzerland; or ProPlan CMF, version 5.7.8.025, Materialise NV, Leuven, Belgium) in all cases (100%). Reconstruction was completed successfully in all patients, with 100% flap survival and R0 margins in all malignant cases. No technical failures occurred during imaging transfer or CAD/CAM fabrication. Conclusions: The integration of PCCT into the surgical workflow proved technically feasible and clinically impactful. This pilot data supports its potential to enhance surgical precision and preoperative planning in complex jaw reconstruction. Full article

With the rapid advancement of deep learning, data acquisition remains a persistent challenge, as model effectiveness heavily relies on the quality and quantity of training data. To address the difficulties of time-consuming and labor-intensive data collection, data augmentation techniques are commonly adopted. In this study, the proposed WACGAN-GP, a Generative Adversarial Network (GAN) architecture, serves as an effective data augmentation tool designed to augment training datasets and bolster model performance. This method integrates the advantages of the Auxiliary Classifier GAN and the Wasserstein GAN with gradient penalty to generate diverse and realistic samples. Experiments were conducted on three image datasets—MNIST, CIFAR-10, and a ship classification dataset—under limited training data conditions. By incorporating WACGAN-GP generated synthetic samples into the original training sets, classification performance was evaluated in both balanced and imbalanced scenarios. The results demonstrate that the proposed GAN-based approach significantly improves recognition accuracy and outperforms conventional augmentation methods, such as horizontal and vertical flipping. Full article

Background and Objective: Positive surgical margins (PSMs) remain a major challenge during radical prostatectomy, particularly in patients with high-risk prostate cancer (HR-PCa), where extracapsular extension, multifocal disease, and aggressive tumor biology substantially increase the likelihood of incomplete resection. In this setting, PSMs are strongly associated with early biochemical recurrence and frequently prompt adjuvant or salvage treatments, potentially exposing patients to overtreatment and added morbidity. Materials and Methods: To review and critically appraise established and emerging intraoperative technologies for surgical margin assessment during radical prostatectomy, with a specific focus on their potential role and relevance in patients with HR-PCa. Evidence Acquisition: A non-systematic literature review was performed using Pubmed, MEDLINE, Web of Science, and Google Scholar, focusing on preoperative, intraoperative ex vivo, and intraoperative in vivo technologies for margin assessment. Emphasis was placed on techniques with potential applicability to HR-PCa, where real-time intraoperative decision-making is particularly consequential. Evidence Synthesis: Preoperative tools, including multiparametric MRI, PSMA-PET imaging, and predictive nomograms, aid surgical planning but show limited sensitivity for microscopic extracapsular extension, especially in high-risk disease. Intraoperative frozen section analysis reduces positive surgical margin rates while enabling selective nerve-sparing (defined as a side-specific, risk-adapted preservation strategy); however, its widespread adoption is constrained by substantial logistical and resource requirements, and robust oncological outcome data in high-risk populations remain limited. Novel ex vivo approaches, such as fluorescence confocal microscopy and specimen-based PSMA PET/CT imaging, offer rapid whole-gland or targeted margin assessment with reduced dependency on dedicated pathology workflows. In parallel, emerging in vivo technologies, particularly PSMA-targeted near-infrared-fluorescence-guided surgery, enable real-time detection of residual tumor and facilitate selective re-resection, representing a biology-driven approach that may be especially suited to HR-PCa. Conclusions: In high-risk prostate cancer, intraoperative margin assessment technologies may extend beyond functional preservation and play a central role in optimizing oncological radicality and multimodal treatment sequencing. While NeuroSAFE remains the reference standard, PSMA-based ex vivo and in vivo technologies are particularly promising in HR-PCa due to their ability to integrate tumor biology into surgical decision-making. Prospective studies focusing on high-risk-specific oncological and patient-reported outcomes are needed before widespread clinical implementation. Full article

Point-of-care ultrasound (POCUS) is portable and radiation-free, but its clinical reliability is constrained by operator-dependent image acquisition and the limited scalability of expert quality assurance (QA) review. As handheld devices proliferate faster than mentorship capacity, trainees increasingly rely on heterogeneous free open access medical education (FOAMed) resources that rarely provide real-time psychomotor feedback. We conducted a structured narrative review (MEDLINE, Embase, Scopus, and Web of Science; last searched on 23 February 2026), with searches performed by H.H. and independently checked by J.J.P. (both POCUS-trained clinicians). After screening, 31 studies were included. We synthesized evidence on artificial intelligence (AI) systems that support bedside image acquisition and automate QA. The primary synthesis centered on key prospective or comparative clinical evaluations of AI-guided acquisition across echocardiography, focused assessment with sonography in trauma, abdominal aortic aneurysm screening, and lung ultrasound, complemented by peer-reviewed studies of FOAMed appraisal tools and online resource quality. These evaluations suggest that real-time probe guidance, view recognition, anatomy labeling, and automated capture may enable novices, after brief training, to acquire diagnostically adequate images for narrowly defined tasks. Early reports of automated QA scoring and program-level triage for expert review suggest potential to reduce expert workload and shorten feedback cycles, but external validation, generalizability across devices and patient habitus, and patient-centered outcomes remain limited. Acquisition-focused AI may therefore serve as an upstream “digital mentor” to improve novice image acquisition. We propose a practical pathway that integrates curated FOAMed resources and simulation with AI-guided bedside acquisition and continuous QA governance for safe deployment. Full article

Although active learning can reduce the effort required to annotate object detection data, many current methods rely on a single selection criterion or combine criteria without accounting for annotation costs or their interactions. This paper presents a multi-criterion, cost-aware active learning framework for detecting small objects in agricultural images. The framework jointly considers prediction uncertainty, object size, scene density, and annotation cost. We evaluate both scalarized and Pareto-based selection strategies across five cost models and conduct an ablation study to examine the role and interactions of each criterion. Experimental results demonstrate that explicit annotation cost modeling improves active learning efficiency by reducing the amount of annotation required to achieve a given level of detection performance. Across multiple cost formulations and selection strategies, cost-aware acquisition reaches comparable accuracy and reduces the estimated annotation effort required to reach comparable detection performance by up to 50% compared to random sampling, where annotation effort is approximated using prediction-derived cost proxies. Full article