Search Results (40,278)

This exploratory mixed-methods research study examined 27 Western YouTubers who are pro-China with over 644 million views and 4.7 million subscribers in 2024. We analyzed their top 10 videos N = 270) with human coding using both qualitative and quantitative content analysis and most-liked comments on these videos using human–machine collaboration with generative AI and audience metrics, showing how influencer marketing techniques can be used to engage audiences and the potential public diplomacy benefits for China on global social media. This study identified the influencer appeals and content attributes in the videos that YouTubers used to promote China’s image and refute the criticisms of China to their Western peers. We found that most videos use Chinese in addition to English, and almost all videos related to China feature at least one China accomplishment icon. Their audiences are mostly a mix of Chinese and Western audiences. These YouTubers’ most common appeals are authenticity and homophily. They present a very friendly character and share their daily experiences in China. Content comparing China with the West or refutation of Western criticism was not found more effective in engaging audiences than videos not using these approaches. By speaking up for China, these YouTubers and their followers form a pro-China community on YouTube, offering perspectives different from what they received from Western mainstream media to global audiences. Full article

Predicting progression from mild cognitive impairment (MCI) to Alzheimer’s disease (AD) requires models that represent both regional brain abnormalities and their evolution across repeated examinations. We developed a longitudinal graph neural network that integrates structural magnetic resonance imaging, FDG-PET, regional imaging biomarkers, and clinical covariates across irregular follow-up visits. The study included 614 participants with baseline MCI from the Alzheimer’s Disease Neuroimaging Initiative: 218 converters to AD within five years and 396 non-converters, with 2438 eligible longitudinal visits. Each visit was represented as an 82-node brain graph based on the Desikan–Killiany atlas. Node features combined a 128-dimensional multimodal convolutional embedding with four regional biomarkers. Graph-attention layers modelled spatial dependencies, a node-wise gated recurrent unit modelled longitudinal dependencies, and masked temporal self-attention aggregated variable-length visit sequences. Participants were divided at the subject level into development and held-out test sets, and hyperparameters were selected by five-fold cross-validation within the development set. On the held-out test set of 123 participants, the model reached an area under the receiver operating characteristic curve of 0.859 (95% CI 0.795–0.915), balanced accuracy of 0.805 (95% CI 0.736–0.862), sensitivity of 0.781, and specificity of 0.832. The AUC was numerically higher than that of the strongest baseline, a CNN–GRU sequence model, which reached 0.832 (95% CI 0.758–0.894); the paired AUC difference was 0.027 (95% CI 0.009–0.098), the unadjusted DeLong p-value was 0.026, and the Holm-adjusted p-value was 0.052, which was not significant at the conventional 0.05 threshold after correction for multiple comparisons. In ablation experiments, removing temporal modelling reduced the AUC to 0.818, and removing the spatial graph structure reduced it to 0.808, the two largest reductions observed. Integrated-gradient analysis placed the highest importance on hippocampal and entorhinal regions. Combining graph-based spatial modelling with recurrent longitudinal reasoning was associated with higher discrimination than sequence modelling alone, though this difference was not statistically significant after correction for multiple comparisons. Validation was restricted to a single research cohort (ADNI), and no independent external dataset was used; prospective external validation on an independent cohort is required before the model can be considered for clinical use Because FDG-PET was unavailable for 19.3% of visits, we report the headline result separately from a sensitivity analysis restricted to participants with complete FDG-PET at every visit (development set cross-validated AUC 0.891 vs. 0.874 for the full cohort with masked missing FDG-PET); multimodal performance should be read as cohort-dependent rather than as a single unconditional figure. Full article

Accurate road graph extraction from satellite imagery is essential for large-scale mapping and geospatial analysis. Recent one-shot graph extraction frameworks based on foundation models have achieved promising performance, but their effectiveness decreases in complex environments where road structures are partially obscured by vegetation, buildings, shadows, and other surface conditions. These occlusion-induced disturbances lead to incomplete connectivity and degraded topology reconstruction, particularly under out-of-domain scenarios. This study proposes an occlusion-aware refinement framework to improve the robustness of satellite image road graph extraction while maintaining the original backbone architecture. The proposed framework introduces three complementary strategies: Synthetic Occlusion Augmentation for explicit occlusion-aware representation learning, an Occlusion-Adaptive Extended-Line strategy with Hard-Mining Topology Optimization for improved connectivity reasoning, and an Occlusion-Adaptive Node-Guided Resampling mechanism for reliable graph node localization. Experiments conducted on the Global-Scale road graph extraction benchmark demonstrate that the proposed method consistently improves topology reconstruction performance. Compared with the reproduced SAM-Road++ baseline, the proposed framework improves TOPO F1 from 61.81 to 62.56 on the in-domain split and from 46.93 to 51.51 on the out-of-domain split. Furthermore, the ID-OOD performance gap is reduced from 14.88 to 11.05, indicating enhanced robustness under unseen geographic conditions. The results demonstrate that explicitly modeling occlusion as a structured factor can effectively improve the generalization capability of satellite road graph extraction systems. Full article

Multimodal studies of prostate cancer typically report aggregate fusion gains from histology and transcriptomics but rarely characterize when each modality is informative. We asked whether morphology and transcriptomics contribute differently to distinct clinical endpoints, using Gleason grading and nodal status prediction. On 401 TCGA-PRAD patients with matched whole-slide images, bulk RNA sequencing, and clinical data, we evaluated 45 morphological configurations, seven RNA configurations, and six fusion strategies across four grading formulations, T-stage, and N-stage. To avoid per-task model selection, morphology used a single fixed configuration (gated ABMIL on UNI features). Contribution estimates used repeated stratified five-fold cross-validation, with a locked conformal-prediction protocol and gene set enrichment analysis. The two endpoints showed opposite modality dependence. For Gleason grade, the fixed morphological model exceeded the best RNA configuration across all 45 configurations (five-class macro-F1: 0.445 versus 0.398). For nodal status, no morphological configuration exceeded AUROC 0.62, whereas transcriptomics reached 0.68 (permutation p = 0.037); a direct interaction test confirmed the reversal (bootstrap 95% CI excluding zero). Fusion gains were modest and task-dependent; enrichment analysis linked the nodal signal to loss of smooth muscle programs, consistent with established dedifferentiation biology. The informative modality is task-dependent: morphology predicts grade, transcriptomics predicts nodal status. Full article

Developing sustainable paradigms for nanomaterial synthesis is essential to reduce the environmental impact of conventional chemical routes. Herein, we report a green, phytochemically mediated synthesis of MnFe₂O₄NPs using aqueous grapefruit (Citrus paradisi) peel extracts. The bioflavonoid and polyphenolic constituents functioned as dual-functional biogenic reducing and stabilizing agents. FTIR analysis verified spinel lattice metal–oxygen bonds and surface-bound biogenic groups, while SEM imaging revealed the surface morphology and structural aggregation of the nanoparticles. Once integrated into an electroanalytical platform (MnFe₂O₄NPs/GCE), the modified interface exhibited robust electrocatalytic activity toward sulfite oxidation in an acetate buffer (pH = 4.0). Mixed-valence dynamics (Mn²⁺/Mn³⁺ and Fe²⁺/Fe³⁺) significantly amplified anodic currents and lowered the kinetic overpotential. Under optimized chronoamperometric conditions, the sensor demonstrated a wide linear range (1–100 μM, r = 0.999), a low detection limit (LOD = 0.045 μM), a limit of quantification (LOQ = 0.136 μM) and excellent selectivity against co-existing ionic and organic interferents. This study provides a simple, eco-friendly, and technically robust strategy for fabricating high-performance electrochemical platforms for food and environmental monitoring. Full article

Near-duplicate diagram detection in electronic documents is challenging because diagram identity depends on graphical structure, spatial composition, and textual labels, while reused images may undergo compression, cropping, rotation, photometric changes, or perspective distortion. This study proposes a cascaded multimodal framework combining perceptual hashing, Siamese Residual Network with 18 layers (Siamese ResNet18), Distillation with No Labels, ver. 2 (DINOv2)visual representations, and a text-similarity classifier. A controlled benchmark was constructed from Artificial Intelligence 2D Diagram Dataset (AI2D) using Light, Medium, and Hard transformations, with source-grouped splitting by base_id to prevent leakage across training, validation, and test sets. Perceptual hashing achieved Area Under the Receiver Operating Characteristic Curve (ROC-AUC) = 0.7483, while Siamese ResNet18 increased ROC-AUC to 0.8599. DINOv2 provided the strongest visual performance, achieving Accuracy = 0.9933, F1-score = 0.9933, ROC-AUC = 0.9992, and Average Precision = 0.9994; F1-score remained 0.9901 for Hard transformations. Visual fusion increased ROC-AUC to 0.9995, and full multimodal fusion reached ROC-AUC = 0.9998. At an early-exit threshold of 0.95, 27.6% of pairs were resolved at the hashing level. These results support the coarse-to-fine design on the constructed AI2D-derived benchmark. A targeted hard-negative stress test revealed substantially higher false-positive rates under deliberately matched spatial layouts, with an overall False Positive Rate (FPR) of 0.48 for DINOv2 and 0.16 for full multimodal fusion. Generalization to naturally reused or redrawn diagrams, larger and more diverse hard-negative collections, and Optical Character Recognition (OCR)-derived text remains to be evaluated. Full article

This study utilizes Sentinel imaging to monitor small farming plots (<5 ha) and objectively estimate their significance in agricultural output. It develops, tests, and evaluates methodologies within and across ten southern European NUTS-3 regions—commonly referred to as prefectures, provinces, or departments—located in Greece, Italy, France, Portugal, and Spain. The study provides stakeholders with the necessary tools to estimate their spatial distribution, crop diversity, crop area extent, and agricultural yields. The approach is expanded as needed to similar landscapes, serving as a model for other European NUTS-3 regions and providing a comprehensive view of the remote sensing solution. It employs random forest crop identification and classification analysis to build crop-type maps and record the countless small agricultural plots. The Sentinel data utilized are from 20 NUTS-3 pilot regions selected across 11 EU countries, focusing on 29 estimated key crop areas. The research demonstrates remarkable classification accuracy for a previously identified wide range of key crop data across each of the selected NUTS-3 regions. Crop area estimates derived from field data were adjusted and stratified. The RS-based methodology reduces propagation errors in estimating the area and production of crop categories with low accuracy levels, enhancing the overall understanding of the crucial role of small farming plots. For highly accurate key crop products (per crop-type categorization, with $FScore$ > 75%), production estimates are calculated by multiplying predicted self-reported crop yields by unbiased key crop small farming plot area estimates. However, for the referenced period of cultivation, only 16 out of the above 29 estimated key crop areas are recorded for SFs by the regional official statistics. In addition, of the 16 reported key crop areas listed above, eleven are from nine NUTS-3 regions spread across four southern EU countries, with the remaining five coming from four NUTS-3 regions of two non-southern EU countries. The analysis shows a strong correlation between estimated key crop areas from SFs reported in regional official statistics and Sentinel-based key crop area estimates obtained from small farming plots, with $R^2=0.96$ for 16 estimated key crop area datasets across all NUTS-3 regions in the EU and $R^2=0.98$ for 11 estimated key crop area datasets across all southern NUTS-3 regions. As a result, in NUTS-3 regions lacking regional official statistics data, RS can serve as a reliable alternative source for estimating the extent of key crop areas. Full article

Multi-view optical reconstruction supports spacecraft inspection, target characterization, and analysis from viewpoints not observed during image acquisition. For time-sensitive space situational awareness, such reconstruction must balance fidelity with the time required to build and render the scene representation. Three-dimensional Gaussian splatting (3DGS) provides efficient novel-view rendering, but its general-purpose initialization and densification do not account for the strong spatial imbalance of spacecraft imagery, where a compact and structurally complex target is surrounded by a largely uninformative background. This mismatch can allocate computation to weakly supported regions, limit the reconstruction of thin structures and object boundaries, and enlarge the representation without a corresponding gain in fidelity. We propose SpaceFast-GS, a foreground-guided framework that introduces RGB-derived target evidence at three successive stages. Foreground-guided allocation (FGA) constructs relaxed multi-view support and places the initial Gaussian population directly in image-supported target regions, reducing the corrective growth required after generic initialization. Object- and boundary-aware refinement (OBR) retains full-image photometric supervision while increasing the contribution of spacecraft regions and silhouette transitions. Guided population control (GPC) combines optimization gradients with support confidence, thin-support evidence, and projected reconstruction residuals to prioritize densification toward a target population and avoid unnecessary primitive growth. All target evidence is obtained from the training RGB images and calibrated cameras, without external segmentation or pretrained reconstruction. On NASA3D Standard-29, SpaceFast-GS ranks second across five reconstruction measures and achieves $41.357\pm 0.015$ dB full-image PSNR in 159.4 s at 932.1 FPS. Relative to standard 3DGS, this represents a 32.3% reduction in training time and a 3.41-fold increase in rendering throughput; stage-wise ablations further verify the contribution of FGA, OBR, and GPC. These results show that coordinating Gaussian allocation, refinement, and population growth with foreground evidence can improve reconstruction turnaround and rendering efficiency while retaining competitive spacecraft reconstruction quality. Full article

Background/Objectives: Dental biofilms are complex microbial communities that could lead to the formation of dental caries and other oral diseases. This in vitro study aimed to evaluate the effects of two commercially available toothpastes containing bioactive glass (BioMin F and Sensodyne Repair & Protection) on biofilm removal from five contemporary restorative dental materials and bovine enamel. The objective of this study was to compare the effects of bioactive glass-containing toothpastes on Streptococcus mutans viability and colonization on different restorative materials and bovine enamel. Methods: A total of 42 disc specimens representing five restorative dental materials and bovine enamel were incubated with Streptococcus mutans for 24 h to develop biofilms. After bacterial adhesion, specimens were brushed with toothpaste slurries (1:2 dilution) and distilled water (control). Biofilm viability and bacterial colonization were evaluated using the MTT assay and colony-forming unit (CFU) counts, respectively. Biofilm morphology on selected specimens was examined by scanning electron microscopy (SEM). Differences among groups were analyzed statistically using Tukey’s test. Results: Statistically significant differences were observed among the study groups following brushing (p < 0.05). EQUIA Forte HT (EF) demonstrated the lowest levels of S. mutans viability and colonization, whereas ACTIVA PRONTO (AP) exhibited the highest levels. Among the toothbrushing methods, BioMin F resulted in a greater reduction in bacterial viability and colony counts independent of the restorative materials. Combined analysis indicated that EF brushed with Sensodyne Repair & Protect yielded the lowest bacterial burden, whereas AP brushed with distilled water exhibited the highest. The results describing bacterial viability and colony counts also correlate with the SEM images taken. Conclusions: Within the limitations of this in vitro study, Sensodyne Repair & Protect was associated with lower S. mutans viability and colonization on most restorative materials, particularly EQUIA Forte HT. Full article

Explainable artificial intelligence (XAI) is increasingly used to improve the transparency of deep learning models for medical image-based diagnosis; however, the reliability and clinical value of the generated explanations remain uncertain. This systematic review followed PRISMA 2020 and examined peer-reviewed studies published from 2022 to 11 June 2026. Following the revised literature search, eligibility assessment, independent methodological appraisal by two reviewers, and consensus resolution of disagreements, 57 primary studies were included in the final evidence synthesis. Of these, 37/57 (64.9%) were published in 2025 or during the eligible portion of 2026, reflecting the recent expansion of research in this field. CNN/convolutional architectures were identified in 56/57 studies (98.2%), while 55/57 (96.5%) produced visual or attribution-based explanations without a textual reporting component; only 2/57 (3.5%) incorporated a textual component. Predictive performance was generally evaluated extensively; however, explanation-quality assessment was less consistently developed: 27/57 studies (47.4%) fully satisfied the review criterion for evaluating explanation quality, whereas 30/57 (52.6%) provided only partial evaluation, frequently relying on qualitative or otherwise incomplete assessment. Recurrent limitations included insufficiently diverse datasets, limited external and multicenter validation, scarce clinician involvement, and the absence of standardized XAI evaluation frameworks. Overall, explanation generation has advanced faster than explanation validation, and clinically dependable XAI requires architecture-aware assessment, external validation, and prospective human-centered evaluation. Full article

Robust traffic sign detection is essential for reliable autonomous driving systems; however, detection performance can be substantially affected by adverse environmental conditions. Although simulation-based approaches are widely used to evaluate robustness, the extent to which simulated weather reproduces real-world environmental effects remains insufficiently understood. This study presents a controlled, comparative evaluation methodology for assessing a YOLOv8-based traffic sign detection model across four environmental conditions: clear (sunny), simulated rain, real-world rain, and simulated snow. The same road segment, camera configuration, and predefined set of 55 traffic-sign instances were maintained across the evaluated conditions, enabling interpretable comparisons while minimizing scene-level variability. Detection performance was assessed using representative detection confidence (RDC) at the traffic-sign-instance level, along with image-quality metrics such as sharpness, saturation, and intensity. Mean RDC was highest under clear conditions (0.825), followed descriptively by simulated snow (0.647), real-world rain (0.408), and simulated rain (0.395). However, simulated and real-world rain did not differ significantly in RDC (Holm-adjusted p = 0.770), while the Friedman test indicated a significant overall difference among conditions (χ²(3) = 98.767, p < 0.001; Kendall’s W = 0.599). Image-quality analysis further revealed substantial differences between rainfall conditions in successfully detected traffic-sign regions, particularly in Sharpness and Mean Saturation. Overall, the findings demonstrate that simulated weather can produce traffic-sign-level detector responses that are statistically comparable to those observed under independently recorded real-world rainfall, while producing substantially different image-level characteristics. The results support the use of controlled weather simulation as a complementary evaluation approach, alongside real-world validation, to investigate the environmental robustness of camera-based traffic-sign detection systems. Full article

The use of artificial intelligence (AI) in endoscopic ultrasonography (EUS) is receiving increasing attention, particularly in the field of pancreatic diseases, where early and accurate diagnosis remains a major clinical challenge. This narrative review focuses on current and emerging applications of AI in pancreatic EUS, covering both image-based diagnostic support and the analysis of samples obtained through EUS-guided tissue acquisition. The first part of the review discusses how AI is being applied to improve EUS image interpretation, ranging from lesion detection to characterization and differentiation between benign and malignant pancreatic findings. The review also discusses early evidence and future perspectives for real-time procedural support, where diagnostic performance remains highly operator-dependent. The second part explores a less frequently discussed but equally relevant area: the use of AI in the analysis of cytological specimens obtained through EUS-guided fine-needle aspiration or fine-needle biopsy. Although cytopathology may appear to lie outside the traditional clinical scope of EUS, it represents an essential step in the diagnostic workflow of pancreatic diseases. Recent developments in AI-assisted digital cytology and pathology have shown promising potential to support and standardize cytological interpretation, with possible benefits in terms of diagnostic consistency, reproducibility, and turnaround time. By bridging imaging and pathology, AI may enhance the entire pancreatic EUS workflow, contributing to more efficient, accurate, and personalized diagnostic pathways in pancreatic disease management. Full article

Non-cooperative sensing using signals of opportunity traditionally requires an explicit reference signal for target detection and localization. This paper introduces a reference-free sensing framework in which target geometry is inferred directly from the received waveform rather than by comparison with an acquired or reconstructed illuminator signal. The proposed framework is implemented using the Ranging, Detection, Imaging, Communications, Approach, and Landing (RaDICAL) architecture, which combines a hybrid Dish–Sparse Uniform Circular Array (SUCA) receiver with Starlink downlink transmissions as spaceborne illuminators of opportunity. Deterministic Multifrequency Dither (DMD) applied across the SUCA elements transforms spatial diversity into unique composite waveform signatures. A unified electromagnetic and signal-processing model is developed that combines spherical-wave propagation, parabolic focusing, deterministic multifrequency modulation, and QR-based waveform-domain hypothesis testing for direct target localization. Numerical simulations together with link-budget analysis demonstrate the feasibility of the proposed approach. Single-dwell detection of 0 dBsm targets is achieved at physical signal-to-noise ratios near 0 dB, while near-unity detection probability is obtained above 10 dB SNR under controlled false-alarm conditions. The results demonstrate that commercial Starlink LEO communication satellites can serve as practical illuminators of opportunity for reference-free non-cooperative sensing without requiring acquisition or reconstruction of the transmitted illuminator waveform. Full article

Architectural colour in public libraries significantly influences indoor atmospheric quality and occupant emotional well-being, yet palette selection in design practice remains predominantly reliant on subjective intuition, precedents, and vendor catalogues. This study develops a computational framework that links multimodal image colour extraction, circular clustering, human perceptual evaluation, and multi-criteria decision weighting into a structured analysis pipeline. Using a corpus of 500 curated library interior images across diverse geographic regions, a multimodal large language model extracted dominant architectural colour records and visual weights, which were deterministically converted to HSV space and clustered using weighted circular k-means. A semantic differential experiment involving fifty design students across five bipolar scales provided empirical perceptual ratings, evaluated through repeated nested ten-fold cross-validation across five machine learning algorithms. The analysis identified five recurrent colour paradigms, demonstrating that spatial hue distributions form distinct chromatic clusters across modern library architecture. Perceptual models revealed that average colour features reliably predict perceived warmth and quietness, whereas modernity and naturalness depend more heavily on non-chromatic spatial cues. By integrating these predictive models with analytic hierarchy process and Delphi expert weights, the framework establishes a transparent decision-support protocol for early schematic design. This allows architects and clients to quantitatively compare alternative colour schemes against functional zoning requirements, providing an objective, accountable foundation for evidence-based interior design. Full article

Electroencephalography (EEG) source localization has evolved from a technical solution to an inverse problem and into a fundamental framework for investigating large-scale brain networks. By reconstructing cortical activity from scalp-recorded signals, source-space EEG improves anatomical interpretability and reduces confounding effects of volume conduction compared with sensor-level analyses. These advances have enabled more reliable estimation of functional and effective connectivity, graph-theoretical characterization of brain networks, and the integration of machine learning for automated analysis and clinical decision support. Nevertheless, the accuracy and reproducibility of source-level analyses remain highly dependent on forward modeling, inverse algorithms, preprocessing strategies, and connectivity estimation methods. Recent methodological developments have further expanded the role of EEG source imaging in computational neuroscience, supporting applications ranging from network-based investigations of brain function to biomarker discovery and disease classification. This review summarizes the computational framework underlying source-level EEG, beginning with the forward and inverse problems and progressing through source reconstruction, connectivity analysis, graph-theoretical network modeling, and emerging machine learning and deep learning techniques. We also highlight current clinical applications and major methodological challenges, including source leakage, model uncertainty, reproducibility, and interpretability, and outline future directions for robust and clinically translatable source-level EEG analysis. Full article