Search Results (16,831)

Background/Objectives: Breast cancer screening such as mammography supports earlier detection, but variability in interpretation can still lead to missed cancers and avoidable follow-up testing. We evaluated ChatGPT-5.2 Thinking (OpenAI) as a stand-alone model for examination-level malignancy classification on standard bilateral mammography views in a biopsy-referred cohort, compared with breast radiologists, and assessed laterality performance. Methods: We conducted a retrospective, multicenter diagnostic-accuracy study across breast imaging centers in Saudi Arabia. From an upstream screened cohort (n = 1225), we constructed a biopsy-referred test set of 100 mammography examinations (four 2D views per exam: bilateral CC and MLO; 400 images), including 61 biopsy-confirmed malignancies and 39 biopsy-negative controls, with pathology as the reference standard. Radiologists were blinded to pathology and AI outputs and assigned BI-RADS (0–5) and suspected laterality. ChatGPT-5.2 interpreted the same de-identified views using a BI-RADS-guided prompt to generate BI-RADS and laterality. The sensitivity, specificity, accuracy, and laterality concordance were then estimated. Results: ChatGPT-5.2 had higher sensitivity than radiologists (95.08% vs. 81.97%) but markedly lower specificity (10.26% vs. 56.41%), resulting in lower overall accuracy (62.00% vs. 72.00%). The AI produced 58 true positives, 35 false positives, and 3 false negatives, while radiologists produced 50 true positives, 17 false positives, and 11 false negatives. Laterality accuracy among malignant examinations was 60.66%. Conclusions: In this pathology-anchored, biopsy-referred evaluation, ChatGPT-5.2 identified more cancers but generated substantially more false-positive classifications and showed only moderate breast-side localization. These findings support use as a concurrent aid or prioritization tool rather than a stand-alone reader and motivate efforts to improve specificity and laterality before prospective validation. Full article

Objectives: Doxorubicin, a widely used chemotherapeutic agent, is known to induce hepatotoxicity through oxidative stress and inflammatory pathways. Vitamin D has been reported to exert antioxidant and immunomodulatory effects; however, its potential protective role in doxorubicin-induced liver injury remains insufficiently characterized. Materials and Methods: Adult male Wistar albino rats were randomly assigned to six groups (n = 7): Control, Vitamin D (5000 IU/kg), Vitamin D (60,000 IU/kg), Doxorubicin, DOX + Vitamin D (5000 IU/kg), and DOX + Vitamin D (60,000 IU/kg). Vitamin D₃ (cholecalciferol) was administered orally either as a daily dose (5000 IU/kg for 12 days) or as a single bolus dose (60,000 IU/kg). Doxorubicin (6 mg/kg/day, cumulative dose 18 mg/kg) was administered intraperitoneally on days 10–12. Hepatic injury was evaluated using ⁹⁹mTc-pyrophosphate (⁹⁹mTc-PYP) scintigraphy, serum liver enzymes (AST, ALT, LDH, total bilirubin), renal markers (BUN, creatinine), calcium and 25-hydroxyvitamin D [25(OH)D], oxidative stress parameters (MDA, TOS, TAS, GSH, SOD, Nrf2), and inflammatory cytokines (TNF-α, IL-6, IL-1β, IL-10). Results: Doxorubicin markedly increased hepatic ⁹⁹mTc-PYP uptake and significantly elevated AST, ALT, LDH, bilirubin, MDA, TOS, TNF-α, IL-6, and IL-1β levels while reducing Nrf2, GSH, SOD, TAS, and IL-10 (all p < 0.001). Vitamin D supplementation significantly increased serum 25-hydroxyvitamin D [25(OH)D] levels compared with controls (32.3 ± 2.7 vs. 74.1 ± 3.8 and 69.3 ± 3.2 ng/mL for the 5000 and 60,000 IU/kg groups, respectively; p < 0.001) and attenuated DOX-induced hepatic injury, as indicated by reduced radiotracer uptake and improved oxidative and inflammatory markers. Vitamin D also mitigated DOX-associated increases in renal injury markers (BUN and creatinine) without inducing hypercalcemia. No significant differences were observed between the two vitamin D dosing regimens in most outcome measures. Conclusion: Vitamin D supplementation exerted protective effects against doxorubicin-induced liver injury, likely through modulation of oxidative stress and inflammatory pathways. Additionally, ⁹⁹mTc-PYP scintigraphy may serve as a useful imaging tool for detecting acute hepatocellular injury and evaluating therapeutic responses. Full article

We investigate 3D-printed composite materials composed of a photosensitive polylactic acid (PLA) resin blended with 10% starch and fabricated by Digital Light Processing. We synthesize the 3D-printed composites by incorporating a post-processing stage involving thermomoulding at various temperatures ranging from 50 °C to 150 °C. The composition, structure, and thermal and mechanical performance of the 3D-printed composites are evaluated using infrared spectroscopy (FTIR), Differential Scanning Calorimetry (DSC), synchrotron X-ray microtomography and tensile testing assisted with digital image correlation. Our results indicate that post-treatment influences the mechanical behaviour of the composites, leading to a moderate increase in stiffness while the tensile strength remains slightly reduced compared with the reference condition, particularly when moulding temperatures reach 100 °C. Our 3D printing approach combined with the photosensitive/starch blend provides a cost-effective alternative for obtaining 3D-printed biosourced components, maintaining technical performance at a reasonable cost. Full article

Introduction: Optical coherence tomography (OCT) is the gold standard for retinal nerve fibre layer (RNFL) assessment; its high cost and limited accessibility hinder widespread use. This study aims to develop deep learning models that generate RNFL thickness maps from fundus images, providing a cost-effective alternative to OCT. Methods: A dataset of 5000 fundus-OCT image pairs from 5000 unique glaucoma patients was used to train and compare the following four U-Net-based deep learning models: ResU-Net, R2U-Net, Nested U-Net, and Dense U-Net. All models were trained for up to 1000 epochs with early stopping (patience = 50 epochs). Performance was evaluated using Mean Squared Error (MSE), Mean Absolute Error (MAE), Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), and Fréchet Inception Distance (FID). Results: ResU-Net demonstrated the best performance, achieving MSE = 0.00061, MAE = 0.01877, SSIM = 0.9163, PSNR = 32.19 dB, and FID = 30.08. These results represent a 108% improvement in SSIM and a 67% improvement in PSNR compared to previously published benchmark for this task. Conclusions: This study demonstrates that deep learning models, particularly ResU-Net, can generate high-fidelity RNFL thickness maps from fundus photographs, substantially outperforming prior published benchmarks. This approach represents a potential contribution toward accessible glaucoma assessment, contingent upon prospective clinical validation and regulatory evaluation. Full article

Lung cancer is one of the most common causes of death from cancer around the world, and catching it early through computed tomography (CT) scans can drastically improve survival. However, automated classification of pulmonary nodule candidates is hard because images do not all have the same intensity across scanners and protocols, resulting in inconsistent performance, more false positives (FP), and a ceiling on how much deep learning models work in an average clinic. In this work, we tackle this by introducing a preprocessing step that corrects intensity differences before feeding images into classification models. We use Contrast-Limited Adaptive Histogram Equalization (CLAHE), but with its key parameters tuned automatically via a modified version of the Covariance Matrix Adaptation Evolution Strategy (CMA-ES). This helps to boost local contrast adaptively, keeps important anatomical details intact, and cuts down on noise. We tested the approach on the public LUNA16 dataset, first checking image quality (Peak Signal-to-Noise Ratio (PSNR) around 53 dB and Structural Similarity Index (SSIM) of 0.9, better than standard methods), then training three popular deep models—namely, ResNet-50, EfficientNet-B0, and InceptionV3—with CutMix augmentation for better generalization. On the enhanced images, ResNet-50 achieved up to 99.0% classification accuracy with substantially less FP than when using the raw scans. Taken together, these results demonstrate that intelligent and optimized preprocessing can effectively mitigate intensity variations via deep learning for lung nodule detection, thus coming closer to realizing the practical toolbox of computer-aided diagnosis in routine clinical practice. Full article

Drought, a major abiotic stressor affecting global agricultural productivity, significantly reduces crop yields and threatens food security worldwide. As the primary organ for perceiving soil moisture signals and absorbing water, the crop root system architecture plays a pivotal role in plant adaptation to drought conditions. With the development of high-throughput imaging technologies (i.e., 2D/3D image acquisition), high-throughput genotyping platforms, and gene-editing technologies, significant progress has been achieved in the characterization of root traits and the dissection of molecular genetic regulatory networks underlying these traits in crops. This review comprehensively synthesizes recent advances in the phenotypic characterization, underlying molecular regulatory networks, and functional roles of key root architectural traits, including the root length, angle, density, and root hair development, in enhancing drought resilience. Finally, we discuss the existing challenges in the current research and provide an outlook on the future trend of integrating multi-omics, high-throughput phenomics, and genome editing technologies to breed new drought-resistant crop varieties with ideal drought-resistant root architectures. Full article

Background: Auricular keloids and ear helix deformities are undesirable and aesthetically unpleasing deformities that can cause significant patient psychologic and self-esteem problems. Pressure therapy for keloids is well documented to be an effective non-invasive treatment modality. However, current devices lack comfort and aesthetic appeal to deliver the pressure forces required effectively and uniformly. This work aims to highlight some different pressure therapy approaches for the management of keloids and irregularities in the ear helix morphology. Methods: A case series of four patients presenting with auricle keloids of various sizes and at different locations secondary to ear piercing and one case of congenital helix deformity were treated successfully with pressure therapy devices. The device designs varied based on the keloids’ characteristics and patients’ preferences and involved wire-based spring-activated appliances resembling ear rings for moderate keloid lesions, modified double-spring systems for large or elongated lesions, and magnet-based devices. A pair of inert magnetic discs of different diameters was positioned on the anterior and posterior aspects of the keloid lesion. The magnets were then encapsulated in acrylic resin to improve retention and adaptation, and the external surface was masked with gold glitter to enhance aesthetics and patient acceptance. The helix-deformity case was treated following a complete digital workflow integration where the sound contralateral ear was digitally scanned, mirror-imaged and then 3D-printed in resin to produce an ear model based on which an anatomically symmetrical pressure device was constructed. Results: All devices were successfully fitted and well tolerated, with no reported discomfort or adverse reactions. The wire spring devices were effective in reducing a large keloids volume; however, frequent reactivation every two weeks was required to ensure continuous pressure application. Incorporating magnets in the customised design allowed controlled and uniform pressure application to small keloid-lesion morphology, with enhanced aesthetics and improved patient acceptance and compliance. The digitally assisted case achieved near-perfect anatomical symmetry with the contralateral ear, reducing operator dependency and fabrication guesswork. Conclusions: Customised pressure therapy devices, of magnetic and spring-based systems, alongside utilising digital technologies, offer effective, non-invasive management for auricular keloids and irregular ear helices as long as the patient is committed to wearing the device. Full article

The Hough transform (HT) is widely used in computer vision, tomography, and neural networks. Numerous algorithms for HT computation have been proposed, making their systematic comparison essential. However, existing comparative methodologies are either non-universal and limited to certain HT formulations or task-oriented, relying on application-specific criteria that do not fully capture algorithmic properties. This paper introduces a novel unified methodology for the systematic comparison of HT algorithms. It evaluates key characteristics, including computational complexity, accuracy, and auxiliary space complexity, while explicitly accounting for the property of self-adjointness. The methodology integrates both implementation-level and theoretical considerations related to the interpretation of HT as a discrete approximation of the Radon transform. A set of mathematically justified evaluation functions, not previously described in the literature, is proposed to support our methodology. Importantly, the methodology is universal, applicable across diverse HT paradigms, encompasses pattern-based and Fourier-based fast HT (FHT) algorithms, and offers a comprehensive alternative to existing task-specific methodologies. Its application to several state-of-the-art FHT algorithms ($FHT2DT$, $FHT2SP$, $ASD2$, $KHM$, and Fast Slant Stack) yields new experimentally confirmed theoretical insights, identifies $ASD2$ as the most balanced algorithm, and provides practical guidelines for algorithm selection. In particular, the methodology reveals that for image sizes up to 3000, the maximum normalized computational complexity increases as follows: $FHT2DT$ (1.1), $ASD2$ (15.3), and $KHM$ (30.6), while the remaining algorithms exhibit at least 1.1 times higher values. The maximum orthotropic approximation error equals 0.5 for $ASD2$, $KHM$, and Fast Slant Stack; lies between 0.5 and 1.5 for $FHT2SP$; and reaches 2.1 for $FHT2DT$. In terms of worst-case normalized auxiliary space complexity, the lowest values are achieved by $FHT2DT$ (2.0), Fast Slant Stack (4.0, lower bound), and $ASD2$ (6.8), with all other algorithms requiring at least 8.2 times more memory. Full article

Referring Remote Sensing Image Segmentation (RRSIS) is a representative multimodal understanding task for remote sensing, which segments designated targets from remote images according to free-form natural language descriptions. However, complex remote sensing characteristics, such as cluttered backgrounds, large-scale variations, small scattered targets and repetitive textures, lead to unstable visual grounding and further spatial grounding drift, resulting in inaccurate segmentation results. Existing approaches typically perform implicit visual–linguistic fusion across encoding and decoding stages, entangling spatial grounding with mask refinement. This tightly coupled formulation lacks explicit structural constraints and is prone to cross-modal ambiguity, especially in complex remote sensing layouts. To address these limitations, we propose a Structurally consistent and Grounding-aware Stagewise Reasoning Framework (SGSRF) that follows a grounding-first, segmentation-second paradigm. The framework decomposes inference into three cascaded stages with progressively imposed structural constraints. First, Cross-modal Consistency Refinement (CCR) lays the foundation for stable spatial grounding by enhancing visual–textual structural alignment via CLIP-based features and Structural Consistency Regularization (SCR), producing well-aligned multimodal representations and reliable grounding cues. Second, Grounding-aware Prompt (GPG) Generation bridges grounding and segmentation by converting aligned representations into complementary sparse and dense prompts, which serve as explicit grounding guidance for the segmentation model. Third, Grounding Modulated Segmentation (GMS) leverages the Segment Anything Model (SAM) to generate fine-grained mask prediction under the joint guidance of prompts and grounding cues, improving spatial grounding stability and robustness to background interference and scale variation. Extensive experiments on three remote sensing benchmarks , namely RefSegRS, RRSIS-D, and RISBench, demonstrate that SGSRF achieves state-of-the-art performance. The proposed stagewise paradigm integrates structural alignment, explicit grounding, and prompt-driven segmentation into a unified framework, providing a practical and robust solution for RRSIS in real-world Earth observation applications. Full article

Background/Objectives: Psoriatic disease is a systemic inflammatory condition associated with increased cardiometabolic risk, but the impact of contemporary systemic therapies and narrowband ultraviolet B (NB-UVB) phototherapy on vascular and systemic inflammatory markers remains incompletely characterized. We aimed to systematically synthesize prospective evidence on treatment-associated changes in vascular inflammation and systemic inflammatory biomarkers in adults with moderate-to-severe psoriatic disease. Specifically, we evaluated changes assessed by 18F-FDG PET/CT imaging and circulating biomarkers following biologic therapies or NB-UVB phototherapy. Methods: We systematically searched MEDLINE, Embase, Web of Science, Scopus, and CENTRAL from inception to 31 January 2026 for prospective interventional and observational studies in adults with psoriasis or psoriatic arthritis treated with biologic agents targeting TNF-α, IL-12/23, IL-17, or IL-23, or with NB-UVB phototherapy. Eligible studies were required to report serial assessments of vascular inflammation by 18F-FDG PET/CT (typically aortic target-to-background ratio) and/or systemic inflammatory markers (high-sensitivity C-reactive protein, interleukin-6, TNF-α, GlycA, or hematologic indices such as the neutrophil-to-lymphocyte ratio) over at least 8 weeks of follow-up. We imposed no language restrictions and included only full-text, peer-reviewed prospective studies. Risk of bias was evaluated using RoB 2 for randomized trials and ROBINS-I for nonrandomized studies. Random-effects meta-analyses were prespecified for outcomes reported by at least two clinically comparable studies; however, because of substantial heterogeneity in reporting and methodology, effect estimates were summarized using a structured narrative synthesis. Results: Thirteen prospective studies (n ≈ 900 adults, published 2015–2025) met inclusion criteria, including four studies with serial 18F-FDG PET/CT imaging and one additional PET/CT study providing baseline observational data on vascular inflammation, as well as eight biomarker-focused prospective cohorts. Across randomized mechanistic trials and observational studies, biologic therapies reduced aortic target-to-background ratio by approximately 6–12% over 12–24 weeks (e.g., mean change from 2.42 to 2.18 with TNF-α inhibition and from 2.51 to 2.20 with IL-17 blockade), and no study reported worsening of PET-derived vascular indices under effective systemic treatment. Biologic and other systemic therapies produced concordant reductions in hs-CRP (typically by 30–50%), IL-6, TNF-α, GlycA, and blood-count-derived indices including neutrophil-to-lymphocyte ratio, with biomarker improvements frequently paralleling reductions in cutaneous disease activity and cardiometabolic risk markers. Two NB-UVB cohorts demonstrated significant hs-CRP reductions of roughly 20–30% and modulation of vitamin D-related inflammatory proteins, suggesting systemic anti-inflammatory effects, although these changes appeared less pronounced than with biologic therapy and were not accompanied by vascular imaging. Conclusions: Contemporary systemic psoriasis therapies, particularly biologic agents targeting the IL-23/Th17 axis and TNF-α, are associated with consistent reductions in aortic vascular inflammation and broad improvements in systemic inflammatory biomarkers, whereas NB-UVB phototherapy confers more modest but measurable systemic anti-inflammatory effects, although the current evidence does not allow differentiation between individual biologic classes in terms of magnitude of effect. Although reductions in vascular and systemic inflammatory markers were observed across therapies targeting TNF-α, IL-12/23, IL-17, and IL-23, the small number of mechanistic imaging studies and absence of head-to-head comparisons do not allow robust differentiation between biologic classes or support a uniform class effect. The convergence of imaging and biomarker data reinforces psoriasis as a clinically relevant model of inflammation-driven atherosclerosis and supports the concept that effective control of psoriatic inflammation may contribute to cardiovascular risk modification, highlighting the need for integrated cardiovascular risk assessment in routine care. However, the imaging evidence base remains limited to four small mechanistic PET/CT studies with relatively short follow-up, which constrains the strength and generalizability of conclusions regarding vascular inflammation. Larger, adequately powered, event-driven prospective trials with standardized imaging and biomarker endpoints are needed to determine whether these vascular and systemic anti-inflammatory effects translate into reduced cardiovascular events in psoriatic disease; because of methodological and reporting heterogeneity across the 13 included studies, these conclusions are based on a structured narrative synthesis rather than a formal quantitative meta-analysis. PROSPERO registration number: CRD420261296646. Full article

Super-resolution microscopy has transformed biological imaging by enabling nanoscale visualization of cellular structures beyond the diffraction limit. However, its effective application in highly dense molecular environments still poses challenges. This is the case for 3D PhotoActivated Localization Microscopy (PALM) achieved through astigmatism in bacterial cells. The limited volume of a single bacterium highly increases the probability of the intensity profiles emitted by single chromophores to overlap, thus strongly decreasing the number of localizations, leading to dramatic undersampling. Dual-color 3D super-resolution in Escherichia coli is achieved through a combination of PALM with Expansion Microscopy (Ex-PALM). PALM provides high specificity through photoactivable (PA) fusion proteins and high localization precision, while ExM physically expands the specimen and separate densely packed molecules. This hybrid approach enables dual-color 3D single-molecule localization with about 3 nm spatial resolution, thus allowing one to measure distances down to the molecular scale. This is achieved by optimizing ExM protocols in bacteria to achieve a 4-fold isotropic expansion, by minimizing both chromatic aberrations and signal crosstalk, and by improving single-molecule sensitivity through highly selective inclined illumination. The method is applied to measure the spatial distribution of HisF and HisH proteins, involved in E. coli histidine biosynthesis. By tagging each protein with a photoactivable fluorescent protein, Ex-PALM reveals that after being synthetized, they co-localize in the bacterial volume with an average 3D distance of 19 nm. By combining labeling specificity with Ex-PALM, an effective method is developed for studying molecular organization in prokaryotes and in high-density samples in general, such as cell organelles or molecular condensates, with broad applications in microbiology, synthetic biology, and cellular biophysics. Full article

Low-count positron emission tomography (PET) is inherently affected by Poisson-dominated noise, which degrades image contrast, structural delineation, and quantitative reliability. This study systematically evaluated residual learning-based deep neural networks to investigate the influence of residual block depth on PET image restoration performance under low-count conditions. We employed a physically controlled striatum phantom, fabricated using 3D printing technology, to ensure reproducible acquisition conditions and controlled physical variability. PET images were acquired using a clinical PET/computed tomography (CT) system with list-mode acquisition. Low-count images reconstructed from short-duration acquisition were paired with high-count reference images reconstructed from extended acquisitions. We compared conventional filtering techniques, including median, Wiener, and modified median Wiener filters, with residual network (ResNet)-based models incorporating 8, 16, and 32 residual blocks. Image quality was quantitatively assessed using contrast-to-noise ratio (CNR), coefficient of variation (COV), line profile analysis, universal quality index (UQI), and perceptual image patch similarity (LPIPS). The results demonstrated that ResNet-based restorations substantially outperformed conventional filtering techniques in contrast recovery, signal stability, and structural preservation. The ResNet-16 model achieved the most balanced performance, yielding the highest CNR (9.02) and lowest COV (0.105), while also demonstrating superior structural and perceptual similarity, as indicated by UQI (0.9224) and LPIPS (0.0174), relative to the high-count reference images. Deeper network configurations exhibited diminishing returns and reduced structural consistencies. These findings indicate that an intermediate residual block depth is optimal for low-count PET image restoration and highlight the importance of architectural optimization in deep learning-based PET image enhancement with phantom-based evaluation frameworks. Full article

Against the global carbon neutrality target, predictive maintenance (PdM) of automotive engines represents a core technical strategy to advance the sustainable development of the automotive industry. Conventional single-modal diagnostic approaches for engine abnormal sound defects suffer from low accuracy and weak anti-interference capability. Existing multi-modal fusion methods fail to deeply mine the physical coupling between cross-modal features and often entail excessive model complexity, hindering deployment on resource-constrained on-board edge devices. To resolve these limitations, this study proposes a Physical Prior-Embedded Cross-Modal Attention (PPE-CMA) mechanism for lightweight multi-modal fusion diagnosis of engine abnormal sound defects. First, wavelet packet decomposition (WPD) and mel-frequency cepstral coefficients (MFCC) are integrated to extract time-frequency features from engine audio signals, while a channel-pruned ResNet18 is employed to extract spatial features from engine thermal imaging and vibration visualization images. Second, the PPE-CMA module is designed to adaptively assign attention weights to audio and image features by exploiting the physical coupling between engine fault acoustic and visual characteristics, enabling efficient cross-modal feature fusion with redundant information suppression. A rigorous theoretical derivation is provided to link cosine similarity with the physical correlation of engine fault acoustic-visual features, justifying the attention weight constraint (β = 1 − α) from the perspective of fault feature physical coupling. Third, an improved lightweight XGBoost classifier is constructed for fault classification, and a hybrid data augmentation strategy customized for engine multi-modal data is proposed to address the small-sample challenge in industrial applications. Ablation experiments on ResNet18 pruning ratios verify the optimal trade-off between diagnostic performance and computational efficiency, while feature distribution analysis validates the authenticity and effectiveness of the hybrid augmentation strategy. Experimental results on a self-constructed multi-modal dataset show that the proposed method achieves 98.7% diagnostic accuracy and a 98.2% F1-score, retaining 96.5% accuracy under 90 dB high-level environmental noise, with an end-to-end inference speed of 0.8 ms per sample (including preprocessing, feature extraction, and classification). Cross-engine and cross-domain validation on a 2.0T diesel engine small-sample dataset and the open-source SEMFault-2024 dataset yield average accuracies of 94.8% and 95.2%, respectively, demonstrating strong generalization. This method effectively enhances the accuracy and robustness of engine abnormal sound defect diagnosis, offering a lightweight technical solution for on-board real-time fault diagnosis and in-plant online quality inspection. By reducing engine fault-induced energy loss and spare parts waste, it further promotes energy conservation and emission reduction in the automotive industry. Quantified experimental data on fuel efficiency improvement and carbon emission reduction are provided to substantiate the ecological benefits of the proposed framework. Full article

Virtual tours are increasingly used in education, particularly when access to real environments is limited. This study examined how display mode and representation format affect subjective user experience in an educational virtual tour of a hospital operating room. A within-subject 2 × 2 design compared two representation formats (360° photographs vs. 3D models) and two display modes (desktop PC vs. immersive virtual reality using Meta Quest 2). Eighty-four university students completed the four visualization conditions and evaluated each experience using an adapted version of the QUXiVE questionnaire. Descriptive statistics and internal consistency indices were calculated, and each questionnaire dimension was analyzed using a two-way repeated-measures ANOVA with display mode and representation format as within-subject factors. A significant main effect of display mode was found for presence, engagement, immersion, flow, emotion, judgment, physical consequences, and perceived educational usefulness (all p < 0.001), but not for usability (p = 0.273). A significant main effect of representation format was observed for presence (p = 0.003), emotion (p = 0.018), and perceived educational usefulness (p = 0.015), whereas no significant interaction effects were found. These findings indicate that immersive VR had the strongest and most consistent effect on subjective user experience across both 360° and 3D virtual tours, although it was also associated with higher physical-consequence scores. By contrast, the effect of representation format was more limited. Overall, both approaches appear to be complementary educational resources, depending on pedagogical goals, available infrastructure, and desired levels of interactivity. Full article

Accurate fault diagnosis of rotating machinery in complex industrial environments demands an optimal trade-off between feature representation capability and computational efficiency. Existing single-modality models relying solely on 1D time-series signals or heavy 2D time-frequency images often fail to simultaneously capture high-frequency transient impacts and long-range degradation trends. CLiST (Complementary Lightweight Spatiotemporal Network), a novel lightweight multimodal framework driven by time-frequency fusion, was proposed to overcome this limitation. The architecture of CLiST employs a synergistic dual-stream design: a LightTS module efficiently extracts global operational trends from 1D vibration signals with linear complexity, while a structurally pruned LiteSwin integrated with Triplet Attention captures local high-frequency textures from 2D continuous wavelet transform (CWT) images. This mechanism establishes explicit cross-dimensional dependencies, effectively eliminating feature blind spots without excessive computational overhead. The experimental results show that CLiST not only achieves perfect accuracy on the fundamental CWRU benchmark but also exhibits exceptional spatial generalization when independently evaluated on non-dominant sensor axes of the XJTUGearbox dataset. Furthermore, validation on the real-world dataset (Guangzhou port) proves that the framework has excellent robustness to the attenuation of the signal transmission path and reduces the performance fluctuation between remote measurement points. Ultimately, CLiST delivers highly reliable AI-driven image and signal-processing solutions for vibration monitoring in industrial equipment. Full article