Search Results (25,555)

Human pose estimation (HPE) has advanced rapidly with deep learning, enabling a transition from specialized sensing and multi-view systems toward monocular RGB-based approaches. These developments have expanded applications in healthcare, robotics, sports analytics, and human–computer interaction. However, the growing diversity of deep learning paradigms, ranging from convolutional and recurrent models to graph-based and Transformer-based approaches, has resulted in a fragmented literature, making it difficult to systematically compare methods and guide system design. This paper addresses this challenge by providing a comprehensive survey of deep learning-based monocular HPE methods published over the past decade and introducing a unified modular framework. The proposed framework organizes HPE systems into six modular estimation paradigms, including single-image-based estimation, multi-frame-based estimation, Top-Down and Bottom-Up pose estimation strategies, 2D-to-3D pose reconstruction, and direct 3D estimation. Each module is analyzed in terms of representative approaches, design trade-offs, and practical considerations, supported by algorithmic formulations that outline the computational pipeline at each stage. Unlike prior surveys that primarily catalog methods or report benchmark results in isolation, this work emphasizes how component-level design choices relate to overall system performance. The paper summarizes performance trends on benchmarks including Human3.6M, COCO, and MPII, highlighting persistent challenges such as occlusion and viewpoint variation, and outlines future research directions including interaction-aware modeling, efficient deployment, and improved robustness under real-world conditions. Full article

Background/Objectives: HER2 overexpression and/or amplification defines a molecularly distinct subset of endometrial carcinomas (ECs) that may benefit from HER2-targeted therapies. However, HER2 testing algorithms remain non-standardized and vary across institutions. This study is a large single-institution audit of EC HER2 testing practices, using both gynecologic (ISGyP) and gastric cancer-specific scoring algorithms at a major academic center with a reference gynecologic oncology service and biomarker laboratory. Methods: HER2 immunohistochemistry (IHC) and whole-slide fluorescence in situ hybridization (FISH) were interpreted by subspecialty breast and gynecologic pathologists, with HER2 IHC performed on 494 tumor samples (2021–2025) and reflex FISH for equivocal cases. Results: Using ISGyP criteria, 15.0% (74/494) of tumors were HER2 IHC 3+, 44.5% (220/494) equivocal (2+), and 40.5% (200/494) were negative (0/1+). Among equivocal cases, 28.2% (58/205) demonstrated ERBB2 amplification, yielding an overall HER2-positive rate of 27.5% (132/480). Re-assessment with gastric scoring criteria demonstrated variability in HER2 classification, with high concordance in cytology specimens (100%) and resections (90.8%; K = 0.842, p < 0.001) but substantially lower concordance in biopsies (60.6%; K = 0.401, p < 0.001), mainly due to reclassification of equivocal cases. Notably, 47.9% (n = 34) of ISGyP-equivocal biopsy specimens were reclassified as HER2 IHC 3+ using gastric biopsy criteria, potentially expanding eligibility for T-DXd therapy. Conclusions: These findings highlight the evolving nature of HER2 testing in EC and demonstrate the significant impact of scoring methodology on HER2 interpretation. Our results support the development of EC-specific HER2 testing guidelines and a dual-reporting approach incorporating both ISGyP and gastric scoring criteria, with selective confirmatory FISH testing, to optimize patient selection for HER2-targeted therapies. Full article

The growing availability of genomic resources is changing how genetic diversity is studied in Vitis vinifera L. At the same time, it has become increasingly clear that a single reference genome cannot fully represent the complexity of a species characterised by high heterozygosity, clonal propagation and a long history of diversification. Recent grapevine pangenomes, super-pangenomes and graph-based resources have revealed forms of variation that are often overlooked in conventional reference-based analyses, including structural variants and gene presence–absence variation. Rather than providing another inventory of available datasets, this review examines how continued reliance on a single reference genome may influence the interpretation of grapevine diversity and what can be gained from a broader pangenomic perspective. Drawing on recent studies in grapevine and other crops, we discuss how these approaches are beginning to improve the representation of genetic diversity, uncover biologically relevant variation and strengthen links between genomic information and adaptive traits. We also examine the challenges that still limit their practical use, particularly the integration of genomic resources with functional studies and breeding programmes. In the end, the value of pangenomics will probably depend not only on generating additional genomic resources, but also on how effectively these can be translated into tools that support grapevine conservation, climate adaptation and varietal improvement. Full article

High levels of low-density lipoprotein cholesterol (LDL-c) have been recognized as the main causal factor of atherosclerotic cardiovascular disease (ASCVD) and are influenced by both genetic and environmental factors. Among genetic determinants, Familial Hypercholesterolemia (FH) is the most common monogenic disorder, caused by rare high-impact variants in genes involved in LDL uptake. Other monogenic causes of hypercholesterolemia include sitosterolemia, cerebrotendinous xanthomatosis and lysosomal acid lipase deficiency (LALD). However, monogenic disorders only account for a small proportion of inherited hypercholesterolemia. In many individuals, increased LDL-c levels are caused by the contemporary presence of different single-nucleotide polymorphisms (SNPs) with a moderate/low impact. These SNPs could be summarized through polygenic risk scores (PRS) that attribute relative weight to each of these. Another genetic determinant of hypercholesterolemic phenotypes is high levels of lipoprotein(a)—Lp(a). Lp(a) is an LDL particle modified by the binding of apolipoprotein(a)—apo(a)—which represents an independent risk factor for ASCVD. Lp(a) levels are mainly genetically determined by variation in the number of kringle IV type 2 (K-IV₂) repeats, as well as by several SNPs, and remain stable throughout life. The aim of this narrative review is to report an updated overview of the genetic mechanisms underlying hypercholesterolemia, including monogenic disorders, PRS and Lp(a), focusing on their potential repercussion in clinical practice by the integration into cardiovascular risk stratification beyond traditional clinical assessment. This integration could lead to a more comprehensive and individualized approach to cardiovascular prevention, with emerging perspectives including the possible use of artificial intelligence (AI). Full article

As the prevalence of myopia among adolescents continues to increase, the design and fabrication of myopia control lenses have become an important research direction in modern optics. Existing myopia control lenses mostly adopt purely refractive structures, which suffer from limited design freedom, insufficient chromatic aberration suppression, and relatively large lens thickness, thereby restricting further improvement of optical performance. This paper proposes a refractive–diffractive hybrid design and fabrication method for myopia control lenses. Centered on a harmonic diffractive optical element (HDOE), an optimization model is established to balance achromatization performance and fabrication feasibility. To address the challenges of small period width, tool shadow effect, and sensitivity to machining tolerances in diffractive lenses with large-aperture and high-additional-power, harmonic design is employed to increase the period width, thereby reducing fabrication difficulty and mitigating the influence of shadowing errors on diffraction efficiency. On this basis, two lenses with different phase structures are designed: one adopts a conventional diffractive correction phase to verify the role of HDOE in achromatization and edge-thickness reduction, while the other adopts a high-degree-of-freedom smooth phase to achieve a continuous multifocal visual effect. Both lenses are fabricated by single-point diamond turning (SPDT), and the effects of surface profile and machining parameters on performance are analyzed. Simulations and measurements show that the proposed method provides stable diffraction efficiency and effective chromatic aberration correction across the design band, while reducing the edge thickness by approximately 37.85% without additional thinning of the aspheric substrate. The results indicate that the refractive–diffractive hybrid design provides a feasible design and fabrication approach for functionally more complex myopia control lenses. Full article

Driving behavior recognition plays a crucial role in intelligent driving systems and road traffic safety. Due to insufficient driving experience and limited ability to allocate visual attention, novice drivers are considered a high-risk group for traffic accidents. Existing approaches primarily focus on experienced drivers and rely on single-modal eye-tracking data, making it difficult to model spatial attention distributions and long-term temporal dependencies simultaneously. Moreover, these methods are often affected by modality asynchrony during multimodal fusion, further limiting performance gains. To address these challenges, this study proposes a novice driver behavior recognition method based on bimodal eye-tracking features and a gated cross-modal attention fusion (GCMAF) mechanism. The model adopts a spatial–temporal dual-branch architecture. The spatial branch employs ResNet34 to extract eye-tracking heatmap features to represent the visual attention distribution. In contrast, the temporal branch integrates a 1D-CNN with the Mamba model to capture local dynamic patterns and long-range temporal dependencies. In the fusion stage, the GCMAF module is introduced to enhance cross-modal interactions, and a gating mechanism is further used to adaptively adjust modality weights, thereby mitigating the adverse effects of modality asynchrony. To validate the effectiveness and generalization ability of the proposed method, repeated experiments and five-fold cross-validation are conducted. The results demonstrate that the model achieves an average classification accuracy of 93.86% across four driving behavior categories, with standard deviations below 0.3%. Compared with baseline methods, paired t-test results show that the performance improvement is statistically significant (p < 0.01). Ablation studies further confirm the independent contribution of each component. Overall, the proposed method outperforms existing approaches in terms of accuracy and stability, providing effective support for driving behavior assessment and proactive safety warning systems. Full article

The outcomes of a standard geochemical, geophysical and petrographical approach to exploration at Lead Trial, a small prospect in central Scotland, have been compared to the interpretation of a parallel gold compositional study describing 703 gold particles from local in situ and alluvial occurrences. Standard exploration approaches identified a 4.5 km² zone hosting an array of numerous auriferous (to 17 g/t Au), vuggy, brecciated quartz-galena ± sphalerite veins culminating in the identification of a drill target. The gold study identified three gold compositional types: two 23–32 wt.% Ag alloys with a Zn-Pb-Cu mineral inclusion assemblage differentiated by sphalerite abundance, and a 5–16 wt.% Ag alloy with a Mo-Bi-Pb-Cu-Fe inclusion signature, yet to be correlated with either float or outcrop. Spatial distribution of the gold types indicates lateral variation and probably vertical variation within a single magmatic hydrothermal system. Integration of gold particle studies with early stages of exploration offers rapid insights into the nature and distribution of mineralization when very limited information is available and is mutually supportive of standard exploration approaches. Full article

The mechanical properties and real-time damage evolution of sustainable concrete (SC) containing 100% recycled concrete aggregate (RCA) under the combined action of hybrid steel and polyolefin fibers were studied. Inspired by solving the massive effects on the environment from construction waste, as well as to improve the lower mechanical performance of lower-grade RCA, the effect of combining high-stiffness hooked-end steel fibers and flexible macro-polyolefin fibers within RCA was investigated. Six different mix designs were considered: plain, single-fiber (100% steel and 100% polyolefin) and three hybrid composites with varying fractions of the steel/polyolefin fibers (25/75, 50/50, and 75/25). Compressive, tensile and flexural strengths were determined by mechanical testing. During compressive testing, the damage evolution was monitored using low-cost acoustic emission (AE) as a non-destructive technique. Cumulative hits analysis, amplitude distributions, and the statistical b-value parameter were used for damage characterization. The results show that steel fiber significantly increased compressive strength (an increase of up to 13.8%), and the 50/50 hybrid mix showed a high synergistic effect, yielding the highest tensile (4.86 MPa) and flexural (25.54 MPa) strengths. AE analysis identified different damage fingerprints: Based on amplitude analysis, steel-fiber composites exhibited high-amplitude events (which may be attributable to fiber pull-out); polyolefin-fiber composites generated medium-amplitude events (may have resulted from distributed microcracking); and hybrid mixes displayed a mixed amplitude distribution. The b-value analysis provided insight into progressive damage and revealed that the hybrid fibers induce stable, diffuse damage that prevents the brittle failure of plain recycled aggregate concrete (RAC). The results show that hybrid fiber reinforcement can be a reliable approach to enhance the mechanical performance and crack resistance of RAC. Furthermore, low-cost acoustic emission (AE) serves as an effective non-destructive method for monitoring damage progression within the material. Full article

In radio-over-fiber (RoF) links, optical single-sideband (OSSB) modulation is an effective method to mitigate power fading caused by chromatic dispersion. However, its low modulation efficiency leads to suboptimal link performance. To address this, we propose a tunable optical carrier-to-sideband ratio (OCSR) OSSB modulation scheme based on a dual-electrode Mach–Zehnder modulator (DEMZM) in a Sagnac loop. Firstly, by adjusting the OCSR, higher radio-frequency (RF) transmission efficiency can be achieved. The experimental results demonstrate that the proposed link provides a 6 dB improvement in received RF power compared to conventional SSB modulation schemes. Furthermore, this approach effectively optimizes nonlinear distortions in the link, achieving a 12.14 dB enhancement in spurious-free dynamic range (SFDR). For tests conducted with a broadband signal featuring a 15 GHz carrier frequency and 500 MHz bandwidth, the optimal error vector magnitude (EVM) reaches 4.88%. Additionally, the link performance can be flexibly improved by adjusting the polarization controller configurations for each channel, making it suitable for multi-user application scenarios. Full article

Wheat is a major staple crop, and storage mold growth poses a severe threat to grain safety and quality stability. Natural mold development in stored wheat exhibits subtle, localized, and highly heterogeneous characteristics. Existing unimodal methods and global fusion approaches generally suffer from insufficient local feature sensitivity, hindering fine-grained mold severity grading. To address this limitation, we propose a Mask-Guided Fine-Grained Fusion Network, a weakly supervised framework based on local RGB–HSI fusion. This framework employs a dynamic parallel A/B experimental design to construct time-matched proxy labels via weakly supervised learning. A standardized preprocessing pipeline including single-kernel extraction, foreground segmentation, and cross-modal registration is established to resolve RGB–HSI spatial misalignment, ensuring physical-level spatial consistency of multimodal features. The model incorporates a Foreground-Aware Spectral Recalibration (FASR) module to suppress background noise, a Mask-Guided Dilated Cross-modal Local Attention (MDCLA) mechanism to establish fine-grained local mappings between RGB visual phenotypes and hyperspectral responses, and a sample-level adaptive fusion strategy to dynamically weight features by modal reliability, enhancing representation of complex samples across all mold stages. Experiments show that the Mask-Guided Fine-Grained Fusion Network achieves 0.9689 classification accuracy, 0.9698 Macro-F1 score, and 0.0593 Mean Absolute Error (MAE), significantly outperforming state-of-the-art unimodal deep models and global attention fusion baselines. This work provides a proof-of-principle framework for fine-grained non-destructive mold risk assessment in stored wheat. Full article

Real-time scheduling in large-scale electric vehicle charging stations is challenged by stochastic vehicle arrivals, dynamic departures, limited charging resources, and station-level power constraints. To address this problem, this paper proposes a two-stage Supervised Service Matching and Reinforcement Power Dispatch (SMPD) framework, termed SMPD, which decomposes the original coupled scheduling problem into supervised service matching and reinforcement learning-based power dispatch. In the first stage, a supervised matching network learns EV-charger service suitability from historical charging-session records and determines service access decisions for feasible EV–charger pairs. In the second stage, a Soft Actor-Critic-based controller allocates continuous charging power to connected EVs under EV-side charging limits, charger capacity constraints, and the station-level total power constraint. The proposed framework is evaluated using public charging-session data from the ElaadNL dataset. Experimental results show that SMPD achieves lower average waiting time, higher average revenue, lower composite penalty, and comparable demand satisfaction compared with rule-based, single-stage reinforcement learning, and multi-agent baselines. Sensitivity and robustness analyses further indicate that SMPD maintains favorable scheduling performance and acceptable online decision time under the tested charger-scale settings and operational disturbance scenarios. These results suggest that the proposed two-stage design provides an effective and computationally tractable approach for real-time scheduling in large-scale EV charging stations. Full article

All rotating electrical machines are susceptible to vibrations arising from electromagnetic (EM) forces, electrical faults, mechanical defects, imbalance, and structural resonance. In Doubly Fed Induction Generators (DFIGs), such electromechanical vibrations are especially important because they can degrade reliability, increase noise, and lead to severe damage if resonance-prone operating conditions are not identified in time. Although fault diagnosis in DFIGs has been widely investigated using current, voltage, and flux signatures, comparatively fewer studies have examined fault-specific vibration behaviour under stator and rotor interturn faults (ITTFs), particularly through a coupled EM structural framework. In addition, prior vibration-based studies have not examined the influence of end winding ITTFs, its location, severity, and modal interaction investigating resonance risk. This paper considers vibration characteristics of a variable-speed 2.8 MW DFIG used in a grid-connected Type-3 wind turbine unit (WTU) at no-load operating condition. The DFIG is modelled in ANSYS Academic Research v 2022 R2 Maxwell for EM behaviour assessment for ITTFs in both stator and rotor windings along with modal analysis (MA) in ANSYS Workbench to examine the undamped stator and rotor modes over a range of frequencies. This coupled approach enables identification of vibration signatures associated with different ITTF types. The results show the magnetic flux density near faulty end-winding region increases with fault severity and ranges from 4.19 T to 4.39 T in proximity to faulty windings. A dominant modal frequency band of 60–65 Hz is identified, where stator and rotor modes coincide, creating probable resonance conditions. A severe vibration response is observed for single-phase stator ITTF, showing an amplitude of 2116 mm/s at 480 Hz for a larger number of shorted turns, indicating that asymmetric faults can produce stronger EM excitation than multi-phase faults. The main contribution of this paper is demonstration of a fault-specific, MA and vibration-based Condition monitoring system (CMS) implementation workflow for a DFIG. Unlike prior vibration-based studies that primarily focus on general machine vibration, mechanical faults, bearings, etc., this paper links stator and rotor ITTF induced EM excitation to modal characteristics, resonance behaviour, and measurable vibration signatures, establishing vibration analysis (VA) as a practical complementary technique for CMS of ITTFs in DFIGs. Full article

Pharmacogenomics AI offers significant potential for individualized drug therapy; however, its clinical benefits remain unevenly distributed. Models trained predominantly on European-ancestry data consistently underperform in non-European populations, with polygenic risk scores (PRS) showing an estimated 39–73% reduction in predictive accuracy in African-ancestry cohorts across complex traits. These disparities have driven increased interest in moving beyond single-layer genomic approaches. Multi-omics frameworks integrating genomic, transcriptomic, proteomic, and metabolomic data have emerged as a promising strategy to improve prediction across heterogeneous clinical populations, as each molecular layer provides distinct and complementary biological information. Among these layers, metabolomics may represent a particularly transferable component across populations. Metabolite profiles capture the downstream functional output of biological systems influenced by genetic, environmental, dietary, and microbiome-related factors, and may therefore be less reliant on ancestry-stratified allele frequency structures that underlie performance disparities in genomic models. This review synthesizes evidence regarding the mechanistic basis of genomic bias in pharmacogenomics AI, the emerging role of multi-omics integration, especially metabolomics, in improving predictive performance, and the current landscape of computational strategies for bias mitigation, including federated learning, transfer learning, domain adaptation, and synthetic data generation. Collectively, current evidence supports metabolomics-inclusive multi-omics frameworks as a biologically plausible, hypothesis-generating strategy to reduce reliance on ancestry-linked genomic features. However, direct evidence that such frameworks reduce ancestry-related bias in clinical AI outputs remains limited, underscoring the need for globally diverse datasets and prospective multi-population validation. Full article

Disentangling the coupled, nonlinear impacts of climate change and anthropogenic activities on vegetation dynamics is critical yet challenging for global change research. The Qinghai–Tibetan Plateau (QTP), a highly climate-sensitive and ecologically strategic region, serves as a vital arena for examining such complex socio-ecological attributions. Based on multi-source environmental datasets from 2000 to 2020, this study developed an integrated, spatially explicit framework coupling residual trend analysis (RESTREND) and GeoDetector to quantify individual drivers and nonlinear climate–human interactions. The QTP exhibited a significant, widespread greening trend during 2000–2020, featuring prominent spatial clustering with “High–High” clusters in the southeast and “Low–Low” clusters in the northwest. Attribution modeling revealed that vegetation dynamics were governed not by isolated variables, but by multifaceted, nonlinear synergies among precipitation, temperature, topography, vegetation type, and land-use change. Key interactive pairs, particularly elevation–temperature and slope–precipitation, dramatically increased explanatory power over single-factor models. Crucially, climate–human synergies explained substantially more variance than climate variables alone, bounded by a distinct elevational threshold: human activities dominated vegetation dynamics at mid-elevations (2500–3500 m), while climate factors took over as the primary controller at high altitudes (above 3500 m). Quantitatively, human activities induced vegetation improvement across 38.6% of the plateau, maintained stability in 35.8%, and caused degradation in 25.6%. By successfully merging trend decomposition with spatial stratified heterogeneity analysis, this study provides a transferable approach to uncoupling complex environmental interactions. These insights highlight the intensifying human footprint on alpine ecosystems and advocate for zone-specific adaptive management: mitigating human disturbances at mid-elevations and fostering climate adaptation in higher zones to preserve plateau resilience. Full article

Bridge decks are continuously subjected to high environmental exposure, traffic loading, and material aging, leading to progressive delamination which can negatively affect structural integrity and public safety. More specifically, subsurface delamination of concrete and corroded steel reinforcement must be repaired to keep the decks operational. Among non-destructive evaluation techniques, Ground-Penetrating Radar (GPR) and Infrared Thermography (IRT) offer complementary capabilities for detecting subsurface and near-surface defects; however, effective GPR-IRT data fusion remains challenging due to fundamental differences in sensing principles, spatial resolution and sensitivity. This study introduces a Physics-Enhanced Multi-Modal Fusion (PE-MMF) framework that integrates GPR and IRT data to improve delamination detection in reinforced concrete bridge decks. The proposed approach leverages transfer learning, cross-modal attention mechanisms, and gated fusion to enable robust learning from heterogeneous sensor inputs. Furthermore, a systematic feature selection protocol is integrated to identify physically meaningful indicators that remain consistent across different bridges, enhancing generalization capability. The framework is trained and validated using the publicly available SDNET2021 dataset, comprising co-registered GPR and IRT measurements from five in-service bridge decks with verified delamination ground truth. Results demonstrate substantial performance improvements, with average F1-score gains of up to 55% over IRT-based methods and 25% over GPR-based methods across all tested bridges. Comparative analysis against state-of-the-art methods confirmed the superior generalization capability of the proposed multi-modal approach over single-modality approaches. The findings highlight the potential of deep learning-based sensor fusion as a scalable and data-efficient decision-support tool to prioritize regions for detailed physical investigation during long-term infrastructure monitoring. Full article