Search Results (26,454)

Drug-induced ocular toxicity is difficult to predict and evaluate, particularly for prostaglandin F2α (PGF2α) analogs used in the management of glaucoma. Traditional global predictive models, which are built with large and various datasets (n = 6187), provide systematic false-negative results for groups of structurally uniform compounds. To address this limitation, we developed special local classification models for PGF2α analogs. A structurally consistent local training dataset (n = 350) was assembled using Murcko scaffold filtering and Tanimoto similarity selection (≥0.5). Binary classification models were built using three different molecular fingerprints in combination with machine learning and two deep learning methods. Although two-dimensional molecular representations do not encode stereochemistry, computational predictions still apply to the shared two-dimensional scaffold of these compounds. These improved models were used to predict the ocular toxicity of latanoprost and its related impurities. The local models (n = 350) provided accurate toxicity predictions for latanoprost, latanoprost acid, 15(S)-latanoprost, and the trans-5,6-latanoprost isomer, whereas all 16 global computational models provided false-negative results. The experimental assessment performed with primary rabbit corneal epithelial cells (pRCECs) and a human corneal epithelial cell line (HCE-T) confirmed the cytotoxic effects, which were in agreement with the predictions made by the models. Among the tested compounds, 15(S)-latanoprost exhibited the highest cytotoxicity (IC50 = 86.22 μM, 95% CI: 82.90–89.56 μM), followed by trans-5,6-latanoprost (IC50 = 106.70 μM, 95% CI: 103.4–110.0 μM) and latanoprost (IC50 = 112.60 μM, 95% CI: 106.7–118.6 μM). At the standard therapeutic dosage (0.005%), no significant toxic response was observed. Virtual molecular docking was employed to explore the mechanism, and all analogs docked favorably into the quinone-binding channel of mitochondrial complex I and the catalytic cleft of SIRT3. These results show how a localized modeling approach is better able to capture structure–toxicity correlations among chemically similar compounds and highlight the critical need for rigorous impurity management in latanoprost products, particularly regarding 15(S)-latanoprost. Full article

Accurate three-dimensional (3D) characterization of geothermal reservoirs is crucial for resource assessment and development, yet it remains a significant challenge due to sparse direct observations and the complex, heterogeneous nature of subsurface geology. While deep learning has emerged as a powerful tool for subsurface modeling, existing approaches often struggle to preserve critical fine-scale structural details and adaptively focus on geologically informative regions, limiting their effectiveness for geothermal applications. To address these limitations, we propose Gen-LEP, a novel deep generative neural network specifically designed for geothermal reservoir characterization. The proposed framework integrates a key component of local edge pattern (LEP) learning module to enhance the preservation of lithological boundaries and structural discontinuities. The LEP learning module is embedded within a conditional generative framework to effectively learn the nonlinear relationships between sparse conditioning data and complex 3D reservoir structures. We evaluate our method on a geothermal reservoir modeling dataset. Experimental results demonstrate that Gen-LEP can achieve accurate reconstruction and preserve complex geological boundaries. Gen-LEP can provide an effective deep learning framework that improves the fidelity of geothermal reservoir reconstructions by explicitly addressing the specific spatial characteristics of subsurface geological systems. Full article

Goal: Deep learning-based motor imagery EEG classification is limited by data scarcity, which constrains model generalization and performance. Methods: We propose a dual-cascade generative adversarial network (dcGAN) framework with a variable focused attention (VFA) module for MI-EEG data augmentation. The first stage learns latent frequency-domain priors from random noise through an adversarial training scheme; the second stage then synthesizes artificial EEG samples with a U-Net generator conditioned on these priors, augmented by the VFA module and a time-domain consistency loss. A VFA-enhanced EEGNet is subsequently trained on the combination of real and generated samples for classification. Results: On the BCI Competition IV 2a and 2b datasets, the proposed method achieves classification accuracies of 84.92% and 91.79%, with Cohen’s Kappa coefficients of 0.79 and 0.81, respectively, outperforming baseline methods. Conclusions: The integration of structured frequency-domain priors and attention mechanisms improves the fidelity of generated EEG samples, which in turn enhances downstream classification performance. Full article

The harvesting of fruit and vegetable crops has long been plagued by prominent issues such as high labor costs, low harvesting efficiency, and high fruit damage rates. The application of object recognition technology has enabled harvesting robots to identify, detect, and locate crops in certain agricultural scenarios, achieving a degree of automated harvesting. However, these systems still suffer from shortcomings such as poor robustness in complex environments, insufficient generalization ability, and high model deployment costs, which significantly limit their large-scale application in agricultural harvesting equipment. This paper comprehensively reviews recent literature in the field of fruit and vegetable target recognition. It summarizes how current research focuses on the implementation principles and directions for the improvement of mainstream methods—including digital image processing, traditional machine learning, and deep learning—while also identifying the remaining issues and challenges facing current technology in terms of algorithmic model real-time performance, robustness, and generalization ability. In the future, target recognition technology is expected to achieve breakthroughs through approaches such as multimodal feature fusion, large-scale models, and semi-supervised learning, evolving toward higher accuracy, faster processing speeds, and easier deployment, thereby providing technical support for the large-scale implementation of smart agriculture. Full article

In grid-connected smart integrated energy systems with high shares of renewable generation, source-side variability and inadequate coordination among battery storage, other energy carriers, and the external grid limit local renewable-electricity utilization and impede deep decarbonization. This study proposes a machine-learning-assisted, renewable-driven framework for multi-energy coupling and scenario-based multi-objective optimization of electricity–heat–hydrogen–storage systems. Historical meteorological and load data are processed using K-means clustering and Latin hypercube sampling to construct representative operating scenarios across multiple volatility regimes and characterize source–load uncertainty. The equipment model includes photovoltaic arrays, wind turbines, heat pumps, electrolyzers, fuel cells, grid-interactive battery energy storage, thermal storage, and hydrogen storage; cross-carrier conversion dynamics and emissions from purchased electricity and natural gas are embedded in the energy-balance constraints. A mixed-integer linear programming formulation then co-optimizes battery charging and discharging, grid exchange, and other multi-energy flows with respect to operating cost, carbon emissions, and renewable-energy curtailment. At 95% renewable-energy penetration, the proposed method achieves a renewable-energy absorption rate of 91.6% and a curtailment rate of 8.4%. Across the carbon-price cases, annualized operating cost ranges from 126.5 × 10⁴ to 141.2 × 10⁴ USD yr⁻¹, while carbon-emission intensity ranges from 26.4 to 38.5 gCO₂/kWh_eq. Under the specified high-risk grid disturbances, the coordinated strategy limits load shedding to 1.8%—73% below deterministic scheduling and 79% below the heuristic benchmark—and maintains 92.6% hydrogen self-sufficiency. These results provide a data-driven modeling and decision framework for battery–grid coordination and deep decarbonization in smart integrated energy systems. Full article

Modern financial risk management increasingly relies on transformer-based anomaly detection, although the sensitivity of these methods to the underlying volatility model specifications remains unexplored. This study thoroughly compares four GARCH variants—symmetric GARCH and asymmetric specifications (EGARCH, GJR-GARCH, APARCH)—across two state-of-the-art transformer architectures (TranAD for point anomalies, VTT for regime detection) using S&P 500 returns spanning 1980 to 2026. The investigation addresses a fundamental question for practitioners integrating econometric and deep learning methods: does the additional complexity of asymmetric volatility modeling yield meaningfully different anomaly detection when passed through transformer architectures? The analysis reveals that the GARCH specification affects anomaly severity rankings rather than detection consensus, with high cross-model agreement (minimum Jaccard of 0.88 for TranAD and 0.76 for VTT) despite statistically significant differences in score distributions. Asymmetric models exhibit extended post-crisis sensitivity, driven by their stronger response to negative shocks—the leverage effect captured by the GJR-GARCH threshold term—rather than by greater persistence; indeed, the symmetric GARCH exhibits the longest half-life. This enables specification selection based on risk philosophy: conservative monitoring via GJR-GARCH or efficient normalization via symmetric specifications. The choice of detection method—point versus regime identification—proves more consequential for anomaly detection performance than volatility model specification. Full article

Background/Objectives: First NIBUT, Average NIBUT, and tear meniscus height (TMH) are routinely used to characterize tear film stability and tear volume at individual visits, whereas their longitudinal behavior across the hospital-attending population is less well characterized. Aggregating routine examinations over time may provide a continuous hospital-level view of ocular surface status and enable the short-term forecasting of expected trajectories. We therefore evaluated a benchmark-first framework for monthly aggregated adult ocular surface indicators. Methods: This retrospective time-series study used de-identified adult eye-level Keratograph 5M records from July 2018 to November 2023. After cleaning, 31,492 records from 13,749 patients and 15,334 examination occasions were aggregated across 65 calendar months (64 observed months; March 2020 had no eligible records). Seven benchmark models and two exploratory deep learning comparators were evaluated using eight rolling-origin 3-month test windows. Results: Linear trend had the lowest mean origin-level macro-normalized RMSE (0.875; 95% bootstrap CI, 0.497–1.421), followed by simple exponential smoothing (0.907) and SARIMA(1,0,0)(1,0,0,12) (0.940). The paired difference between linear trend and simple exponential smoothing was small and did not show clear superiority (mean difference, −0.032; 95% bootstrap CI, −0.217 to 0.135; p = 0.789). The model with the lowest pooled error differed by target, while patient-month and sample-size-weighted sensitivity analyses gave a similar overall benchmark pattern. The exploratory LSTM and Transformer did not show a consistent advantage over the leading simple models. Conclusions: In this short hospital-level monthly series, simple forecasting models remained competitive, while no single model showed consistent superiority across forecast origins and sensitivity analyses. By extending ocular surface assessment from isolated examinations to longitudinal hospital-level trajectories, this framework provides a methodological basis for monitoring temporal changes in tear film stability and tear volume and for future quality monitoring, clinical, and epidemiological applications. Full article

Monitoring forest regeneration is vital for sustainable management, yet traditional ground surveys and early aerial imagery methods face cost, labor, and resolution limitations. Unmanned Aerial Vehicles (UAVs) offer cost-effective, flexible platforms for acquiring ultra-high-resolution data to address these gaps. This paper reviews recent research on UAV applications in forest regeneration surveys (FRS), tracing the evolution from field-based surveys and conventional aerial approaches to current UAV practices, and synthesizing developments in data acquisition, processing workflows, and analysis. It contrasts established Canopy Height Model (CHM) and point-cloud approaches with the growing use of deep learning, particularly Convolutional Neural Networks (CNNs) for seedling detection, crown delineation, density, height estimation, and species classification. Particular attention is given to accuracy assessment, examining sampling design, reference data, prediction-to-reference matching, and evaluation metrics, and highlighting the disconnect between traditional map-validation principles and standard deep-learning metrics that often neglect background classes. The case study applying Mask R-CNN to red pine seedlings in an Adirondack Park plantation achieved stand-level recall of 70.3% and precision of 98.7%, while plot-level DL detections represented only 36.8% of the field-observed seedling count. These results demonstrate the potential of DL for reliably identifying visible red pine seedlings while highlighting its limitations for complete regeneration inventories, particularly when seedlings have small crown sizes. Full article

Bearings are core components of rotating machinery, and the development of precise and efficient fault diagnosis technologies is of paramount importance for realizing early warning and accurate localization of faults. This paper briefly analyzes bearing fault mechanisms and provides a comprehensive review and critical commentary on the research progress of bearing fault diagnosis methods, outlining future development trends. Specifically, this study begins by introducing common bearing fault types and reviewing the advancements in fault signal acquisition techniques. Subsequently, it categorizes fault diagnosis methods based on vibration signals and critically evaluates their respective research methodologies. Furthermore, focusing on non-contact acoustic signal diagnosis, the paper summarizes mainstream technical pathways for acoustic signal denoising and highlights innovative applications of deep-learning models tailored to acoustic characteristics. Finally, addressing the urgent demands for industrial deployment, future research directions are projected from three perspectives: the deep integration of physics-driven and data-driven multi-modal fusion, interpretable diagnosis assisted by Large Language Models (LLMs), and lightweight engineering deployment. This work aims to provide a reference for constructing an all-scenario intelligent monitoring system. Full article

The development of accurate, robust and adaptive methods for monitoring the optical parameters of fiber-optic sensors (FOS) is one of the priority tasks in the field of precision measurements, especially in the context of rapidly growing requirements for intelligent monitoring systems. This paper presents a comprehensive approach to the use of modern deep learning algorithms for analyzing and processing spectral data coming from FOS. The proposed solution is based on the use of convolutional neural networks (CNN) for the automatic extraction of informative features, as well as autoencoders for noise suppression and signal restoration. A hybrid architecture combining CNN and recurrent neural networks (RNN) was developed. The experiments conducted confirmed the effectiveness of the evaluated models. On the independent regression test set, the CNN-only model achieved a macro-averaged R²-based prediction score of 98.2% without added noise and 89.4% under high-noise conditions; on a separate temporal test sequence, the hybrid CNN + RNN model achieved 95.0% compared with 88.0% for CNN alone. The presented approach has high resistance to noise and the ability to scale to various types of FOS. At the conclusion, the prospects for the practical applications of the proposed system are discussed: the structural monitoring of buildings and structures and the automation of processes in industry and energy, with an emphasis on reliability, autonomy and integration with existing platforms. Full article

Wrist fractures present significant challenges in clinical diagnosis, often leading to treatment delays and compromised recoveries. Manual diagnosis is resource-intensive and error-prone. To address these challenges, we develop DeepWristFNet, a compact convolutional architecture designed for end-to-end wrist fracture classification using a small dataset of 193 wrist X-ray images. The DeepWristFNet architecture integrates multi-scale convolutional operations with Fire and Shuffle modules within a compact network design, followed by fully connected layers for binary classification. We applied data pre-processing techniques such as data augmentation, image enhancement, and image resizing to increase the number of images, improve image quality, and resize images to match the DeepWristFNet input size. The proposed method comprised three phases. In the first phase, we trained, validated, and tested end-to-end and achieved validation and testing accuracies of 99.04% and 87.93%, respectively. Testing was performed on a hold-out subset of image instances that was kept separate from model development. The evaluated hold-out images originated from the same dataset distribution and included the corresponding augmented variants. In the second phase, we further evaluated the learned representation by extracting deep features from the first fully connected layer of DeepWristFNet. ReliefF was then used to select informative features, which were subsequently evaluated using 10 conventional machine learning classifiers. Out of 10 classifiers, 5 classifiers, i.e., Efficient linear SVM, quadratic SVM, Narrow NN, wide NN, and medium NN, achieved 100% testing accuracy on unseen samples. In the third phase, an auxiliary Fuzzy Inference System provides an intensity-based foreground-background representation of the X-ray images. This representation provides complementary visual information for interpretation but is not intended to directly classify or localize fractures. Grad-CAM is additionally used to visualize image regions contributing to the DeepWristFNet predictions, providing a model-specific explanation of the classification decision. Additionally, we evaluated how well the proposed DeepWristFNet approach performed against cutting-edge deep transfer learning models. In the evaluated experiments, DeepWristFNet outperformed the compared pre-trained deep learning architectures on the unseen hold-out subset from the same dataset distribution (test set). This study demonstrates the potential of DeepWristFNet for wrist fracture classification under a small-data setting. However, further evaluation on larger, independently collected clinical datasets is required to establish its robustness, generalizability, and suitability for clinical decision support. Full article

Cross-condition bearing fault diagnosis suffers from severe performance degradation due to domain shift across different operating conditions, especially when only a few labeled target-domain samples are available. To address this challenge, this paper proposes a progressive attention-guided two-stage transfer learning framework for few-shot bearing fault diagnosis across different fixed operating points. First, the raw time-domain vibration signals are fused with frequency-domain representations extracted by short-time Fourier transform (STFT) to enhance fault feature representation. Then, a progressive attention-guided feature learning strategy is developed by integrating dual efficient channel attention (ECA) modules into a deep one-dimensional convolutional neural network (1D-CNN), enabling the network to adaptively emphasize fault-sensitive features while suppressing redundant information. Subsequently, a two-stage transfer learning strategy is designed, consisting of transferable feature learning from the source domain and few-shot adaptation to the target domain. During target-domain adaptation, key feature extraction layers are frozen, and a sample-balanced optimization mechanism is introduced to alleviate the dominance of source-domain samples during joint training. Experimental results on the Case Western Reserve University (CWRU) bearing dataset demonstrate that the proposed method achieves an average accuracy of 99.96% across three cross-condition transfer tasks. Furthermore, experiments conducted on a self-built shaft system dataset show that the proposed method achieves an average accuracy of 87.11% under three representative transfer scenarios. The results verify that the proposed framework effectively mitigates domain shift and enables accurate bearing fault diagnosis with limited labeled target-domain samples. Full article

Artificial intelligence (AI) and machine learning (ML) are transforming the landscape of rheumatological and systemic inflammatory disease management, offering unprecedented capacity to integrate complex, multidimensional data for diagnostic support, disease monitoring, and therapeutic decision-making. This comprehensive narrative review, based on a non-systematic literature search of PubMed/MEDLINE and Google Scholar combined with the authors’ clinical expertise, provides a clinically oriented synthesis of current and emerging AI applications across the full spectrum of immune-mediated inflammatory diseases—including rheumatoid arthritis, systemic lupus erythematosus, vasculitis, inflammatory bowel disease, psoriatic arthritis, systemic sclerosis, inflammatory myopathies, and sarcoidosis—with particular attention to applications that have demonstrated or are approaching clinical utility. We discuss deep learning-based image analysis, natural language processing of electronic health records, multi-omic biomarker discovery, and the application of Fourier transform-based signal processing to biological time series as a novel approach to continuous disease monitoring. Fourier transform methods—already foundational in MRI reconstruction, cardiac electrophysiology, and clinical neurophysiology—are here systematically extended to rheumatological and inflammatory disease signals, including accelerometry, electromyography, heart rate variability, and longitudinal biomarker time series. The phenomenon of large language model hallucination—particularly critical in rare inflammatory diseases—is addressed alongside retrieval-augmented generation as a mitigation strategy. We further argue that AI-driven methods do not merely improve the interpretation of clinical data, but fundamentally expand what is observable—with profound epistemological implications for clinical knowledge transmitted through generations of medical tradition. Ethical considerations and future directions toward precision inflammatory disease medicine are outlined. Full article

This survey presents a structured review of deep learning-based techniques for image data hiding, proposing a three-paradigm taxonomy organized by the method’s operational relationship to the carrier image. We classify existing methods into modification-based, synthesis-based, and logic-based approaches. In the modification-based tier, we trace the architectural progression from foundational Convolutional Neural Networks and Generative Adversarial Networks to high-capacity Invertible Neural Networks and Transformers, analyzing their distinct trade-offs between embedding capacity, imperceptibility, and robustness. In the synthesis-based tier, we examine how Diffusion Probabilistic Models and generative adversarial frameworks reframe data hiding as a carrier generation problem rather than a pixel editing task. This paradigm encompasses both generative steganography (where carriers are synthesized from scratch) and proactive watermarking (where provenance is embedded during AI content generation). In the logic-based tier, we review zero-watermarking and coverless steganography, where ownership is established through feature extraction and semantic mapping without modifying any image, a critical property for sensitive domains such as medical imaging. Finally, we identify four persistent infrastructure gaps: benchmarking fragmentation, narrow robustness evaluation, domain generalization failures, and computational infeasibility that prevent real-world deployment despite architectural progress, and we propose concrete research directions. Full article

Timely and accurate classification of lung diseases from chest X-ray images remains an important healthcare challenge. Machine-learning and deep-learning methods can support automated classification, but their evaluation may be affected by class imbalance, feature redundancy, dataset leakage, and computational cost. This paper evaluates an artificial immune system (AIS)-evolved decision-tree ensemble using fold-specific ResNet18 features. All within-dataset experiments use duplicate-family-aware five-fold splits. Within each fold, standardization and adaptive principal component analysis (PCA) are fitted to the training features, and the Synthetic Minority Over-sampling Technique (SMOTE) is applied only to the reduced training data. Each candidate tree is assigned an affinity based on out-of-bag macro-F1. In the primary run, mean within-dataset macro-F1 was 98.07%, 99.48%, and 98.26% for Datasets 1–3, respectively, and 96.66% for the exploratory Dataset 4. Because of extensive cross-dataset image reuse and conflicting labels, Dataset 4 does not provide independent evidence of clinical lung-cancer detection. A five-seed repeated-initialization analysis repeated the complete fold-specific feature and classification pipeline while preserving the same folds. Mean macro-F1 differences between AIS and the prespecified static comparator for each dataset, calculated as AIS minus the comparator, were $-0.18$, 0.00, $-0.35$, and $-0.13$ percentage points for Datasets 1–4, respectively. Using the same sign convention, mean differences between AIS and the fixed random tree ensemble ranged from $-0.05$ to $+0.05$ percentage points. Population diagnostics showed that evolution improved individual-tree macro-F1 but reduced pairwise disagreement, without a consistent majority-vote gain. Median latency from an already decoded image to prediction ranged from 38.86 to 61.20 ms on one CPU thread and from 3.43 to 6.36 ms on an RTX 4090. The results do not establish a practically important or consistent predictive advantage from AIS evolution. The study provides a reproducible and duplicate-controlled framework for evaluating AIS-based tree ensembles. Full article