Search Results (495)

Fly ash-based geopolymer concrete (FAGC) is a sustainable alternative to Portland cement concrete, offering significant reductions in carbon emissions while maintaining sufficient strength. This study proposes a three-stage framework for developing empirical formulae to accurately and interpretably predict FAGC compressive strength. In the first stage, predictive models were developed using linear regression (LR), deep neural network (DNN), and residual neural network (ResNet) approaches. Among these, the ResNet model achieved the highest predictive accuracy and effectively captured the complex nonlinear relationship between mix components, curing conditions, and compressive strength. In the second stage, global sensitivity analysis identified sodium silicate content, curing time, sodium hydroxide molarity, and water content as the most influential variables. Additionally, the interaction between fine aggregate content and curing temperature was found to have a substantial effect on strength development. In the final stage, an empirical formula was developed based on key variables and their interactions, providing a simple yet reliable tool for practical strength prediction with reduced computational requirements. The proposed framework is expected to bridge the gap between machine-learning prediction and applicability to support mix design optimisation and promote the wider adoption of sustainable geopolymer concrete in construction applications. Full article

With the increasing integration of distributed energy resources and the growing variability of multiple loads, distribution networks face significant uncertainties in measurement data, line parameters, and topology. Traditional state estimation methods, such as weighted least squares, rely on accurate network parameters and are therefore highly sensitive to measurement noise and topology variations. To address these challenges, this work proposes a comprehensive data-driven framework for ADN state estimation that features a novel integration of an improved deep residual network (i-ResNet) and transfer learning. An improved deep residual network (i-ResNet) is developed to enable fast and robust state estimation without dependence on online parameters, even under uncertain data conditions. Furthermore, a transfer learning–based model is introduced to accommodate topology changes by leveraging historical data from multiple network configurations. Experimental studies on the IEEE 33-bus and 118-bus test systems are conducted to evaluate the performance of the proposed approach. The results demonstrate that the proposed method achieves higher accuracy and faster convergence than conventional techniques, with voltage magnitude errors consistently maintained below 1%. Full article

Flywheel energy storage systems play an important role in frequency regulation and power quality control within modern power grids, yet the fault signals generated by defects in their rolling bearings are typically indistinct, making direct diagnosis difficult. Raw noisy signals often yield unsatisfactory diagnostic performance when directly processed by neural networks. Although MOMEDA (Multipoint Optimal Minimum Entropy Deconvolution Adjusted) can effectively extract impulsive fault components, its performance is highly dependent on the selected fault period and filter length. To address these issues, this paper proposes an improved fault diagnosis method that integrates MOMEDA-based periodic extraction with a neural network classifier. The Artificial Fish Swarm Algorithm (AFSA) is employed to adaptively determine the key parameters of MOMEDA using multi-point kurtosis as the optimization objective, and the optimized parameters are used to enhance impulsive fault features. The filtered signals are then converted into image representations and fed into a ResNet-18 network (a compact 18-layer deep convolutional neural network from the residual network family) to achieve intelligent identification and classification of bearing faults. Experimental results demonstrate that the proposed method can effectively extract and diagnose bearing fault signals. Full article

Iris recognition continues to pose challenges for deep learning models, despite its status as one of the most reliable biometric authentication techniques. These challenges become more pronounced when training data is limited, as subtle, high-dimensional patterns are easily missed. To address this issue and strengthen both feature extraction and recognition accuracy, this study introduces IRIS-QResNet, a customized ResNet-18 architecture augmented with a quanvolutional layer. The quanvolutional layer simulates quantum effects such as entanglement and superposition and incorporates sinusoidal feature encoding, enabling more discriminative multilayer representations. To evaluate the model, we conducted 14 experiments on the CASIA-Thousands, IITD, MMU, and UBIris datasets, comparing the performance of the proposed IRIS-QResNet with that of the IResNet baseline. While IResNet occasionally yielded subpar accuracy, ranging from 50.00% to 98.66%, and higher loss values ranging from 0.1060 to 2.0640, comparative analyses showed that IRIS-QResNet consistently outperformed it. IRIS-QResNet achieved lower loss (ranging from 0.0570 to 1.8130), higher accuracy (ranging from 66.67% to 99.55%), and demon-started improvement margins spanning from 0.1870% in the CASIA End-to-End subject recognition with eye-side to 16.67% in the MMU End-to-End subject recognition with eye-side. Loss reductions ranged from 0.0360 in the CASIA End-to-End subject recognition without eye-side to 1.0280 in the UBIris Non-End-to-End subject recognition. Overall, the model exhibited robust generalization across recognition tasks despite the absence of data augmentation. These findings indicate that quantum-inspired modifications provide a practical and scalable approach for enhancing the discriminative capacity of residual networks, offering a promising bridge between classical deep learning and emerging quantum machine learning paradigms. Full article

Background: The mortality and morbidity rates of cardiovascular disease (CVD) are rising sharply in many developed and developing countries. CVD is a fatal disease that requires early and timely diagnosis to prevent further damage and ultimately save patients’ lives. In recent years, numerous studies have explored the automated identification of different categories of CVDs using various deep learning classifiers. However, they often rely on a substantial amount of data. The lack of representative training samples in real-world scenarios, especially in developing countries, poses a significant challenge that hinders the successful training of accurate predictive models. In this study, we introduce a framework to address this gap. Methods: The core novelty of our framework is the combination of Multi-Resolution Wavelet Features with Scale-Invariant Feature Transform (SIFT) keypoint density maps and a lightweight residual attention neural network (ResAttNet). Our hybrid approach transforms one-dimensional ECG signals into a three-channel image representation. Specifically, the CWT is used to extract hidden features in the time-frequency domain to create the first two image channels. Subsequently, the SIFT algorithm is implemented to capture additional significant features to generate the third channel. These three-channel images are then fed to our custom residual attention neural network to enhance classification performance. To tackle the challenge of class imbalance present in our dataset, we employed a hybrid strategy combining the Synthetic Minority Over-sampling Technique (SMOTE) with Edited Nearest Neighbors (ENN) to balance class samples and integrated Focal Loss into the training process to help the model focus on hard-to-classify instances. Results: The performance metrics achieved using five-fold cross-validation are 99.60% accuracy, 97.38% precision, 98.53% recall, and 97.37% F1-score. Conclusions: The experimental results showed that our proposed method outperforms current state-of-the-art methods. The primary practical implication of this work is that by combining a novel, information-rich feature representation with a lightweight classifier, our framework offers a highly accurate and computationally efficient solution, making it a significant step towards developing accessible and scalable computer-aided screening tools. Full article

Industrial predictive maintenance at the edge faces persistent challenges such as extreme class imbalance, limited labeled failure data, and the need for efficient yet scalable AI models. This paper proposes a transfer learning-based edge AI framework that addresses these challenges through a signal-to-image transformation and fine-tuning of deep residual networks (ResNet). One-dimensional sensor signals are converted into two-dimensional RGB images, enabling the use of powerful convolutional architectures originally trained on large-scale datasets. The approach emulates an edge–cloud synergy, where knowledge distilled from large pre-trained models is efficiently adapted and executed on resource-constrained edge environments. Trained on less than 5% of the original dataset, the model achieves a negative predictive value of 96.53%, significantly reducing classification cost and outperforming both conventional deep learning and traditional machine learning methods. The results demonstrate that transfer learning-driven edge intelligence offers a cost-effective, scalable, and generalizable solution for predictive maintenance and industrial automation under data scarcity. Full article

Open-winding permanent-magnet synchronous motors feature flexible control and a high fault-tolerance capability, making them widely used in high-reliability and high-power scenarios such as military equipment and electric locomotives. To address the issues that traditional model predictive control fails to balance, such as zero-sequence current suppression, system loss optimization and the reliance of weight parameter design on experience (with online optimization consuming excessive resources), this paper proposes an OW-PMSM MPC strategy for loss optimization and a weight design method based on a residual neural network. Specifically, the former strategy adds a zero-sequence current suppression term and a loss quantification term to the MPC cost function, enabling coordinated control of the two objectives; the latter establishes a mapping between weight parameters and motor performance via ResNet (which avoids the gradient vanishing problem in deep networks) and outputs optimal weight parameters offline to save online computing resources. Comparative experiments under two operating conditions show that the improved MPC strategy reduces system loss by 25%, while the ResNet-based weight design improves the performance of the drive system by 30%, fully verifying the effectiveness of the proposed methods. Full article

Deep learning methods based on convolutional neural networks (CNNs) and Mamba have advanced medical image segmentation, yet two challenges remain: (1) trade-off in feature extraction, where CNNs capture local details but miss global context, and Mamba captures global dependencies but overlooks fine structures, and (2) limited feature aggregation, as existing methods insufficiently integrate inter-layer common information and delta details, hindering robustness to subtle structures. To address these issues, we propose a hybrid cascade and dual-path adaptive aggregation network (HCDAA-Net). For feature extraction, we design a hybrid cascade structure (HCS) that alternately applies ResNet and Mamba modules, achieving a spatial balance between local detail preservation and global semantic modeling. We further employ a general channel-crossing attention mechanism to enhance feature expression, complementing this spatial modeling and accelerating convergence. For feature aggregation, we first propose correlation-aware aggregation (CAA) to model correlations among features of the same lesions or anatomical structures. Second, we develop a dual-path adaptive feature aggregation (DAFA) module: the common path captures stable cross-layer semantics and suppresses redundancy, while the delta path emphasizes subtle differences to strengthen the model’s sensitivity to fine details. Finally, we introduce a residual-gated visual state space module (RG-VSS), which dynamically modulates information flow via a convolution-enhanced residual gating mechanism to refine fused representations. Experiments on diverse datasets demonstrate that our HCDAA-Net outperforms some state-of-the-art (SOTA) approaches. Full article

With the increasing demand for indoor location-based services, magnetic-fingerprint-based positioning has emerged as a promising complementary solution in scenarios lacking WiFi coverage. However, the dynamic nature of indoor environments, architectural complexity, and variations in pedestrian walking speeds can lead to stretching, compression, and distortion of magnetic fingerprint sequences, making it challenging for traditional sequence-matching algorithms to maintain stable positioning performance. To address these challenges, this paper proposes a magnetic-fingerprint-based positioning model that integrates residual networks (ResNet), transformer, and LSTM, referred to as Res-T-LSTM. Within the overall architecture, the ResNet module extracts deep local spatial features of magnetic fingerprints, and its residual connections effectively mitigate gradient attenuation during deep network training. The transformer module leverages self-attention mechanisms to model long-range dependencies and global contextual information, adaptively emphasizing key magnetic variations to enhance the discriminability of the feature representations. The LSTM module further captures the dynamic temporal evolution of magnetic sequences, improving robustness to variations in walking speed and sequence stretching or compression. Experimental results show that the proposed model achieves excellent performance across four smartphone-carrying postures, yielding an average positioning error of 0.21 m. Full article

In controlled environment agriculture (CEA), CO₂ enrichment can promote photosynthesis while simultaneously reducing evapotranspiration, but the optimal settings vary depending on crop type, growth stage, and microclimate. This study presents a near-field remote sensing framework that fuses RGB image features with environmental variables to predict the CO₂ uptake/respiration dynamics of five leafy vegetables grown in a hydroponic culture system and evaluate their impact on resource efficiency under CO₂ control. A hybrid deep model incorporating You Only Look Once version 11 (YOLOv11) and a Residual Network with 50 layers (ResNet50) extracts growth-related visual cues and integrates them with tabular features (CO₂, temperature, and light conditions) to predict chamber CO₂ dynamics. Performance was evaluated by Mean Absolute Error (MAE)/Mean Squared Error (MSE) on withheld data, and the system-level impacts on water use (ET), pumping energy, and relative yield were analyzed using a conventional greenhouse model. The model exhibited high accuracy (MAE = 0.95; MSE = 1.62). Scenario analysis results showed that increasing ambient CO₂ concentration from 400 to 1200 ppm reduced modeled water demand by approximately 11%, increased modeled yield by approximately 9%, and resulted in a corresponding reduction in pumping energy per unit area. Unlike conventional single-crop, table-based approaches, this study demonstrates multi-crop generalization and image-environment fusion for CO₂ dynamic prediction, establishing proximity sensing as a viable decision-making layer for CEA. While yield/ET results were simulated rather than measured in long-term trials, and leaf area normalization was not available, the proposed framework provides a viable path for data-driven CO₂ control in indoor farms by linking image-based monitoring with operational optimization. Full article

This paper provides an overview of artificial intelligence (AI) applications in ophthalmology, with a focus on diagnosing dry eye disease (DED). We aim to synthesize studies that explicitly compare AI-based diagnostic models with clinical tests employed by ophthalmologists, examine results obtained using similar imaging modalities, and identify recurring limitations to propose recommendations for future work. We conducted a systematic literature search following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines across four databases: Google Scholar, PubMed, ScienceDirect, and the Cochrane Library. We targeted studies published between 2020 and 2025 and applied predefined inclusion criteria to select 30 original peer-reviewed articles. We then analyzed each study based on the AI models used, development strategies, diagnostic performance, correlation with clinical parameters, and reported limitations. The imaging modalities covered include videokeratography, smartphone-based imaging, tear film interferometry, anterior segment optical coherence tomography, infrared meibography, in vivo confocal microscopy, and slit-lamp photography. Across modalities, deep learning models (e.g., U-shaped Convolutional Network (U-Net), Residual Network (ResNet), Densely Connected Convolutional Network (DenseNet), Generative Adversarial Networks (GANs), transformers) demonstrated promising performance, often matching or surpassing clinical assessments, with reported accuracies ranging from 82% to 99%. However, few studies performed external validations or addressed inter-expert variability. The findings confirm AI’s potential in DED diagnosis, but emphasize gaps in data diversity, clinical use, and reproducibility. It offers practical recommendations for future research to bridge these gaps and support AI deployment in routine eye care. Full article

Rice diseases pose a critical threat to global food security. While deep learning offers a promising path toward automated diagnosis, clear guidelines for model selection in resource-constrained agricultural environments are still lacking. This study presents a systematic benchmark of five deep convolutional neural networks (CNNs)—Visual Geometry Group (VGG)16, VGG19, Residual Network (ResNet)101V2, Xception, and Densely Connected Convolutional Network (DenseNet)121—for rice disease identification using a public leaf image dataset. The models, initialized with ImageNet pre-trained weights, were rigorously evaluated under a unified framework, including 5-fold cross-validation and a challenging out-of-distribution (OOD) generalization test. Our results demonstrate a clear performance hierarchy, with DenseNet121 emerging as the superior model. It achieved the highest OOD accuracy and F1-score (both 85.08%) while exhibiting the greatest parameter efficiency (8.1 million parameters), making it ideally suited for edge deployment. In contrast, architectures with large fully connected layers (VGG) or less efficient feature learning mechanisms (Xception, ResNet101V2) showed lower performance in this specific task. This study confirms the critical impact of architectural design choices, provides a reproducible performance baseline, and identifies DenseNet121 as a robust, efficient, and highly recommendable CNN for practical rice disease diagnosis in precision agriculture. Full article

Precise classification of brain tumors is crucial for early diagnosis and treatment, but obtaining tumor masks is extremely challenging, limiting the application of traditional methods. This paper proposes a brain tumor classification model based on whole-brain images, combining a spatial block–residual block cooperative architecture with striped pooling feature fusion to achieve multi-scale feature representation without requiring tumor masks. The model extracts fine-grained morphological features through three shallow VGG spatial blocks while capturing global contextual information between tumors and surrounding tissues via four deep ResNet residual blocks. Residual connections mitigate gradient vanishing. To effectively fuse multi-level features, strip pooling modules are introduced after the third spatial block and fourth residual block, enabling cross-layer feature integration—particularly optimizing representation of irregular tumor regions. The fused features undergo cross-scale concatenation, integrating both spatial perception and semantic information, and are ultimately classified via an end-to-end Softmax classifier. Experimental results demonstrate that the model achieves an accuracy of 97.29% in brain tumor image classification tasks, significantly outperforming traditional convolutional neural networks. This validates its effectiveness in achieving high-precision, multi-scale feature learning and classification without brain tumor masks, holding potential clinical application value. Full article

Additive manufacturing has emerged as one of the revolutionary technologies of today, enabling quick prototyping, customized production, and reduced material waste. However, its reliability is often weakened due to faults arising during printing, which remain undetected and, thus, give rise to product defects, waste generation, and safety issues. Most of the existing fault detection methods suffer from limited accuracy, poor adaptability within different printing conditions, and a lack of real-time monitoring capability. These factors critically limit their effectiveness in practical deployment. To address these limitations, the current study proposes a novel process control approach for additive manufacturing with the integration of advanced segmentation, detection, and monitoring strategies. The implemented framework involves segmentation of layer regions using MaskLab-CRFNet, integrating Mask R-CNN, DeepLabv3, and Conditional Random Fields for precise defect location; detection is performed by MoShuResNet, hybridizing MobileNetV3, ShuffleNet, and Residual U-Net for lightweight yet robust fault classification; and monitoring is done by BLC-MonitorNet, which incorporates Bayesian deep networks, ConvAE-LSTM, and convolutional autoencoders together for reliable real-time anomaly detection. Experimental evaluation demonstrates superior performance, with the achievement of 99.31% accuracy and 97.73% sensitivity. This work presents a reliable and interpretable process control framework for additive manufacturing that will improve safety, efficiency, and sustainability. Full article

Internal combustion engines (ICEs) are prone to faults such as abnormal injection pressure and valve clearance, but traditional diagnosis methods struggle with feature extraction and require large data volumes, limiting real-time applications. Deep learning approaches like CNN and LSTM have improved accuracy but often fail to capture both time and frequency domain features efficiently. This study proposes a Time–Frequency Domain Diagnosis Network (TFDN) that integrates a time-domain path (using residual networks and self-attention mechanisms for sequential feature extraction) and a frequency-domain path (using CNNs for spatial feature extraction). The model employs Swish activation functions and batch normalization to enhance training efficiency. Validated on a six-cylinder diesel engine with 12 fault types, TFDN achieved an accuracy of 98.12%~99.79% in full-load conditions, outperforming baselines like CNN, ResNet, and LSTM. Under mixed operating conditions, TFDN maintained high accuracy, precision, and recall, and demonstrated robustness with limited data (60%~70% accuracy at 5 samples per fault). TFDN effectively combines time-frequency features to improve diagnostic accuracy and stability, enabling real-time fault detection with reduced data dependency. It offers a practical solution for ICE condition monitoring. Full article