Search Results (47)

Long scan times remain a fundamental challenge in Magnetic Resonance Imaging (MRI). Accelerated MRI, which undersamples k-space, requires robust reconstruction methods to solve the ill-posed inverse problem. Recent methods have shown promise by processing image-domain features to capture global spatial context. However, these approaches are often limited, as they fail to fully leverage the unique, sequential characteristics of the k-space data themselves, which are critical for disentangling aliasing artifacts. This study introduces a novel, hybrid, dual-domain deep learning architecture that combines a ViT-based autoencoder with Bidirectional Recurrent Neural Networks (BiRNNs). The proposed architecture is designed to synergistically process information from both domains: it uses the ViT to learn features from image patches and the BiRNNs to model sequential dependencies directly from k-space data. We conducted a comprehensive comparative analysis against a standard ViT with only an MLP head (Model 1), a ViT autoencoder operating solely in the image domain (Model 2), and a competitive UNet baseline. Evaluations were performed on retrospectively undersampled neuro-MRI data using R = 4 and R = 8 acceleration factors with both regular and random sampling patterns. The proposed architecture demonstrated superior performance and robustness, significantly outperforming all other models in challenging high-acceleration and random-sampling scenarios. The results confirm that integrating sequential k-space processing via BiRNNs is critical for superior artifact suppression, offering a robust solution for accelerated MRI. Full article

Human Activity Recognition (HAR) is widely used for healthcare, but few works focus on Manual Material Handling (MMH) activities, despite their diffusion and impact on the workers’ health. We propose four Deep Learning algorithms for HAR in MMH: Bidirectional Long Short-Term Memory (BiLSTM), Sparse Denoising Autoencoder (Sp-DAE), Recurrent Sp-DAE, and Recurrent Convolutional Neural Network (RCNN). We explored different hyperparameter combinations to maximize the classification performance (F1-score,) using wearable sensors’ data gathered from 14 subjects. We investigated the best three-parameter combinations for each network using the full dataset to select the two best-performing networks, which were then compared using 14 datasets with increasing subject numerosity, 70–30% split, and Leave-One-Subject-Out (LOSO) validation, to evaluate whether they may perform better with a larger dataset. The benchmarking network DeepConvLSTM was tested on the full dataset. BiLSTM performs best in classification and complexity (95.7% 70–30% split; 90.3% LOSO). RCNN performed similarly (95.9%; 89.2%) with a positive trend with subject numerosity. DeepConvLSTM achieves similar classification performance (95.2%; 90.3%) but requires $\times 57.1$ and $\times 31.3$ more Multiply and ACcumulate (MAC) and $\times 100.8$ and $\times 28.3$ more Multiplication and Addition (MA) operations, which measure the complexity of the network’s inference process, than BiLSTM and RCNN, respectively. The BILSTM and RCNN perform close to DeepConvLSTM while being computationally lighter, fostering their use in embedded systems. Such lighter algorithms can be readily used in the automatic ergonomic and biomechanical risk assessment systems, enabling personalization of risk assessment and easing the adoption of safety measures in industrial practices involving MMH. Full article

As port operations rapidly evolve toward intelligent and heavy-duty applications, fault diagnosis for core equipment demands higher levels of real-time performance and robustness. Deep learning, with its powerful autonomous feature learning capabilities, demonstrates significant potential in mechanical fault prediction and health management. This paper first provides a systematic review of deep learning research advances in rotating machinery fault diagnosis over the past eight years, focusing on the technical approaches and application cases of four representative models: Deep Belief Networks (DBNs), Convolutional Neural Networks (CNNs), Auto-encoders (AEs), and Recurrent Neural Networks (RNNs). These models, respectively, embody four core paradigms, unsupervised feature generation, spatial pattern extraction, data reconstruction learning, and temporal dependency modeling, forming the technological foundation of contemporary intelligent diagnostics. Building upon this foundation, this paper delves into the unique challenges encountered when transferring these methods from generic laboratory components to specialized port equipment such as shore cranes and yard cranes—including complex operating conditions, harsh environments, and system coupling. It further explores future research directions, including cross-condition transfer, multi-source information fusion, and lightweight deployment, aiming to provide theoretical references and implementation pathways for the technological advancement of intelligent operation and maintenance in port equipment. Full article

The increasing demands for the reliability of modern industrial equipment and structures necessitate advanced techniques for design, monitoring, and analysis. This review article presents the latest research advancements in the application of machine learning techniques to vibration and acoustic signal analysis from 2015 to 2024. A total of 96 peer-reviewed scientific publications were examined, selected using a systematic Scopus-based search. The main research areas include processes such as modeling and design, health management, condition monitoring, non-destructive testing, damage detection, and diagnostics. In the context of these processes, a review of machine learning techniques was conducted, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), long short-term memory (LSTM), autoencoders, support vector machines (SVMs), decision trees (DTs), nearest neighbor search (NNS), K-means clustering, and random forests. These techniques were applied across a wide range of engineering domains, including civil infrastructure, transportation systems, energy installations, and rotating machinery. Additionally, this article analyzes contributions from different countries, highlighting temporal and methodological trends in this field. The findings indicate a clear shift towards deep learning-based methods and multisensor data fusion, accompanied by increasing use of automatic feature extraction and interest in transfer learning, few-shot learning, and unsupervised approaches. This review aims to provide a comprehensive understanding of the current state and future directions of machine learning applications in vibration and acoustics, outlining the field’s evolution and identifying its key research challenges and innovation trajectories. Full article

During crises, people use X to share real-time updates. These posts reveal public sentiment and evolving emergency situations. However, the changing sentiment in tweets coupled with anomalous patterns may indicate significant events, misinformation or emerging hazards that require timely detection. By using a neural network, and employing deep learning techniques for crisis observation, this study proposes a pipeline for sentiment analysis and anomaly detection in crisis-related tweets. The authors used pre-trained BERT to classify tweet sentiment. For sentiment anomaly detection, autoencoders and recurrent neural networks (RNNs) with an attention mechanism were applied to capture sequential relationships and identify irregular sentiment patterns that deviate from standard crisis talk. Experimental results show that neural networks are more accurate than traditional machine learning methods for both sentiment categorization and anomaly detection tasks, with higher precision and recall for identifying sentiment shifts in the public. This study indicates that neural networks can be used for crisis management and the early detection of significant sentiment anomalies. This could be beneficial to emergency responders and policymakers and support data-driven decisions. Full article

Video anomaly detection plays a crucial role in intelligent transportation systems by enhancing urban mobility and safety. This review provides a comprehensive analysis of recent advancements in artificial intelligence methods applied to traffic anomaly detection, including convolutional and recurrent neural networks (CNNs and RNNs), autoencoders, Transformers, generative adversarial networks (GANs), and multimodal large language models (MLLMs). We compare their performance across real-world applications, highlighting patterns such as the superiority of Transformer-based models in temporal context understanding and the growing use of multimodal inputs for robust detection. Key challenges identified include dependence on large labeled datasets, high computational costs, and limited model interpretability. The review outlines how recent research is addressing these issues through semi-supervised learning, model compression techniques, and explainable AI. We conclude with future directions focusing on scalable, real-time, and interpretable solutions for practical deployment. Full article

Various uncertainties such as communication delay, packet loss and disconnection in the Industrial Internet, as well as the asynchronous sampling of sensors, can cause irregularity, sparsity, and misalignment of sampling sequences, and thereby seriously affect the training and prediction performance of a digital twin model. Sequence reconstruction is an effective way to deal with the above problems, but if the measurement data become sparse or contain significant noise due to packet loss and electromagnetic interference, existing methods struggle to achieve ideal results. Therefore, a novel variational autoencoder model based on a parallel reference network and neural controlled differential equation (PRN-NCDE) is proposed in this article to solve the problem of reconstructing irregular series under sparse measurements and high noise levels. First, a multi-channel self-attention module is established, which can not only analyze the position and feature information of the sampled data to improve the reconstruction accuracy under sparse measurements, but also effectively tackle the misalignment and irregularity of the observation sequence through multi-channel and mask mechanisms. Second, to improve the accuracy of sequence reconstruction under large noise levels, a PRN is established to obtain reference features, which are weighted and fused with the features of observed data. Third, we use the NCDE to construct a decoder that can combine the control input of the system to predict the output values to solve the problem of sequence reconstruction in a controlled system. Finally, a weighted loss function is constructed to better train the network parameters of the model. This article takes the furnace of the boiler system in a coal-fired power plant as the test object to verify the effectiveness and fitting accuracy of the proposed PRN-NCDE model compared to the existing methods for a controlled system under sparse measurements and large noise levels. Simulation results show that the proposed PRN-NCDE model can improve the estimation accuracy by more than 50% and 70% compared with the recurrent neural network-NCDE (RNN-NCDE) under different sampling numbers and noise levels, and by more than 80% and 60% compared with the recurrent neural network-NODE (RNN-NODE). Full article

Deep learning has revolutionised medical image analysis, offering the possibility of automated, efficient, and highly accurate diagnostic solutions. This article explores recent developments in deep learning techniques applied to medical imaging, including convolutional neural networks (CNNs) for classification and segmentation, recurrent neural networks (RNNs) for temporal analysis, autoencoders for feature extraction, and generative adversarial networks (GANs) for image synthesis and augmentation. Additionally, U-Net models for segmentation, vision transformers (ViTs) for global feature extraction, and hybrid models integrating multiple architectures are explored. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) process were used, and searches on PubMed, Google Scholar, and Scopus databases were conducted. The findings highlight key challenges such as data availability, interpretability, overfitting, and computational requirements. While deep learning has demonstrated significant potential in enhancing diagnostic accuracy across multiple medical imaging modalities—including MRI, CT, US, and X-ray—factors such as model trust, data privacy, and ethical considerations remain ongoing concerns. The study underscores the importance of integrating multimodal data, improving computational efficiency, and advancing explainability to facilitate broader clinical adoption. Future research directions emphasize optimising deep learning models for real-time applications, enhancing interpretability, and integrating deep learning with existing healthcare frameworks for improved patient outcomes. Full article

Human activity recognition (HAR) is the task of classifying sequences of data into defined movements. Taking advantage of deep learning (DL) methods, this research investigates and optimizes neural network architectures to effectively classify physical activities from smartphone accelerometer data. Unlike traditional machine learning (ML) methods employing manually crafted features, our approach employs automated feature learning with three deep learning architectures: Convolutional Neural Networks (CNN), CNN-based autoencoders, and Long Short-Term Memory Recurrent Neural Networks (LSTM RNN). The contribution of this work is primarily in optimizing LSTM RNN to leverage the most out of temporal relationships between sensor data, significantly improving classification accuracy. Experimental outcomes for the WISDM dataset show that the proposed LSTM RNN model achieves 96.1% accuracy, outperforming CNN-based approaches and current ML-based methods. Compared to current works, our optimized frameworks achieve up to 6.4% higher classification performance, which means that they are more appropriate for real-time HAR. Full article

This review explores advanced sensing technologies and deep learning (DL) methodologies for monitoring airborne particulate matter (PM), which is critical for environmental health assessments. It begins with discussing the significance of PM monitoring and introduces surface plasmon resonance (SPR) as a promising technique in environmental applications, alongside the role of DL neural networks in enhancing these technologies. This review analyzes advancements in airborne PM sensing technologies and the integration of DL methodologies for environmental monitoring. This review emphasizes the importance of PM monitoring for public health, environmental policy, and scientific research. Traditional PM sensing methods, including their principles, advantages, and limitations, are discussed, covering gravimetric techniques, continuous monitoring, optical and electrical methods, and microscopy. The integration of DL with PM sensing offers potential for enhancing monitoring accuracy, efficiency, and data interpretation. DL techniques, such as convolutional neural networks (CNNs), autoencoders, recurrent neural networks (RNNs), and their variants, are examined for applications like PM estimation from satellite data, air quality prediction, and sensor calibration. This review highlights the data acquisition and quality challenges in developing effective DL models for air quality monitoring. Techniques for handling large and noisy datasets are explored, emphasizing the importance of data quality for model performance, generalizability, and interpretability. The emergence of low-cost sensor technologies and hybrid systems for PM monitoring is discussed, acknowledging their promise while recognizing the need for addressing data quality, standardization, and integration issues. This review identifies areas for future research, including the development of robust DL models, advanced data fusion techniques, applications of deep reinforcement learning, and considerations of ethical implications. Full article

Drug–target affinity (DTA) prediction is a critical aspect of drug discovery. The meaningful representation of drugs and targets is crucial for accurate prediction. Using 1D string-based representations for drugs and targets is a common approach that has demonstrated good results in drug–target affinity prediction. However, these approach lacks information on the relative position of the atoms and bonds. To address this limitation, graph-based representations have been used to some extent. However, solely considering the structural aspect of drugs and targets may be insufficient for accurate DTA prediction. Integrating the functional aspect of these drugs at the genetic level can enhance the prediction capability of the models. To fill this gap, we propose GramSeq-DTA, which integrates chemical perturbation information with the structural information of drugs and targets. We applied a Grammar Variational Autoencoder (GVAE) for drug feature extraction and utilized two different approaches for protein feature extraction as follows: a Convolutional Neural Network (CNN) and a Recurrent Neural Network (RNN). The chemical perturbation data are obtained from the L1000 project, which provides information on the up-regulation and down-regulation of genes caused by selected drugs. This chemical perturbation information is processed, and a compact dataset is prepared, serving as the functional feature set of the drugs. By integrating the drug, gene, and target features in the model, our approach outperforms the current state-of-the-art DTA prediction models when validated on widely used DTA datasets (BindingDB, Davis, and KIBA). This work provides a novel and practical approach to DTA prediction by merging the structural and functional aspects of biological entities, and it encourages further research in multi-modal DTA prediction. Full article

Brain tumor detection is crucial in medical research due to high mortality rates and treatment challenges. Early and accurate diagnosis is vital for improving patient outcomes, however, traditional methods, such as manual Magnetic Resonance Imaging (MRI) analysis, are often time-consuming and error-prone. The rise of deep learning has led to advanced models for automated brain tumor feature extraction, segmentation, and classification. Despite these advancements, comprehensive reviews synthesizing recent findings remain scarce. By analyzing over 100 research papers over past half-decade (2019–2024), this review fills that gap, exploring the latest methods and paradigms, summarizing key concepts, challenges, datasets, and offering insights into future directions for brain tumor detection using deep learning. This review also incorporates an analysis of previous reviews and targets three main aspects: feature extraction, segmentation, and classification. The results revealed that research primarily focuses on Convolutional Neural Networks (CNNs) and their variants, with a strong emphasis on transfer learning using pre-trained models. Other methods, such as Generative Adversarial Networks (GANs) and Autoencoders, are used for feature extraction, while Recurrent Neural Networks (RNNs) are employed for time-sequence modeling. Some models integrate with Internet of Things (IoT) frameworks or federated learning for real-time diagnostics and privacy, often paired with optimization algorithms. However, the adoption of eXplainable AI (XAI) remains limited, despite its importance in building trust in medical diagnostics. Finally, this review outlines future opportunities, focusing on image quality, underexplored deep learning techniques, expanding datasets, and exploring deeper learning representations and model behavior such as recurrent expansion to advance medical imaging diagnostics. Full article

Medical image segmentation plays a critical role in accurate diagnosis and treatment planning, enabling precise analysis across a wide range of clinical tasks. This review begins by offering a comprehensive overview of traditional segmentation techniques, including thresholding, edge-based methods, region-based approaches, clustering, and graph-based segmentation. While these methods are computationally efficient and interpretable, they often face significant challenges when applied to complex, noisy, or variable medical images. The central focus of this review is the transformative impact of deep learning on medical image segmentation. We delve into prominent deep learning architectures such as Convolutional Neural Networks (CNNs), Fully Convolutional Networks (FCNs), U-Net, Recurrent Neural Networks (RNNs), Adversarial Networks (GANs), and Autoencoders (AEs). Each architecture is analyzed in terms of its structural foundation and specific application to medical image segmentation, illustrating how these models have enhanced segmentation accuracy across various clinical contexts. Finally, the review examines the integration of deep learning with traditional segmentation methods, addressing the limitations of both approaches. These hybrid strategies offer improved segmentation performance, particularly in challenging scenarios involving weak edges, noise, or inconsistent intensities. By synthesizing recent advancements, this review provides a detailed resource for researchers and practitioners, offering valuable insights into the current landscape and future directions of medical image segmentation. Full article

To improve the reliability and maintainability of the nuclear safety-class digital control system (DCS), this paper conducts a study on the fault prediction of critical components in the output circuit of the nuclear safety-class signal conditioning module. To address the issue of insufficient feature extraction for the minor offset fault feature and the low accuracy of fault prediction, a predictive model based on stacked denoising autoencoder (SDAE) feature extraction and an improved gated recurrent unit (GRU) is proposed. Therefore, fault simulation modeling is performed for critical components of the signal output circuit to obtain fault datasets of critical components, and the SDAE model is used to extract fault features. The fault prediction model based on GRU is established, and the number of hidden layers, the number of hidden layer nodes, and the learning rate of the GRU model are optimized using the adaptive gray wolf optimization algorithm (AGWO). The prediction performance evaluation metrics include the root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and absolute error (EA), which are used for evaluating the prediction results of models such as the AGWO-GRU model, recurrent neural network (RNN) model, and long short-term memory network (LSTM). The results show that the GRU model optimized by AGWO has a better prediction accuracy (errors range within 0.01%) for the faults of the circuit critical components, and, moreover, can accurately and stably predict the fault trend of the circuit. Full article

The remarkable human ability to predict others’ intent during physical interactions develops at a very early age and is crucial for development. Intent prediction, defined as the simultaneous recognition and generation of human–human interactions, has many applications such as in assistive robotics, human–robot interaction, video and robotic surveillance, and autonomous driving. However, models for solving the problem are scarce. This paper proposes two attention-based agent models to predict the intent of interacting 3D skeletons by sampling them via a sequence of glimpses. The novelty of these agent models is that they are inherently multimodal, consisting of perceptual and proprioceptive pathways. The action (attention) is driven by the agent’s generation error, and not by reinforcement. At each sampling instant, the agent completes the partially observed skeletal motion and infers the interaction class. It learns where and what to sample by minimizing the generation and classification errors. Extensive evaluation of our models is carried out on benchmark datasets and in comparison to a state-of-the-art model for intent prediction, which reveals that classification and generation accuracies of one of the proposed models are comparable to those of the state of the art even though our model contains fewer trainable parameters. The insights gained from our model designs can inform the development of efficient agents, the future of artificial intelligence (AI). Full article