Deep Learning Approach Research Articles

Enhancing SNR in CEST imaging: A deep learning approach with a denoising convolutional autoencoder.

To develop a SNR enhancement method for CEST imaging using a denoising convolutional autoencoder (DCAE) and compare its performance with state-of-the-art denoising methods. The DCAE-CEST model encompasses an encoder and a decoder network. The encoder learns features from the input CEST Z-spectrum via a series of one-dimensional convolutions, nonlinearity applications, and pooling. Subsequently, the decoder reconstructs an output denoised Z-spectrum using a series of up-sampling and convolution layers. The DCAE-CEST model underwent multistage training in an environment constrained by Kullback-Leibler divergence, while ensuring data adaptability through context learning using Principal Component Analysis-processed Z-spectrum as a reference. The model was trained using simulated Z-spectra, and its performance was evaluated using both simulated data and in vivo data from an animal tumor model. Maps of amide proton transfer (APT) and nuclear Overhauser enhancement (NOE) effects were quantified using the multiple-pool Lorentzian fit, along with an apparent exchange-dependent relaxation metric. In digital phantom experiments, the DCAE-CEST method exhibited superior performance, surpassing existing denoising techniques, as indicated by the peak SNR and Structural Similarity Index. Additionally, in vivo data further confirm the effectiveness of the DCAE-CEST in denoising the APT and NOE maps when compared with other methods. Although no significant difference was observed in APT between tumors and normal tissues, there was a significant difference in NOE, consistent with previous findings. The DCAE-CEST can learn the most important features of the CEST Z-spectrum and provide the most effective denoising solution compared with other methods.

Fault detection in single-hop and multi-hop wireless sensor networks using a deep learning algorithm

<p>The wireless sensor network (WSN) has received significant recognition for its positive impact on environmental monitoring, yet its reliability remains prone to faults. Common factors contributing to faults include connectivity loss from malfunctioning node interfaces, disruptions caused by obstacles, and increased packet loss due to noise or congestion. This research employs a variety of machine learning and deep learning techniques to identify and address these faults, aiming to enhance the overall lifespan and scalability of the WSN. Classification models such as support vector machine (SVM), gradient boosting clasifer (GBC), K-nearest neighbours (KNN), random forest, and decision tree were employed in model training, with the decision tree emerging as the most accurate at 90.23%. Additionally, a deep learning approach, the recurrent neural network (RNN), effectively identified faults in sensor nodes, achieving an accuracy of 93.19%.</p>

Artificial Intelligence Application for Anti-tumor Drug Synergy Prediction.

Currently, the main therapeutic methods for cancer include surgery, radiation therapy, and chemotherapy. However, chemotherapy still plays an important role in tumor therapy. Due to the variety of pathogenic factors, the development process of tumors is complex and regulated by many factors, and the treatment of a single drug is easy to cause the human body to produce a drug-resistant phenotype to specific drugs and eventually leads to treatment failure. In the process of clinical tumor treatment, the combination of multiple drugs can produce stronger anti-tumor effects by regulating multiple mechanisms and can reduce the problem of tumor drug resistance while reducing the toxic side effects of drugs. Therefore, it is still a great challenge to construct an efficient and accurate screening method that can systematically consider the synergistic anti- tumor effects of multiple drugs. However, anti-tumor drug synergy prediction is of importance in improving cancer treatment outcomes. However, identifying effective drug combinations remains a complex and challenging task. This review provides a comprehensive overview of cancer drug synergy therapy and the application of artificial intelligence (AI) techniques in cancer drug synergy prediction. In addition, we discuss the challenges and perspectives associated with deep learning approaches. In conclusion, the review of the AI techniques' application in cancer drug synergy prediction can further advance our understanding of cancer drug synergy and provide more effective treatment plans and reasonable drug use strategies for clinical guidance.

Accurate and robust ammonia level forecasting of aeration tanks using long short-term memory ensembles: A comparative study of Adaboost and Bagging approaches

As wastewater treatment aeration systems are embracing innovative solutions to data management for operational sustainability, deep learning approaches like long short-term memory (LSTM) networks become imperative. However, how to enhance LSTMs to forecast aeration status through ensemble learning is still in its infancy. This study tackles this challenge by comprehensively comparing two ensemble learning algorithms, AdaBoost and Bagging. Both one-step and multi-step predictions were compared using performance metrics like Z-score derived from aeration set-points. The robustness of models was evaluated under quantified extreme events, such as sudden spikes in ammonia concentration. The results indicate that while AdaBoost-LSTM models slightly outperformed Bagging-LSTM models in one-step-ahead predictions, their true advantage lies in enabling precise decisions for switching aeration blowers on or off, which could avoid excess energy usage of aeration systems. This advantage was even more pronounced in multi-step forecasting. In 4-step-ahead prediction, the AdaBoost-LSTM model attained an optimal precision of 92.77%, marking an 8.88% improvement over the Bagging-LSTM model. Furthermore, AdaBoost-LSTM models showed greater resilience to fluctuations in ammonia levels, ensuring continued stable aeration. Therefore, AdaBoost-LSTM ensembles demonstrate greater suitability for accurate and robust ammonia forecasting of aeration tanks, leading to sustainable operation and target costs/energy savings.

Evaluating Retinal Blood Vessels for Predicting White Matter Hyperintensities in Ischemic Stroke: A Deep Learning Approach

This study aims to investigate whether a deep learning approach incorporating retinal blood vessels can effectively identify ischemic stroke patients with a high burden of White Matter Hyperintensities (WMH) using Nuclear Magnetic Resonance Imaging (MRI) as the gold standard. In this cross-sectional study, we evaluated 263 ischemic stroke inpatients who had acquired both retinal fundus images and MRI images. The primary outcome was the diagnostic WMH on MRI brain, defined as different degrees of the age-related white matter changes (ARWMC) grade (<2 or ≥2). We developed a deep-learning network model with retinal fundus images to estimate WMH. The mean age of the patient cohort was 60.8 years, with 196 individuals (74.5%) being male. The prevalence of risk factors was as follows: hypertension in 237 (90.1%), diabetes in 109 (41.4%), hyperlipidemias in 84 (31.9%), coronary heart disease in 37 (14.1%), hyperhomocysteinemia in 70 (26.6%), and hyperuricemia in 73 (27.8%). Severe WMH defined as global ARWMC grade ≥2 was found in 139 (52.9%) participants. Using binocular fundus images, we achieved an F1 score of 0.811 and a Macro Accuracy of 0.811 in the ARWMC classification task. Additionally, we conducted experiments by progressively occluding fundus images to assess the relationship between different areas of the fundus images and ARWMC prediction. Our study presents a novel deep learning model designed to detect a high burden of WMH using binocular fundus images in ischemic stroke patients. We have conducted initial investigations into the predictive significance of various fundus image areas for WMH identification. These findings underscore the need for broader data collection, further model training, and prospective data validation.

Federated inception-multi-head attention models for cyber-attacks detection

&lt;span lang="EN-US"&gt;With the proliferation of internet of things (IoT) devices, ensuring the security of these interconnected systems has become a critical concern. Cyberattacks targeting IoT devices pose significant threats to individuals and organizations due to the generation of vast amounts of data across many connected devices, which traditional centralized methods cannot solve. Federated learning (FL) could be a promising solution to mitigate privacy concerns associated with centralized approaches and address cybersecurity concerns. This paper uses FL and deep learning (DL) approaches to cybersecurity in IoT applications. The goal of cyber security is achieved by forming a federation of acquired and shared models at the head of the various participants. We use inception time and multi-head attention (CNN) algorithm based on FL to detect cyber-attacks and avoid data privacy leaks under two distribution modes, namely IID and Non-IID. In contrast, the FedAvg and FedMA algorithms aggregate local model updates. A global model is produced after several communication rounds between the IoT devices and the model parameter server. Cyber threats are simulated using edge-IIoT datasets. Experiment results show that the federated inception model's best global accuracy was 93, 91%, and 93, 49% using multi-head attention.&lt;/span&gt;

Underwater Multi-Target Node Path Planning in Hybrid Action Space: A Deep Reinforcement Learning Approach

Underwater Multi-Target Node Path Planning in Hybrid Action Space: A Deep Reinforcement Learning Approach

Improving laryngeal cancer detection using chaotic metaheuristics integration with squeeze-and-excitation resnet model.

Laryngeal cancer (LC) represents a substantial world health problem, with diminished survival rates attributed to late-stage diagnoses. Correct treatment for LC is complex, particularly in the final stages. This kind of cancer is a complex malignancy inside the head and neck region of patients. Recently, researchers serving medical consultants to recognize LC efficiently develop different analysis methods and tools. However, these existing tools and techniques have various problems regarding performance constraints, like lesser accuracy in detecting LC at the early stages, additional computational complexity, and colossal time utilization in patient screening. Deep learning (DL) approaches have been established that are effective in the recognition of LC. Therefore, this study develops an efficient LC Detection using the Chaotic Metaheuristics Integration with the DL (LCD-CMDL) technique. The LCD-CMDL technique mainly focuses on detecting and classifying LC utilizing throat region images. In the LCD-CMDL technique, the contrast enhancement process uses the CLAHE approach. For feature extraction, the LCD-CMDL technique applies the Squeeze-and-Excitation ResNet (SE-ResNet) model to learn the complex and intrinsic features from the image preprocessing. Moreover, the hyperparameter tuning of the SE-ResNet approach is performed using a chaotic adaptive sparrow search algorithm (CSSA). Finally, the extreme learning machine (ELM) model was applied to detect and classify the LC. The performance evaluation of the LCD-CMDL approach occurs utilizing a benchmark throat region image database. The experimental values implied the superior performance of the LCD-CMDL approach over recent state-of-the-art approaches.

Dependency-Aware Microservice Deployment for Edge Computing: A Deep Reinforcement Learning Approach With Network Representation

Dependency-Aware Microservice Deployment for Edge Computing: A Deep Reinforcement Learning Approach With Network Representation

Deep Learning Based Method for Dynamic Tracking of Bubbles in Microfluidic Gas-Driven Water Experiments

Using microfluidic chips for gas displacement experiments is an important approach for studying CO2 displacement mechanisms and gas-liquid exchange processes. Analyzing the micro-scale distribution of gas and water during gas displacement enables the optimization of injection strategies, offering valuable guidance for enhancing oil and gas recovery rates. Traditional image processing techniques often face challenges in effectively handle the complex gas-liquid flow dynamics and irregular bubble shapes encountered in gas displacement experiments. To address this challenge, this paper proposes a gas bubble detection and tracking method for gas-driven water microfluidic experiments based on domain adaptation in deep learning and improved YOLOv8. Using the CycleGAN style transfer network to generate numerous simulated bubble training samples significantly reduces the manual labeling workload for bubbles. Additionally, the SE attention mechanism is integrated into the YOLOv8 backbone network to enhance the feature extraction capability of bubbles, and the Fast-PConv Head structure is incorporated into the detection head to enhance detection efficiency. Experimental results demonstrate that using just 20 annotated experimental images, the bubble detection accuracy of the proposed method can reach 94%. Lastly, DeepSORT is employed to track bubbles and visualize the dynamic trends of bubble changes in the form of visual charts. The deep learning approach proposed in this paper provides a new perspective for intelligent analysis of bubbles in microfluidic chip experiments, and future work aiming to its application to the complex multiphase flow field in porous media.

SSR-2D: Semantic 3D Scene Reconstruction From 2D Images.

Most deep learning approaches to comprehensive semantic modeling of 3D indoor spaces require costly dense annotations in the 3D domain. In this work, we explore a central 3D scene modeling task, namely, semantic scene reconstruction without using any 3D annotations. The key idea of our approach is to design a trainable model that employs both incomplete 3D reconstructions and their corresponding source RGB-D images, fusing cross-domain features into volumetric embeddings to predict complete 3D geometry, color, and semantics with only 2D labeling which can be either manual or machine-generated. Our key technical innovation is to leverage differentiable rendering of color and semantics to bridge 2D observations and unknown 3D space, using the observed RGB images and 2D semantics as supervision, respectively. We additionally develop a learning pipeline and corresponding method to enable learning from imperfect predicted 2D labels, which could be additionally acquired by synthesizing in an augmented set of virtual training views complementing the original real captures, enabling more efficient self-supervision loop for semantics. As a result, our end-to-end trainable solution jointly addresses geometry completion, colorization, and semantic mapping from limited RGB-D images, without relying on any 3D ground-truth information. Our method achieves state-of-the-art performance of semantic scene completion on two large-scale benchmark datasets MatterPort3D and ScanNet, surpasses baselines even with costly 3D annotations in predicting both geometry and semantics. To our knowledge, our method is also the first 2D-driven method addressing completion and semantic segmentation of real-world 3D scans simultaneously.

Deep learning utilization in Sundanese script recognition for cultural preservation

This study addresses the challenge of preserving the Sundanese script, a traditional writing system of the Sundanese community in Indonesia, which is at risk of being forgotten due to technological advancements. To tackle this problem, we propose a deep learning approach using the YOLOv8 model for the automatic recognition of Sundanese characters. Our methodology includes creating a comprehensive dataset, applying augmentation techniques, and annotating the characters. The trained model achieved a precision of 95% after 150 epochs, demonstrating its effectiveness in recognizing Sundanese characters. While some variability in accuracy was observed for certain characters and real-time applications, the results indicate the feasibility and promise of using deep learning for Sundanese script recognition. This research highlights the potential of technological solutions to digitize and preserve the Sundanese script, ensuring its continued legacy and accessibility for future generations. Thus, we contribute to cultural preservation by providing a method to safeguard the Sundanese script against obsolescence.

Deep learning approach for forensic facial reconstruction depends on unidentified skull

Facial reconstruction, or facial approximation, is an essential problem in a criminal investigation involving reconstructing a victim's face from his skull to determine the victim's identification at a crime scene. Facial approximation plays a crucial part when there is a lack of clues with investigators. Investigators utilize facial approximation to guess the victims' identities. This research attempted to use computer-aided face reconstruction rather than traditional approaches. Traditional methods of face reconstruction include the use of clay or gypsum. Traditional procedures necessitate forensic professionals to rebuild the victim's face. This research uses the convolution neural network skull part with sift (CNNSPS) model is employed to reconstruct facial features from a skull image utilizing public datasets CelebAMask-HQ and MUG500+. The proposed algorithm was tested on unidentified skull databases, and celebrity faces were used. The genuine datasets are not available, which is the key issue in this research.

Advanced drone-based weed detection using feature-enriched deep learning approach

This research addresses the pressing challenge of weed identification in agriculture, crucial for ensuring food security in anticipation of a global population exceeding 9.7 billion by 2050. Utilizing drone imagery, we collected a dataset and proposed a customized model to achieve optimal performance. Our proposed model uses strategically modified backbone, neck, and head components, leveraging elements such as Ghost Convolution, BottleNeckCSP, and ECA (Efficient Channel Attention) layers. These modifications enhance the model’s capability to discern intricate patterns in drone imagery, ultimately leading to improved precision in weed detection. We introduce a purposefully crafted dataset to complement the model’s training, and our experiments demonstrate superior performance compared to the baseline models. Our model achieves a precision of 72.5%, recall of 68.0%, and mAP@0.5 of 73.9, showcasing the effectiveness of our approach in enhancing detection accuracy. Leveraging a unique blend of feature extraction mechanisms, our model achieves remarkable accuracy in real-time soybean detection, outperforming established models like RT-DETR (Real-Time DEtection TransfoRmer) and YOLOv10. A detailed ablation study and comparative analysis with different YOLO versions and the transformer-based RT-DETR showcase the effectiveness of the proposed enhancements. Our work signifies a significant step toward advancing the field of precision agriculture, offering a model that is not only adaptive but also robust in identifying and localizing weeds in soybean fields.

Deep Learning Approaches for Similarity Computation: A Survey

Deep Learning Approaches for Similarity Computation: A Survey

End-to-end deep learning method to reconstruct full-field stress distribution for ship hull structure with stress concentrations

An image-based deep learning framework is developed to reconstruct full-field stress distribution of ship hull structure from discrete measurements in this paper. The work is motivated by the lack of data to assess the strength state of vessel. The proposed deep learning framework is based on a conditional generative adversarial network (cGAN) which contains a generator and a discriminator. They are mutually trained in a supervised manner based on an adversarial mechanism. The hull structure of interest is selected to be an opening deck of a bulk carrier. A novel image encoding algorithm is also designed to fuse global stress distribution and high gradient stresses at hatch corner regions into an image. A total of 3400 stress fields are generated and simulated using finite element method (FEM) to provide a sufficiently large dataset for training and testing the network. It is shown that the proposed deep learning approach can efficiently reconstruct the full-field stress distribution of hull structure. Furtherly, the influence of measurement noise on reconstruction accuracy is discussed. The results demonstrate that the well-trained network can directly generate robust solutions to full-field stress distribution under given discrete measurements.

Advances in Protein-Ligand Binding Affinity Prediction via Deep Learning: A Comprehensive Study of Datasets, Data Preprocessing Techniques, and Model Architectures.

Drug discovery is a complex and expensive procedure involving several timely and costly phases through which new potential pharmaceutical compounds must pass to get approved. One of these critical steps is the identification and optimization of lead compounds, which has been made more accessible by the introduction of computational methods, including deep learning (DL) techniques. Diverse DL model architectures have been put forward to learn the vast landscape of interaction between proteins and ligands and predict their affinity, helping in the identification of lead compounds. This survey fills a gap in previous research by comprehensively analyzing the most commonly used datasets and discussing their quality and limitations. It also offers a comprehensive classification of the most recent DL methods in the context of protein-ligand binding affinity prediction (BAP), providing a fresh perspective on this evolving field. We thoroughly examine commonly used datasets for BAP and their inherent characteristics. Our exploration extends to various preprocessing steps and DL techniques, including graph neural networks, convolutional neural networks, and transformers, which are found in the literature. We conducted extensive literature research to ensure that the most recent deep learning approaches for BAP were included by the time of writing this manuscript. The systematic approach used for the present study highlighted inherent challenges to BAP via DL, such as data quality, model interpretability, and explainability, and proposed considerations for future research directions. We present valuable insights to accelerate the development of more effective and reliable DL models for BAP within the research community. The present study can considerably enhance future research on predicting affinity between protein and ligand molecules, hence further improving the overall drug development process.

Weighted ensemble deep learning approach for classification of gastrointestinal diseases in colonoscopy images Aided by explainable AI

Gastrointestinal diseases are significant health issues worldwide, requiring early diagnosis due to their serious health implications. Therefore, detecting these diseases using artificial intelligence-based medical decision support systems through colonoscopy images plays a critical role in early diagnosis. In this study, a deep learning-based method is proposed for the classification of gastrointestinal diseases and colon anatomical landmarks using colonoscopy images. For this purpose, five different Convolutional Neural Network (CNN) models, namely Xception, ResNet-101, NASNet-Large, EfficientNet, and NASNet-Mobile, were trained. An ensemble model was created using class-based recall values derived from the validation performances of the top three models (Xception, ResNet-101, NASNet-Large). A user-friendly Graphical User Interface (GUI) was developed, allowing users to perform classification tasks and use Gradient-weighted Class Activation Mapping (Grad-CAM), an explainable AI tool, to visualize the regions from which the model derives information. Grad-CAM visualizations contribute to a better understanding of the model’s decision-making processes and play an important role in the application of explainable AI. In the study, eight labels, including anatomical markers such as z-line, pylorus, and cecum, as well as pathological findings like esophagitis, polyps, and ulcerative colitis, were classified using the KVASIR V2 dataset. The proposed ensemble model achieved a 94.125% accuracy on the KVASIR V2 dataset, demonstrating competitive performance compared to similar studies in the literature. Additionally, the precision and F1 score values ​​of this model are equal to 94.168% and 94.125%, respectively. These results suggest that the proposed method provides an effective solution for the diagnosis of GI diseases and can be beneficial for medical education.

Artificial Intelligence-Based Classification of CT Images Using a Hybrid SpinalZFNet.

The kidney is an abdominal organ in the human body that supports filtering excess water and waste from the blood. Kidney diseases generally occur due to changes in certain supplements, medical conditions, obesity, and diet, which causes kidney function and ultimately leads to complications such as chronic kidney disease, kidney failure, and other renal disorders. Combining patient metadata with computed tomography (CT) images is essential to accurately and timely diagnosing such complications. Deep Neural Networks (DNNs) have transformed medical fields by providing high accuracy in complex tasks. However, the high computational cost of these models is a significant challenge, particularly in real-time applications. This paper proposed SpinalZFNet, a hybrid deep learning approach that integrates the architectural strengths of Spinal Network (SpinalNet) with the feature extraction capabilities of Zeiler and Fergus Network (ZFNet) to classify kidney disease accurately using CT images. This unique combination enhanced feature analysis, significantly improving classification accuracy while reducing the computational overhead. At first, the acquired CT images are pre-processed using a median filter, and the pre-processed image is segmented using Efficient Neural Network (ENet). Later, the images are augmented, and different features are extracted from the augmented CT images. The extracted features finally classify the kidney disease into normal, tumor, cyst, and stone using the proposed SpinalZFNet model. The SpinalZFNet outperformed other models, with 99.9% sensitivity, 99.5% specificity, precision 99.6%, 99.8% accuracy, and 99.7% F1-Score in classifying kidney disease.

Utilizing deep learning algorithms for the resolution of partial differential equations

Partial differential equations (PDEs) are mathematical equations that are used to model physical phenomena around us, such as fluid dynamics, electrodynamics, general relativity, electrostatics, and diffusion. However, solving these equations can be challenging due to the problem known as the dimensionality curse, which makes classical numerical methods less effective. To solve this problem, we propose a deep learning approach called deep Galerkin algorithm (DGA). This technique involves training a neural network to approximate a solution by satisfying the difference operator, boundary conditions and an initial condition. DGA alleviates the curse of dimensionality through deep learning, a meshless approach, residue-based loss minimisation and efficient use of data. We will test this approach for the transport equation, the wave equation, the Sine-Gordon equation and the Klein-Gordon equation.