Synthetic Graphic Well Log Generation Using an Enhanced Deep Learning Workflow: Imbalanced Multiclass Data, Sample Size, and Scalability Challenges

SPE Journal, Год журнала: 2023, Номер 29(01), С. 1 - 20

Опубликована: Сен. 7, 2023

Summary The present study introduces an enhanced deep learning (DL) workflow based on transfer (TL) for producing high-resolution synthetic graphic well logs (SGWLs). To examine the scalability of proposed workflow, a carbonate reservoir with high geological heterogeneity has been chosen as case study, and developed is evaluated unseen data (i.e., blind well). Data sources include conventional graphical (GWLs) from neighboring wells. During drilling operations, GWLs are standard practice collecting data. GWL provides rapid visual representation subsurface lithofacies to establish correlations. This investigation examines five wells in southwest Iranian oil field. Due heterogeneities, primary challenge this research lies addressing imbalanced facies distribution. traditional artificial intelligence strategies that manage [e.g., modified minority oversampling technique (M-SMOTE) Tomek link (TKL)] mainly designed solve binary problems. However, adapt these methods upcoming multiclass situation, one-vs.-one (OVO) one-vs.-all (OVA) decomposition ad-hoc techniques used. Well-known VGG16-1D ResNet18-1D used adaptive very-deep algorithms. Additionally, highlight robustness efficiency algorithms, shallow approaches support vector machine (SVM) random forest (RF) classification also other main need enough points train very resolved through TL. After identifying well, four wells’ entered model training. average kappa statistic F-measure, appropriate imbalance evaluation metrics, implemented assess workflows’ performance. numerical comparison analysis shows TL performs better set when combined OVA scheme TKL combat tactic. An 86.33% mean F-measure 92.09% demonstrate superiority. Considering prevalence different distributions, scalable can be efficient productive generating SGWL.

Язык: Английский

---

Analyzing the performance of a bio-sensor integrated improved blended learning model for accurate pneumonia prediction

Results in Engineering, Год журнала: 2024, Номер 22, С. 102063 - 102063

Опубликована: Апрель 2, 2024

Pneumonia has been considered a life-threatening disease for elderly human beings and those with weakened immune systems in the present medical era. The contemporary scenario highlights significance of intelligent automatic handheld devices to detect pneumonia other pulmonary diseases. Hence, this research designed an improved blended learning paradigm (IBLP) real-time detection from chest X-rays, early lung diseases alveolar gas using biosensors graphical processing unit (GPU) developed overcome resolve such challenges. It emphasizes applications techniques, particularly identifying X-ray images exhaled breath support vector machine (SVM). experimental findings indicate that based VGG16 (91.99%) consistently outperforms VGG19 (88.91%) ResNet50 (87.02%) model diagnostic accuracy. IBLP provided 95.5% precision, 97.69% F1 score, 100% recall rate no false-negative results. future classification diagnosis will likely involve artificial intelligence-based can provide accurate timely analysis images, thereby improving patient outcomes reducing healthcare costs.

Язык: Английский

---

EO-CNN: Equilibrium Optimization-Based hyperparameter tuning for enhanced pneumonia and COVID-19 detection using AlexNet and DarkNet19

Journal of Applied Biomedicine, Год журнала: 2024, Номер 44(3), С. 635 - 650

Опубликована: Июль 1, 2024

Язык: Английский

---

Unlocking the Power of 3D Convolutional Neural Networks for COVID-19 Detection: A Comprehensive Review

Deleted Journal, Год журнала: 2025, Номер unknown

Опубликована: Янв. 23, 2025

The advent of three-dimensional convolutional neural networks (3D CNNs) has revolutionized the detection and analysis COVID-19 cases. As imaging technologies have advanced, 3D CNNs emerged as a powerful tool for segmenting classifying in medical images. These demonstrated both high accuracy rapid capabilities, making them crucial effective diagnostics. This study offers thorough review various CNN algorithms, evaluating their efficacy across range modalities. systematically examines recent advancements methodologies. process involved comprehensive screening abstracts titles to ensure relevance, followed by meticulous selection research papers from academic repositories. evaluates these based on specific criteria provides detailed insights into network architectures algorithms used detection. reveals significant trends use segmentation classification. It highlights key findings, including diverse employed compared other diseases, which predominantly utilize encoder/decoder frameworks. an in-depth methods, discussing strengths, limitations, potential areas future research. reviewed total 60 published repositories, Springer Elsevier. this implications clinical diagnosis treatment strategies. Despite some efficiency underscore advancing image findings suggest that could significantly enhance management COVID-19, contributing improved healthcare outcomes.

Язык: Английский

---

Two-Parallel-Step CNN Framework for Detection of COVID-19 Based on Segmented CT-Scan and Chest X-Ray Images

Vietnam Journal of Computer Science, Год журнала: 2025, Номер unknown, С. 1 - 27

Опубликована: Март 28, 2025

COVID-19 is a disease that infects people and quickly isolates the entire world. The new variants of continue to cause high mortality rates. Therefore, many scientists worldwide still are looking for solution accurately detect COVID-19. This paper aims using chest CT-Scan Chest X-ray images. In this work, we design bimodal convolutional neural network (CNN) requires two inputs. first modality image segmented by U-Net deep learning technique infected areas in lung. second input image. proposed CNN combines features extracted from these Feature extraction performed on images parallel feature layers. vectors will be combined perceptron attention mechanism taken as fully connected layers classify patient COVID-19, non-COVID, pneumonia. results have shown newly designed outperforms other similar state-of-the-art methods especially distinguishing between pneumonia cases. has achieved 98.79% classification accuracy 43.20% loss. framework could particularly beneficial telemedicine, enabling remote diagnosis with limited access medical specialists.

Язык: Английский

---

EffiCOVID-Net: A Highly Efficient Convolutional Neural Network for COVID-19 Diagnosis Using Chest X-ray Imaging

Methods, Год журнала: 2025, Номер unknown

Опубликована: Апрель 1, 2025

Язык: Английский

---

Optimized Xception Learning Model and XgBoost Classifier for Detection of Multiclass Chest Disease from X-ray Images

Diagnostics, Год журнала: 2023, Номер 13(15), С. 2583 - 2583

Опубликована: Авг. 3, 2023

Computed tomography (CT) scans, or radiographic images, were used to aid in the early diagnosis of patients and detect normal abnormal lung function human chest. However, lungs infected with coronavirus disease 2019 (COVID-19) was made more accurately from CT scan data than a swab test. This study uses chest radiography pictures identify categorize lungs, opacities, COVID-19-infected viral pneumonia (often called pneumonia). In past, several CAD systems using image processing, ML/DL, other forms machine learning have been developed. those did not provide general solution, required huge hyper-parameters, computationally inefficient process datasets. Moreover, DL models high computational complexity, which requires memory cost, complexity experimental materials' backgrounds, makes it difficult train an efficient model. To address these issues, we developed Inception module, improved recognize four classes Chest X-ray this research by substituting original convolutions architecture based on modified-Xception (m-Xception). addition, model incorporates depth-separable convolution layers within layer, interlinked linear residuals. The model's training utilized two-stage transfer produce effective Finally, XgBoost classifier multiple X-rays. evaluate m-Xception model, 1095 dataset converted augmentation technique into 48,000 including 12,000 normal, pneumonia, COVID-19 opacity images. balance classes, technique. Using public datasets three distinct train-test divisions (80-20%, 70-30%, 60-40%) our work, attained average 96.5% accuracy, 96% F1 score, recall, precision. A comparative analysis demonstrates that method outperforms comparable existing methods. results experiments indicate proposed approach is intended assist radiologists better diagnosing different diseases.

Язык: Английский

---

IRCM‐Caps: An X‐ray image detection method for COVID‐19

The Clinical Respiratory Journal, Год журнала: 2023, Номер 17(5), С. 364 - 373

Опубликована: Март 15, 2023

Abstract Objective COVID‐19 is ravaging the world, but traditional reverse transcription‐polymerase reaction (RT‐PCR) tests are time‐consuming and have a high false‐negative rate lack of medical equipment. Therefore, lung imaging screening methods proposed to diagnose due its fast test speed. Currently, commonly used convolutional neural network (CNN) model requires large number datasets, accuracy basic capsule for multiple classification limital. For this reason, paper proposes novel based on CNN CapsNet. Methods The integrates And attention mechanism module multi‐branch lightweight applied enhance performance. Use contrast adaptive histogram equalization (CLAHE) algorithm preprocess image contrast. preprocessed images input into training, ReLU was as activation function adjust parameters achieve optimal. Result dataset includes 1200 X‐ray (400 COVID‐19, 400 viral pneumonia, normal), we replace VGG16, InceptionV3, Xception, Inception‐Resnet‐v2, ResNet50, DenseNet121, MoblieNetV2 integrate with Compared CapsNet, improves 6.96%, 7.83%, 9.37%, 10.47%, 10.38% in accuracy, area under curve (AUC), recall, F1 scores, respectively. In binary experiment, compared AUC, recall rate, score were increased by 5.33%, 5.34%, 2.88%, 8.00%, 5.56%, Conclusion embedded advantages has good effect small dataset.

Язык: Английский

---

DBM-ViT: A multiscale features fusion algorithm for health status detection in CXR / CT lungs images

Biomedical Signal Processing and Control, Год журнала: 2023, Номер 87, С. 105365 - 105365

Опубликована: Авг. 25, 2023

Язык: Английский

---

Deep learning‐based COVID‐19 diagnosis using CT scans with laboratory and physiological parameters

IET Image Processing, Год журнала: 2023, Номер 17(11), С. 3127 - 3142

Опубликована: Май 30, 2023

Abstract The global economy has been dramatically impacted by COVID‐19, which spread to be a pandemic. COVID‐19 virus affects the respiratory system, causing difficulty breathing in patient. It is crucial identify and treat infections as soon possible. Traditional diagnostic reverse transcription‐polymerase chain reaction (RT‐PCR) methods require more time find infection. A high infection rate, slow laboratory analysis, delayed test results caused widespread uncontrolled of disease. This study aims diagnose epidemic leveraging modified convolutional neural network (CNN) quickly safely predict disease's appearance from computed tomography (CT) scan images physiological parameters dataset. dataset representing 500 patients was used train, test, validate CNN model with detecting having an accuracy, sensitivity, specificity, F1‐score 99.33%, 99.09%, 99.52%, 99.24%, respectively. These experimental suggest that our strategy performs better than previously published approaches.

Язык: Английский

---