Deep learning framework for rapid and accurate respiratory COVID-19 prediction using chest X-ray images

Journal of King Saud University - Computer and Information Sciences, Journal Year: 2023, Volume and Issue: 35(7), P. 101596 - 101596

Published: May 25, 2023

COVID-19 is a contagious disease that affects the human respiratory system. Infected individuals may develop serious illnesses, and complications result in death. Using medical images to detect from essentially identical thoracic anomalies challenging because it time-consuming, laborious, prone error. This study proposes an end-to-end deep-learning framework based on deep feature concatenation Multi-head Self-attention network. Feature involves fine-tuning pre-trained backbone models of DenseNet, VGG-16, InceptionV3, which are trained large-scale ImageNet, whereas network adopted for performance gain. End-to-end training evaluation procedures conducted using COVID-19_Radiography_Dataset binary multi-classification scenarios. The proposed model achieved overall accuracies (96.33% 98.67%) F1_scores (92.68% multi classification scenarios, respectively. In addition, this highlights difference accuracy (98.0% vs. 96.33%) F_1 score (97.34% 95.10%) when compared with against highest individual performance. Furthermore, virtual representation saliency maps employed attention mechanism focusing abnormal regions presented explainable artificial intelligence (XAI) technology. provided better prediction results outperforming other recent learning same dataset.

Language: Английский

---

Chest X-ray in Emergency Radiology: What Artificial Intelligence Applications Are Available?

Diagnostics, Journal Year: 2023, Volume and Issue: 13(2), P. 216 - 216

Published: Jan. 6, 2023

Due to its widespread availability, low cost, feasibility at the patient’s bedside and accessibility even in low-resource settings, chest X-ray is one of most requested examinations radiology departments. Whilst it provides essential information on thoracic pathology, can be difficult interpret prone diagnostic errors, particularly emergency setting. The increasing availability large datasets has allowed development reliable Artificial Intelligence (AI) tools help radiologists everyday clinical practice. AI integration into workflow would benefit patients, radiologists, healthcare systems terms improved standardized reporting accuracy, quicker diagnosis, more efficient management, appropriateness therapy. This review article aims provide an overview applications for X-rays setting, emphasizing detection evaluation pneumothorax, pneumonia, heart failure, pleural effusion.

Language: Английский

---

COVID-19 detection and analysis from lung CT images using novel channel boosted CNNs

Expert Systems with Applications, Journal Year: 2023, Volume and Issue: 229, P. 120477 - 120477

Published: May 17, 2023

In December 2019, the global pandemic COVID-19 in Wuhan, China, affected human life and worldwide economy. Therefore, an efficient diagnostic system is required to control its spread. However, automatic poses challenges with a limited amount of labeled data, minor contrast variation, high structural similarity between infection background. this regard, new two-phase deep convolutional neural network (CNN) based proposed detect minute irregularities analyze infection. first phase, novel SB-STM-BRNet CNN developed, incorporating channel Squeezed Boosted (SB) dilated convolutional-based Split-Transform-Merge (STM) block infected lung CT images. The STM blocks performed multi-path region-smoothing boundary operations, which helped learn variation specific patterns. Furthermore, diverse boosted channels are achieved using SB Transfer Learning concepts texture COVID-19-specific healthy second images provided COVID-CB-RESeg segmentation identify infectious regions. methodically employed region-homogeneity heterogeneity operations each encoder-decoder boosted-decoder auxiliary simultaneously low illumination boundaries region. yields good performance terms accuracy: 98.21 %, F-score: 98.24%, Dice Similarity: 96.40 IOU: 98.85 % for would reduce burden strengthen radiologist's decision fast accurate diagnosis.

Language: Английский

---

Joint Diagnosis of Pneumonia, COVID-19, and Tuberculosis from Chest X-ray Images: A Deep Learning Approach

Diagnostics, Journal Year: 2023, Volume and Issue: 13(15), P. 2562 - 2562

Published: Aug. 1, 2023

Pneumonia, COVID-19, and tuberculosis are some of the most fatal common lung diseases in current era. Several approaches have been proposed literature for diagnosis individual diseases, since each requires a different feature set altogether, but few studies joint diagnosis. A patient being diagnosed with one disease as negative may be suffering from other disease, vice versa. However, said related to lungs, there might likelihood more than present same patient. In this study, deep learning model that is able detect mentioned chest X-ray images patients proposed. To evaluate performance model, multiple public datasets obtained Kaggle. Consequently, achieved 98.72% accuracy all classes general recall score 99.66% 99.35% No-findings, 98.10% Tuberculosis, 96.27% respectively. Furthermore, was tested using unseen data augmented dataset proven better state-of-the-art terms metrics.

Language: Английский

---

OPT-CO: Optimizing pre-trained transformer models for efficient COVID-19 classification with stochastic configuration networks

Information Sciences, Journal Year: 2024, Volume and Issue: 680, P. 121141 - 121141

Published: July 8, 2024

Building upon pre-trained ViT models, many advanced methods have achieved significant success in COVID-19 classification. Many scholars pursue better performance by increasing model complexity and parameters. While these can enhance performance, they also require extensive computational resources extended training times. Additionally, the persistent challenge of overfitting, due to limited dataset sizes, remains a hurdle. To address challenges, we proposed novel method optimize transformer models for efficient classification with stochastic configuration networks (SCNs), referred as OPT-CO. We two optimization methods: sequential (SeOp) parallel (PaOp), incorporating optimizers manner, respectively. Our without necessitating parameter expansion. introduced OPT-CO-SCN avoid overfitting problems through adoption random projection head augmentation. The experiments were carried out evaluate our based on publicly available datasets. Based evaluation results, superior, surpassing other state-of-the-art methods.

Language: Английский

---

Explainable deep learning diagnostic system for prediction of lung disease from medical images

Computers in Biology and Medicine, Journal Year: 2024, Volume and Issue: 170, P. 108012 - 108012

Published: Jan. 19, 2024

Language: Английский

---

Spiking neural network classification of X-ray chest images

Knowledge-Based Systems, Journal Year: 2025, Volume and Issue: unknown, P. 113194 - 113194

Published: Feb. 1, 2025

Language: Английский

---

Traffic Flow Management of Autonomous Vehicles Using Deep Reinforcement Learning and Smart Rerouting

IEEE Access, Journal Year: 2021, Volume and Issue: 9, P. 51005 - 51019

Published: Jan. 1, 2021

Autonomous Vehicles (AVs) promise to disrupt the traditional systems of transportation. An autonomous driving environment requires an uninterrupted, continuous stream data and information based on complex traffic sets predictive measurements make critical real-time decisions in uncertain situations. Such fosters a self-organizing system where vehicles must be seamlessly connected various other services intelligent manage flow executed emergent manner. To proceed towards this vision, paper, we develop management model which is novel two-phase approach for AVs optimize during congestion periods. In first phase our approach, build adaptive signal control, using Deep Reinforcement Learning (DRL) road intersections periods when congested. second phase, implement Smart Re-routing (SR) technique approaching intersections. used carry out load-balancing alternate paths avoid congested The experimental evaluation proposed validated simulations that demonstrate up 31% improved performance efficiency compared settings pre-timed signals without re-routing. improves overall while reducing delays minimizing long queues' lengths. This useful making infrastructure enough handle balance efficiently.

Language: Английский

---

Generative adversarial network based data augmentation for CNN based detection of Covid-19

Scientific Reports, Journal Year: 2022, Volume and Issue: 12(1)

Published: Nov. 10, 2022

Abstract Covid-19 has been a global concern since 2019, crippling the world economy and health. Biological diagnostic tools have developed to identify virus from bodily fluids causes pneumonia, which results in lung inflammation, presence of can also be detected using medical imaging by expert radiologists. The success each method is measured hit rate for identifying Covid infections. However, access people diagnosis tool limited, depending on geographic region and, treatment denotes race against time, duration plays an important role. Hospitals with X-ray opportunities are widely distributed all over world, so investigating images possible infections would offer itself. Promising achieved literature automatically detecting like CT scans X-rays supervised artificial neural network algorithms. One major drawbacks learning models that they require enormous amounts data train, generalize new data. In this study, we develop Swish activated, Instance Batch normalized Residual U-Net GAN dense blocks skip connections create synthetic augmented training. proposed architecture, due instance normalization swish activation, deal randomness luminosity, arises different sources better than classical architecture generate realistic-looking Also, radiology equipment not generally computationally efficient. They cannot efficiently run state-of-the-art deep networks such as DenseNet ResNet effectively. Hence, propose novel CNN 40% lighter more accurate networks. Multi-class classification three classes chest (CXR), ie Covid-19, healthy Pneumonia, performed model had extremely high test accuracy 99.2% any previous studies literature. Based mentioned criteria developing Corona infection diagnosis, present Artificial Intelligence based proposed, resulting rapid generative adversarial convolutional benefit will identification 99% accuracy. This could lead support helps accessible CXR images.

Language: Английский

---

ChestX-Ray6: Prediction of multiple diseases including COVID-19 from chest X-ray images using convolutional neural network

Expert Systems with Applications, Journal Year: 2022, Volume and Issue: 211, P. 118576 - 118576

Published: Aug. 27, 2022

Language: Английский

---