A methodical exploration of imaging modalities from dataset to detection through machine learning paradigms in prominent lung disease diagnosis: a review

BMC Medical Imaging, Год журнала: 2024, Номер 24(1)

Опубликована: Фев. 1, 2024

Abstract Background Lung diseases, both infectious and non-infectious, are the most prevalent cause of mortality overall in world. Medical research has identified pneumonia, lung cancer, Corona Virus Disease 2019 (COVID-19) as prominent diseases prioritized over others. Imaging modalities, including X-rays, computer tomography (CT) scans, magnetic resonance imaging (MRIs), positron emission (PET) others, primarily employed medical assessments because they provide computed data that can be utilized input datasets for computer-assisted diagnostic systems. used to develop evaluate machine learning (ML) methods analyze predict diseases. Objective This review analyzes ML paradigms, modalities' utilization, recent developments Furthermore, also explores various available publically being Methods The well-known databases academic studies have been subjected peer review, namely ScienceDirect, arXiv, IEEE Xplore, MDPI, many more, were search relevant articles. Applied keywords combinations procedures with primary considerations such COVID-19, ML, convolutional neural networks (CNNs), transfer learning, ensemble learning. Results finding indicates X-ray preferred detecting while CT scan predominantly favored cancer. COVID-19 detection, datasets. analysis reveals X-rays scans surpassed all other techniques. It observed using CNNs yields a high degree accuracy practicability identifying Transfer complementary techniques facilitate analysis. is metric assessment.

Язык: Английский

---

COVID-19 CT-images diagnosis and severity assessment using machine learning algorithm

Cluster Computing, Год журнала: 2023, Номер 27(1), С. 547 - 562

Опубликована: Янв. 24, 2023

Язык: Английский

---

A Novel Approach to Detect COVID-19: Enhanced Deep Learning Models with Convolutional Neural Networks

Applied Sciences, Год журнала: 2022, Номер 12(18), С. 9325 - 9325

Опубликована: Сен. 17, 2022

The novel coronavirus (COVID-19) is a contagious viral disease that has rapidly spread worldwide since December 2019, causing the disruption of life and heavy economic losses. Since beginning virus outbreak, polymerase chain reaction been used to detect virus. However, it an expensive slow method, artificial intelligence researchers have attempted develop quick, inexpensive alternative methods diagnosis help doctors identify positive cases. Therefore, are starting incorporate chest X-ray scans (CXRs), easy examination method. This study approach uses image cropping deep learning technique (updated VGG16 model) classify three public datasets. had four main steps. First, data were split into training testing sets (70% 30%, respectively). Second, in processing step, each was cropped show only area. images then resized 150 × 150. third step build updated convolutional neural network (VGG16-CNN) model using multiple classifications (three classes: COVID-19, normal, pneumonia) binary classification (COVID-19 normal). fourth evaluate model’s performance accuracy, sensitivity, specificity. obtained 97.50% accuracy for 99.76% classification. also got best COVID-19 (99%) both models. It can be considered scientific contribution this research summarized as: reduced from approximately 138 million parameters around 40 parameters. Further, tested on different datasets proved highly efficient performance.

Язык: Английский

---

Advances in artificial intelligence for accurate and timely diagnosis of COVID-19: A comprehensive review of medical imaging analysis

Scientific African, Год журнала: 2023, Номер 22, С. e01961 - e01961

Опубликована: Ноя. 1, 2023

In December 2019, the first case of coronavirus 2019 (COVID-19) appeared in China, quickly leading to a global pandemic. Early and accurate diagnosis is crucial for effective disease management. While reverse transcription polymerase chain reaction (RT-PCR) standard diagnostic test, it may yield false negative misleading results. Artificial intelligence (AI) systems are accelerating transformation medical field, particularly early detection diagnosis. Recent research has combined AI with imaging modalities, such as chest X-ray (CXR) computed tomography (CT), detect virus, aiding doctors making decisions reducing misdiagnosis rates. this article, we conducted systematic review high-quality articles published high-impact journals that examined convolutional neural network (CNN)-based methods detecting COVID-19 from radiographic or CT images discussed associated issues. We synthesized publicly available datasets evaluation measures, including accuracy, sensitivity, specificity, F1 score, each system used automatic using several well-performing CNN architectures. Furthermore, identified key questions future directions field. Our results show use considerable potential improve accuracy reduce Nevertheless, important challenges must be addressed, limited access need rigorous model validation. Additionally, generalization models different populations contexts needs examined. findings underscore directions, exploration deep learning smaller datasets, enhancing performance complex cases, designing practical deployment clinical settings.

Язык: Английский

---

Activity-aware electrocardiogram biometric verification utilising deep learning on wearable devices

EURASIP Journal on Information Security, Год журнала: 2025, Номер 2025(1)

Опубликована: Фев. 25, 2025

Язык: Английский

---

Classification of the ICU Admission for COVID-19 Patients with Transfer Learning Models Using Chest X-Ray Images

Diagnostics, Год журнала: 2025, Номер 15(7), С. 845 - 845

Опубликована: Март 26, 2025

Objectives: Predicting intensive care unit (ICU) admissions during pandemic outbreaks such as COVID-19 can assist clinicians in early intervention and the better allocation of medical resources. Artificial intelligence (AI) tools are promising for this task, but their development be hindered by limited availability training data. This study aims to explore model strategies data-limited scenarios, specifically detecting need ICU admission using chest X-rays patients leveraging transfer learning data extension improve performance. Methods: We explored convolutional neural networks (CNNs) pre-trained on either natural images or X-rays, fine-tuning them a relatively dataset (COVID-19-NY-SBU, n = 899) lung-segmented X-ray classification. To further address scarcity, we introduced strategy that integrates an additional (MIDRC-RICORD-1c, 417) with different clinically relevant labels. Results: The TorchX-SBU-RSNA ELIXR-SBU-RSNA models, X-ray-pre-trained models our approach, enhanced classification performance from baseline AUC 0.66 (56% sensitivity 68% specificity) AUCs 0.77-0.78 (58-62% 78-80% specificity). gradient-weighted class activation mapping (Grad-CAM) analysis demonstrated focused more precisely lung regions reduced distractions non-relevant areas compared image-pre-trained without expansion. Conclusions: demonstrates benefits image-specific pre-training strategic expansion enhancing imaging AI models. Moreover, approach potential diverse sources alleviate limitations AI. developed may facilitate effective efficient patient management resource future infectious respiratory diseases.

Язык: Английский

---

Machine Learning First Response to COVID-19: A Systematic Literature Review of Clinical Decision Assistance Approaches during Pandemic Years from 2020 to 2022

Electronics, Год журнала: 2024, Номер 13(6), С. 1005 - 1005

Опубликована: Март 7, 2024

Background: The declaration of the COVID-19 pandemic triggered global efforts to control and manage virus impact. Scientists researchers have been strongly involved in developing effective strategies that can help policy makers healthcare systems both monitor spread mitigate impact pandemic. Machine Learning (ML) Artificial Intelligence (AI) applied several fronts fight. Foremost is diagnostic assistance, encompassing patient triage, prediction ICU admission mortality, identification mortality risk factors, discovering treatment drugs vaccines. Objective: This systematic review aims identify original research studies involving actual data construct ML- AI-based models for clinical decision support early response during years. Methods: Following PRISMA methodology, two large academic publication indexing databases were searched investigate use ML-based technologies their applications combat Results: literature search returned more than 1000 papers; 220 selected according specific criteria. illustrate usefulness ML with respect supporting professionals (1) triage patients depending on disease severity, (2) predicting hospital or Intensive Care Units (ICUs), (3) new repurposed treatments (4) factors. Conclusion: ML/AI community was able propose develop a wide variety solutions hospitalizations recommendations diagnostic, opening door further integration practices fighting this forecoming pandemics. However, translation practice impeded by heterogeneity datasets methodological computational approaches. lacks robust model validations desired translation.

Язык: Английский

---

Added value of chest CT in a machine learning-based prediction model to rule out COVID-19 before inpatient admission: A retrospective university network study

European Journal of Radiology, Год журнала: 2023, Номер 163, С. 110827 - 110827

Опубликована: Апрель 7, 2023

During the coronavirus disease 2019 (COVID-19) pandemic, hospitals still face challenge of timely identification infected individuals before inpatient admission. An artificial intelligence approach based on an established clinical network may improve prospective pandemic preparedness.Supervised machine learning was used to construct diagnostic models predict COVID-19. A pooled database retrospectively generated from 4437 participant data that were collected between January 2017 and October 2020 at 12 German centers belong radiological cooperative COVID-19 (RACOON) consortium. total 692 (15.6 %) participants positive according reference reverse transcription-polymerase chain reaction test. The included chest CT features (model R), examination laboratory test CL), or all three feature categories RCL). Performance outcomes accuracy, sensitivity, specificity, negative predictive value, area under receiver operating curve (AUC).Performance improved significantly by adding evaluation features. Without CL) with inclusion RCL), sensitivity 0.82 0.89 (p < 0.0001), specificity 0.84 value 0.96 0.97 AUC 0.92 0.95 proportion false classifications 2.6 % 1.7 respectively.Addition learning-based improves effectiveness in ruling out admission regular wards.

Язык: Английский

---

Analysis of Swin-UNet vision transformer for Inferior Vena Cava filter segmentation from CT scans

Artificial Intelligence in the Life Sciences, Год журнала: 2023, Номер 4, С. 100084 - 100084

Опубликована: Авг. 18, 2023

The purpose of this study is to develop an accurate deep learning model capable Inferior Vena Cava (IVC) filter segmentation from CT scans. does a comparative assessment the impact Residual Networks (ResNets) complemented with reduced convolutional layer depth and also analyzes using vision transformer architectures without performance degradation. This experimental retrospective on 84 scans consisting 54618 slices involves design, implementation, evaluation algorithm which can be used generate clinical report for presence IVC filters abdominal performed any reason. Several variants patch-based 3D-Convolutional Neural Network (CNN) Swin UNet Transformer (Swin-UNETR) retrieve signature filters. Dice Score as metric compare models. Model trained variant four ResNet layers showed higher achieving median = 0.92 [Interquartile range(IQR): 0.85, 0.93] compared plain having 0.89 [IQR: 0.83, 0.92]. Segmentation results two achieved 0.93 0.87, 0.94] was than at 0.87 0.77, 0.90]. Models SWIN-based transformers significantly better in both training validation datasets CNN variants. highest 4 UNETR 0.88 followed by 2 0.85. Utilization based Swin-UNETR output low bias variance thereby solving real-world problem within healthcare advanced Artificial Intelligence (AI) image processing recognition. will reduce time spent manually tracking centralizing electronic health record. Link GitHub repository.

Язык: Английский

---

Computational Simulation of Virtual Patients Reduces Dataset Bias and Improves Machine Learning-Based Detection of ARDS from Noisy Heterogeneous ICU Datasets

IEEE Open Journal of Engineering in Medicine and Biology, Год журнала: 2023, Номер 5, С. 611 - 620

Опубликована: Фев. 8, 2023

Machine learning (ML) technologies that leverage large-scale patient data are promising tools predicting disease evolution in individual patients. However, the limited generalizability of ML models developed on single-center datasets, and their unproven performance real-world settings, remain significant constraints to widespread adoption clinical practice. One approach tackle this issue is base large multi-center datasets. such heterogeneous datasets can introduce further biases driven by origin, as structures cohorts may differ between hospitals.

Язык: Английский

---