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

BMC Medical Imaging, Journal Year: 2024, Volume and Issue: 24(1)

Published: Feb. 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.

Language: Английский

---

IEViT: An enhanced vision transformer architecture for chest X-ray image classification

Computer Methods and Programs in Biomedicine, Journal Year: 2022, Volume and Issue: 226, P. 107141 - 107141

Published: Sept. 16, 2022

Chest X-ray imaging is a relatively cheap and accessible diagnostic tool that can assist in the diagnosis of various conditions, including pneumonia, tuberculosis, COVID-19, others. However, requirement for expert radiologists to view interpret chest images be bottleneck, especially remote deprived areas. Recent advances machine learning have made possible automated scans. In this work, we examine use novel Transformer-based deep model task image classification.We first performance Vision Transformer (ViT) state-of-the-art classification classification, then propose evaluate Input Enhanced (IEViT), enhanced achieve improved on associated with pathologies.Experiments four data sets containing pathologies (tuberculosis, COVID-19) demonstrated proposed IEViT outperformed ViT all variants examined, achieving an F1-score between 96.39% 100%, improvement over up +5.82% terms across examined sets. IEViT's maximum sensitivity (recall) ranged 93.50% 100% sets, +3%, whereas precision 97.96% +6.41%.Results showed ViT's demonstrating its superiority generalisation ability. Given low cost widespread accessibility imaging, potentially offer powerful, but method assisting using images.

Language: Английский

---

IoMT with Deep CNN: AI-Based Intelligent Support System for Pandemic Diseases

Electronics, Journal Year: 2023, Volume and Issue: 12(2), P. 424 - 424

Published: Jan. 13, 2023

The Internet of Medical Things (IoMT) is an extended version the (IoT). It mainly concentrates on integration medical things for servicing needy people who cannot get services easily, especially rural area and aged peoples living alone. main objective this work to design a real time interactive system providing do not have sufficient infrastructure. With help system, will at their end with minimal infrastructure less treatment cost. However, designed could be upgraded address family SARs viruses, experimentation, we taken COVID-19 as test case. proposed comprises many modules, such user interface, analytics, cloud, etc. interface data collection. At initial stage, it collects preliminary information, pulse oxygen rate RT-PCR results. oximeter, they level. swap kit, find positivity. That information uploaded via UI. If identifies COVID positivity, requests that person upload X-ray/CT images ranking severity disease. multi-model data. Hence, can deal X-ray, CT images, textual (RT-PCR results). Once are collected UI, those forwarded AI module analytics. multi-disease classification. classifies patients affected or pneumonia any other viral infection. also measures intensity level lung infection suitable patients. Numerous deep convolution neural network (DCNN) architectures available image We used ResNet-50, ResNet-100, ResNet-101, VGG 16, 19 better From observed ResNet101 outperform, accuracy 97% images. outperforms 98% X-ray For obtaining enhanced accuracy, major voting classifier. combines all classifiers result presents majority voted one. results in reduced classifier bias. Finally, automatic summary report textually. accessed user-friendly graphical (GUI). generation individual

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

---

Deep Learning for Pneumonia Detection in Chest X-ray Images: A Comprehensive Survey

Journal of Imaging, Journal Year: 2024, Volume and Issue: 10(8), P. 176 - 176

Published: July 23, 2024

This paper addresses the significant problem of identifying relevant background and contextual literature related to deep learning (DL) as an evolving technology in order provide a comprehensive analysis application DL specific pneumonia detection via chest X-ray (CXR) imaging, which is most common cost-effective imaging technique available worldwide for diagnosis. particular key period associated with COVID-19, 2020–2023, explain, analyze, systematically evaluate limitations approaches determine their relative levels effectiveness. The context applied both aid automated substitute existing expert radiography professionals, who often have limited availability, elaborated detail. rationale undertaken research provided, along justification resources adopted relevance. explanatory text subsequent analyses are intended sufficient detail being addressed, solutions, these, ranging from more general. Indeed, our evaluation agree generally held view that use transformers, specifically, vision transformers (ViTs), promising obtaining further effective results area using CXR images. However, ViTs require extensive address several limitations, specifically following: biased datasets, data code ease model can be explained, systematic methods accurate comparison, notion class imbalance possibility adversarial attacks, latter remains fundamental research.

Language: Английский

---

SSP: self-supervised pertaining technique for classification of shoulder implants in x-ray medical images: a broad experimental study

Artificial Intelligence Review, Journal Year: 2024, Volume and Issue: 57(10)

Published: Aug. 18, 2024

Abstract Multiple pathologic conditions can lead to a diseased and symptomatic glenohumeral joint for which total shoulder arthroplasty (TSA) replacement may be indicated. The long-term survival of implants is limited. With the increasing incidence surgery, it anticipated that revision surgery will become more common. It challenging at times retrieve manufacturer in situ implant. Therefore, certain systems facilitated by AI techniques such as deep learning (DL) help correctly identify implanted prosthesis. Correct identification reduce perioperative complications complications. DL was used this study categorise different based on X-ray images into four classes (as first case small dataset): Cofield, Depuy, Tornier, Zimmer. Imbalanced public datasets poor performance model training. Most methods literature have adopted idea transfer (TL) from ImageNet models. This type TL has been proven ineffective due some concerns regarding contrast between features learnt natural (ImageNet: colour images) (greyscale images). To address that, new approach (self-supervised pertaining (SSP)) proposed resolve issue datasets. SSP training models (ImageNet models) large number unlabelled greyscale medical domain update features. are then trained labelled data set implants. shows excellent results five models, including MobilNetV2, DarkNet19, Xception, InceptionResNetV2, EfficientNet with precision 96.69%, 95.45%, 98.76%, 98.35%, 96.6%, respectively. Furthermore, shown domains (such ImageNet) do not significantly affect images. A lightweight scratch achieves 96.6% accuracy, similar using standard extracted train several ML classifiers show outstanding obtaining an accuracy 99.20% Xception+SVM. Finally, extended experimentation carried out elucidate our approach’s real effectiveness dealing imaging scenarios. Specifically, tested without SSP, 99.47% CT brain stroke 98.60%.

Language: Английский

---

Screening Lung Diseases Using Cascaded Feature Generation and Selection Strategies

Healthcare, Journal Year: 2022, Volume and Issue: 10(7), P. 1313 - 1313

Published: July 14, 2022

The global pandemic COVID-19 is still a cause of health emergency in several parts the world. Apart from standard testing techniques to identify positive cases, auxiliary tools based on artificial intelligence can help with identification and containment disease. need for development alternative smart diagnostic combat has become more urgent. In this study, framework machine learning (ML) proposed; it medical practitioners COVID-19-affected patients, among others pneumonia healthy individuals, monitoring status cases using X-ray images. We investigated application transfer-learning (TL) networks various feature-selection improving classification accuracy ML classifiers. Three different TL were tested generate relevant features images; these include AlexNet, ResNet101, SqueezeNet. generated further refined by applying methods that iterative neighborhood component analysis (iNCA), chi-square (iChi2), maximum relevance-minimum redundancy (iMRMR). Finally, was performed convolutional neural network (CNN), linear discriminant (LDA), support vector (SVM) Moreover, study exploited stationary wavelet (SW) transform handle overfitting problem decomposing each image training set up three levels. Furthermore, enhanced dataset, operations as data-augmentation techniques, including random rotation, translation, shear operations. revealed combination SqueezeNet, iChi2, SVM very effective images, producing 99.2%. Similarly, along iChi2 proposed CNN network, yielded 99.0% accuracy. results showed cascaded feature generator selection strategies significantly affected performance classifier.

Language: Английский

---

Conv-CapsNet: capsule based network for COVID-19 detection through X-Ray scans

Multimedia Tools and Applications, Journal Year: 2023, Volume and Issue: 82(18), P. 28521 - 28545

Published: Feb. 21, 2023

Language: Английский

---

MV-MS-FETE: Multi-view multi-scale feature extractor and transformer encoder for stenosis recognition in echocardiograms

Computer Methods and Programs in Biomedicine, Journal Year: 2024, Volume and Issue: 245, P. 108037 - 108037

Published: Jan. 21, 2024

Background: aortic stenosis is a common heart valve disease that mainly affects older people in developed countries. Its early detection crucial to prevent the irreversible progression and, eventually, death. A typical screening technique detect uses echocardiograms; however, variations introduced by other tissues, camera movements, and uneven lighting can hamper visual inspection, leading misdiagnosis. To address these issues, effective solutions involve employing deep learning algorithms assist clinicians detecting classifying developing models predict this pathology from single views. Although promising, information conveyed image may not be sufficient for an accurate diagnosis, especially when using automatic system; thus, indicates different should explored. Methodology: following rationale, paper proposes novel architecture, composed of multi-view, multi-scale feature extractor, transformer encoder (MV-MS-FETE) parasternal long short-axis In particular, starting latter, designed model extracts relevant features at multiple scales along its extractor component takes advantage perform final classification. Results: experiments were performed on recently released Tufts medical echocardiogram public dataset, which comprises 27,788 images split into training, validation, test sets. Due recent release collection, tests also conducted several state-of-the-art create multi-view single-view benchmarks. For all models, standard classification metrics computed (e.g., precision, F1-score). The obtained results show proposed approach outperforms methods terms accuracy F1-score has more stable performance throughout training procedure. Furthermore, highlight generally better than their counterparts. Conclusion: introduces recognition, as well three benchmarks evaluate it, effectively providing comparisons fully model's effectiveness aiding performing diagnoses while producing baselines recognition task.

Language: Английский

---

Real-time deep learning method for automated detection and localization of structural defects in manufactured products

Computers & Industrial Engineering, Journal Year: 2022, Volume and Issue: 172, P. 108512 - 108512

Published: July 29, 2022

Language: Английский

---