Investigating the Effect of Using Deep Learning Algorithms in CT scan and Chest X-ray Images on Improving the Diagnosis of Covid-۱۹

سال انتشار: 1402
نوع سند: مقاله کنفرانسی
زبان: انگلیسی
مشاهده: 92

نسخه کامل این مقاله ارائه نشده است و در دسترس نمی باشد

این مقاله در بخشهای موضوعی زیر دسته بندی شده است:

استخراج به نرم افزارهای پژوهشی:

لینک ثابت به این مقاله:

شناسه ملی سند علمی:

AIMS01_075

تاریخ نمایه سازی: 1 مرداد 1402

چکیده مقاله:

Background and aims: Advances in artificial intelligence and rapid changes in medical equipmenthave provided many opportunities in medical radiology. Advanced algorithms based ondeep learning, known as Convolutional Neural Network, play a significant role in extracting basicfeatures in medical images. The aim of this study was to investigate and compare machine learningand deep learning algorithms used to solve a clinical problem in medical images in order toimprove the diagnosis of Covid-۱۹ and increase the speed and accuracy in diagnosing this disease.Method: In this study, articles that used machine learning and deep learning algorithms to solvea clinical problem in medical images were reviewed. To identify related articles, PubMed, Web ofScience, Scopus and Google scholar search engines were searched based on keywords. Differentalgorithms based on artificial intelligence were extracted in the situation of the spread of the coronavirus. Automated techniques for classifying chest X-rays into pneumonia class or disease-freeclass are used from ۹ deep learning architectures including basic CNN, DenseNet۲۰۱, VGG۱۶,VGG۱۹, Inception_ResNet_V۲, Inception_V۳, Xception, Resnet۵۰. Finally, the type of algorithmused and its accuracy percentage were examined and then the algorithms were compared.Results: Overall, ۱۳ studies had used deep learning algorithms on COVID-۱۹ and healthy individuals’data. The most advanced COVID-۱۹ diagnostic systems that were used to identify andclassify COVID-۱۹ patients and normal individuals based on accuracy were: ResNet۵۰ (۹۶.۰۱)ResNet۱۰۱ (۹۶.۰۱) ResNet۱۵۲ (۹۳.۰۹) Inception network from V۱ to V۴ (۹۸.۷۰ and ۹۷.۹۷%) andCNN (۹۷.۶۲ and ۹۹.۴%). Therefore, the review of studies shows that deep learning with ConvolutionalNeural Networks (CNN) can have significant effects in automatic detection and extractionof very essential features from chest images related to the diagnosis of the corona virus. Advancedalgorithms based on deep learning, known as Convolutional Neural Network, play a significantrole in extracting basic features in high-precision medical images, and this method has been appliedusing CT and X-Ray image scans with significant results.Conclusion: Advanced algorithms based on deep learning, known as Convolutional Neural Networks,play a significant role in extracting essential features in high-precision medical images,and deep learning-based diagnosis systems can be useful in areas where experts and wellequippedclinics are not available.

نویسندگان

M Abbasi

MSc in Information Technology Management, Information Technology and Statistics Management Center, Mashhad university of medical sciences, Mashhad, Iran

A Kheirdoust

MSc in Medical Informatics, Vascular and Endovascular Surgery research center, Mashhad university of medical sciences, Mashhad, Iran

F Golmakani

MSc in Health Information Technology, Vascular and Endovascular Surgery research center, Mashhad university of medical sciences, Mashhad, Iran

M.R Mazaheri Habibi

Ph.D .in Medical Informatics, Department of Health Information Technology, Varastegan Institute for Medical Sciences, Mashhad, Iran

M.M Kamyar

Assistant professor in vascular surgery, Vascular and Endovascular Surgery research center, Mashhad university of medical sciences, Mashhad, Iran