Classification of X-Ray chest images by vision transformers and covid-19 detection

Dilek Civil; Ozge Oztimur Karadag

Makale detayları

Classification of X-Ray chest images by vision transformers and covid-19 detection

Yazarlar:

Dilek Civil | Ozge Oztimur Karadag

Yayıncı:

Bursa Uludağ University

Ana Başlıklar:

Yapay Zeka

Dergi:

Uludağ Üniversitesi Mühendislik Fakültesi Dergisi Vol.28 Issue 2, pp.349-364

Belge Tipi:

Makale

Yayın Tarihi:

8/2023

Belge Dili:

Türkçe

Document ID (Org No.):

2001296972

pISSN:

2148-4155

Anahtar kelimeler:

Özet:

 The COVID-19 virus, which caused the declaration of a worldwide pandemic in 2019, caused serious health problems and even death due to the rapid progression of the disease. Hence, early diagnosis of the disease has been of great importance to prevent the emergence of these adverse conditions. Medical imaging methods, such as X-Ray chest images, Computed Tomography CT, etc. have played an important role in the detection. Deep learning methods have proven their superior image analysis ability. In this study, it is proposed to use the Vision Transformer VIT, which has recently been popular in computer vision applications, for Covid-19 detection using X-ray chest images. The effectiveness of the VIT architecture is compared with a popular Convolutional Neural Network architecture, Resnet50 through various criteria, as well as VGG16 and InceptionV3 architectures, which were recorded to be successful in multi-class datasets of COVID-19. As a result of the experiments, it has been observed that the VIT architecture outperforms the Resnet50 and VGG16 architectures trained with the transfer learning method, and it has been observed that it achieves head-to-head results with InceptionV3, and it has been concluded that it can be used as an alternative decision support system for COVID-19 detection.

 The COVID-19 virus, which caused the declaration of a worldwide pandemic in 2019, caused serious health problems and even death due to the rapid progression of the disease. Hence, early diagnosis of the disease has been of great importance to prevent the emergence of these adverse conditions. Medical imaging methods, such as X-Ray chest images, Computed Tomography CT, etc. have played an important role in the detection. Deep learning methods have proven their superior image analysis ability. In this study, it is proposed to use the Vision Transformer VIT, which has recently been popular in computer vision applications, for Covid-19 detection using X-ray chest images. The effectiveness of the VIT architecture is compared with a popular Convolutional Neural Network architecture, Resnet50 through various criteria, as well as VGG16 and InceptionV3 architectures, which were recorded to be successful in multi-class datasets of COVID-19. As a result of the experiments, it has been observed that the VIT architecture outperforms the Resnet50 and VGG16 architectures trained with the transfer learning method, and it has been observed that it achieves head-to-head results with InceptionV3, and it has been concluded that it can be used as an alternative decision support system for COVID-19 detection.

 The COVID-19 virus, which caused the declaration of a worldwide pandemic in 2019, caused serious health problems and even death due to the rapid progression of the disease. Hence, early diagnosis of the disease has been of great importance to prevent the emergence of these adverse conditions. Medical imaging methods, such as X-Ray chest images, Computed Tomography CT, etc. have played an important role in the detection. Deep learning methods have proven their superior image analysis ability. In this study, it is proposed to use the Vision Transformer VIT, which has recently been popular in computer vision applications, for Covid-19 detection using X-ray chest images. The effectiveness of the VIT architecture is compared with a popular Convolutional Neural Network architecture, Resnet50 through various criteria, as well as VGG16 and InceptionV3 architectures, which were recorded to be successful in multi-class datasets of COVID-19. As a result of the experiments, it has been observed that the VIT architecture outperforms the Resnet50 and VGG16 architectures trained with the transfer learning method, and it has been observed that it achieves head-to-head results with InceptionV3, and it has been concluded that it can be used as an alternative decision support system for COVID-19 detection.

 2019 yılında dünya çapında pandemi ilan edilmesine sebep olan COVİD-19 virüsü, bulaştığı bazı kişilerde hastalığın çok hızlı ilerlemesi sebebi ile çok ciddi sağlık problemlerine, hatta ölümlere neden olmuştur. Hastalığın hızlı teşhisi bu olumsuz durumların ortaya çıkmasını engellemek için büyük önem arz etmiştir. X-Ray göğüs görüntüleri, Bilgisayarlı Tomografi BT vb. gibi tıbbi görüntüleme yöntemleri covid-19 un tespit edilmesinde önemli rol oynamıştır. Derin öğrenme yöntemleri, insan faktörünü ve insandan kaynaklı hata payını minimuma indirerek üstün görüntü analizi yeteneğini kanıtlamıştır. Bu çalışmada, son dönemlerde bilgisayarlı görü uygulamalarında öne çıkan görüntü dönüştürücülerin Vision Transformers, VIT X-ray göğüs görüntüleri üzerinde Covid-19 tespiti için kullanılması önerilmiştir. Önerilen sistemin eğitim ve test performansı literatürde yaygın kullanılan bir Konvolüsyonel Sinir Ağı, Resnet50, ile ayrıntılı olarak karşılaştırılarak analiz edilmiştir. Ayrıca VİT mimarisinin etkinliği doğruluk, kesinlik, hassasiyet ve F1 skoru kriterleri aracılığı ile Resnet50nin yanı sıra COVİD-19un çok sınıflı veri kümelerinde başarı gösterdiği kaydedilen VGG16 ve InceptionV3 mimarileri ile de karşılaştırılmıştır. Deneyler sonucunda VİT mimarisinin öğrenme aktarımı yöntemi ile eğitilen Resnet50 ve VGG16 mimarilerinden daha iyi performans gösterdiği ve InceptionV3 ile başa baş sonuçlar elde ettiği gözlemlenmiş ve COVİD-19 tespitinde alternatif bir karar destek sistemi olarak kullanılabileceği tespit edilmiştir.

Makale detayları

Classification of X-Ray chest images by vision transformers and covid-19 detection

Yazarlar:

Dilek Civil | Ozge Oztimur Karadag

Yayıncı:

Bursa Uludağ University

Ana Başlıklar:

Yapay Zeka

Dergi:

Uludağ Üniversitesi Mühendislik Fakültesi Dergisi Vol.28 Issue 2, pp.349-364

Belge Tipi:

Makale

Yayın Tarihi:

8/2023

Belge Dili:

Türkçe

Document ID (Org No.):

2001296972

pISSN:

2148-4155

COVID-19

Vision Transformers

Deep Learning

CNN

Transfer LearningCovid-19

Görüntü Dönüştürücü

Derin Öğrenme

CNN

Aktarım Öğrenme

Sayı

Makale detayları

Yazarlar: Dilek Civil | Ozge Oztimur Karadag

Yayıncı: Bursa Uludağ University

Ana Başlıklar: Yapay Zeka

Dergi: Uludağ Üniversitesi Mühendislik Fakültesi Dergisi Vol.28 Issue 2, pp.349-364

Belge Tipi: Makale

Yayın Tarihi: 8/2023

Belge Dili: Türkçe

Document ID (Org No.): 2001296972

pISSN: 2148-4155

Sayı

Yazarlar:

Dilek Civil | Ozge Oztimur Karadag

Yayıncı:

Bursa Uludağ University

Ana Başlıklar:

Yapay Zeka

Dergi:

Uludağ Üniversitesi Mühendislik Fakültesi Dergisi Vol.28 Issue 2, pp.349-364

Belge Tipi:

Makale

Yayın Tarihi:

8/2023

Belge Dili:

Türkçe

Document ID (Org No.):

2001296972

pISSN:

2148-4155