Detection of Pneumonia with a Novel CNN-based Approach

Ebru ERDEM; Tolga AYDİN

doi:https://doi.org/10.35377/saucis.04.01.787030

Makale detayları

Detection of Pneumonia with a Novel CNN-based Approach

Yazarlar:

Ebru ERDEM | Tolga AYDİN

Ana Başlıklar:

Özel bilgisayar yöntemleri

Dergi:

Sakarya University Journal of Computer and Information Sciences 2021 Vol.4 Issue 1, pp.26-34

Belge Tipi:

Makale

Yayın Tarihi:

4/2021

Belge Dili:

İngilizce

Document ID (Org No.):

2000551926

DOI:

https://doi.org/10.35377/saucis.04.01.787030

Özet:

 

Pneumonia is a seasonal infectious lung tissue inflammatory disease. According to the World Health Organization (WHO), early diagnosis of the disease reduces the risk of its transmission and death. Various deep learning and machine learning algorithms were used for pneumonia detection. This study aims to analyze the lung images and diagnose pneumonia disease by employing deep learning approaches. We have suggested a novel deep learning framework for the detection of pneumonia in lung. A comparison was made between the proposed new deep learning model and pre-trained deep learning models. 88.62% accuracy rate has been obtained from the proposed deep learning structure. It was observed that by utilizing the new deep neural network developed, the accuracy results of VGG16 (88.78%) and VGG19 (88.30%), which are among the popular deep learning architectures, can be approximated. The test results show that our proposed model has a better recall value (97.43%) (VGG16 (93.33%) and VGG19 (96.92%)), and a better F1-Score (91.45%) (VGG16 (91.22%) and VGG19 (91.19%)).

Makale detayları

Detection of Pneumonia with a Novel CNN-based Approach

Yazarlar: Ebru ERDEM | Tolga AYDİN

Ana Başlıklar: Özel bilgisayar yöntemleri

Dergi: Sakarya University Journal of Computer and Information Sciences 2021 Vol.4 Issue 1, pp.26-34

Belge Tipi: Makale

Yayın Tarihi: 4/2021

Belge Dili: İngilizce

Document ID (Org No.): 2000551926

DOI: https://doi.org/10.35377/saucis.04.01.787030

Sayı

Yazarlar:

Ebru ERDEM | Tolga AYDİN

Ana Başlıklar:

Özel bilgisayar yöntemleri

Dergi:

Sakarya University Journal of Computer and Information Sciences 2021 Vol.4 Issue 1, pp.26-34

Belge Tipi:

Makale

Yayın Tarihi:

4/2021

Belge Dili:

İngilizce

Document ID (Org No.):

2000551926

DOI:

https://doi.org/10.35377/saucis.04.01.787030