Coronavirus detection from chest x-rays images using CNN and GLSM methods

• Published in: AIP conference proceedings Vol. 2834; no. 1
• Main Authors: Mohammed, Faisel G., Yassir, Yassir Hussein
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Melville American Institute of Physics 04.12.2023
• Subjects: Abnormalities; Algorithms; Artificial neural networks; Chest; Coronaviruses; COVID-19; Lungs; Machine learning; Medical imaging; Neural networks; Viral diseases
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0162008

In global population, the wellbeing and health of individuals are heavily affected by the Corona virus Disease 2019 (COVID19). Efficient screening regarding the infected patients is considered as one of the essential steps for fighting against the pandemic, with radiology examination with the use of chest radiography being one of the key screening method. Early researches discovered that COVID19 patients have abnormalities in the chest radiography images. The modus operandi of the proposed system’s components will be clarified in this article. There are two sections to the system. Detecting and recognizing COVID-19. The recognition part of the proposed work used histogram orientation gradient (HOG) algorithm to specify lungs, and convolution neural networks (CNNs) used to classify lung have COVID-19 or not. Convolutional neural networks (CNNs) have been widely used in the applications of deep learning that have rapidly evolved over the last decade, most notably as a method for analyzing medical pictures. Detection part consist gray level co-occurrence matrix (GLCM) to detect the feature of COVID-19 region. The results of proposed method get 100% in training, 99.86 in validation, and 98.5 in testing with dataset contains two classes (each class contains 200 image)