Detecting COVID-19 from Chest X-Ray Images using Deep Learning

Abstract

COVID-19 is a deadly disease that may cause lasting harm to the lungs and other organs. It can be life-threatening if timely action is not taken and therefore detection at an early is vital. The objective of this paper is to use chest X-rays to identify COVID-19 images using deep learning models. Since COVID-19 harms the respiratory epithelial cells, we could make use of X-rays to determine the patient's lungs condition. When the number of patients is exceptionally high and thus the required radiological expertise is low, deep learning-based recommender systems can be extremely useful. The goal is to use pre-trained models to develop an image classification model that can predict Covid-19 in Chest X-Ray scans with reasonably high accuracy. CNN's are primarily used for medical image classification tasks as they can easily detect the important features and classify them accordingly. Four distinct pre-trained models were used for this purpose. In this work, the analysis of the results showed that compared with other models, the DenseNet201 model provides the highest accuracy (96.54%) in detecting chest X-rays. This model can be used by any medical professional on any system to quickly identify Covid +ve patients using chest X-ray scans.

Ähnliche Arbeiten