Publication Date: 2020/10/29
Abstract: COVID-19 has led to a worldwide surge of patients with acute respiratory distress syndrome (ARDS) in intensive care units. Milder cases of the virus that do not reach the ARDS stage are still often characterized by inflammation of the lung, causing shortness of breath. A salient step in fighting COVID-19 is the ability to detect infected patients early enough to be able to put them under special care. Detecting the virus from radiology images can be a much-needed, expeditious method to diagnose patients. Given this, we propose a method to detect COVID-19 using chest X-ray images as well as embeddings generated from such images. The model used to train this COVID-19 data is a Support Vector Machine (SVM) Classifier. We achieved an accuracy of 55% on raw image data and 63% on embeddings of X-ray images generated using Resnet. Further refinement is possible by training a larger image data set, extra pre-processing steps and data image refining techniques, and more sophisticated modelling to improve accuracy
Keywords: SVM, COVID-19, X-ray, Machine Learning
DOI: No DOI Available
PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT20OCT501.pdf
REFERENCES