Visualization of COVID Bimodal scan using DNN

ABSTRACT

COVID-19 is a disease caused by SARS-CoV-2 that can arouse a respiratory tract infection. Therefore, a rapid identification of clearly visualized infections is urgently needed, which can assist early diagnosis and save the lives of suspected COVID-19 patients.Recent technological progress has made it possible to fuse deep learning classification and medical images that can accelerate and improve the accuracy of results when leveraged. This could particularly be important for disease where faster result and increased accuracy can help early detection of COVID-19 cases vis-à-vis the traditional RT-PCR tests. DNN classifier is designed such that, it automatically detects virus present in lungs using chest image is termed as Bimodal. This research article proposes an automatic frame work for identifying COVID -19 as early using chest X-ray images and CT Scan Images. For this experiment, 3 types of data set are used, 1) COVID X-ray chest 2) CT-scan SARS-COV-2, 3), X-Ray images in the chest (Pneumonia). This deep learning model can detect positive COVID-19 cases more quickly than RT-PCR tests for the detection of COVID-19 cases. The proposed model provides a relationship between COVID-19 patients and pneumonia patients. Color visualization approach on the basis of Grad-CAM is used to clearly interpret image radiology detection. The proposed deep learning model has achieved a total accuracy of 92.33%, with precision and recall of 0.94% and 0.93%. © 2021 IEEE.