Detection & Diagnosis of COVID-19 from CXR Images Through VGG19 Transfer Learning Model

Abstract

COVID-19 must be diagnosed rapidly and precisely to control the current pandemic effectively. Using a dataset of 21,165 chest X-ray images, this study suggests a transfer learning approach for identifying and stratification of COVID-19. Three subsets of the dataset—training, validation, and test—each including 16,932, 2,116, and 2,117 images—are created. The VGG19 transfer learning model uses deep learning approaches to extract pertinent image characteristics and accurately classify the input images. By reaching a high accuracy rate, the key objective of this study is to distinguish COVID-19-infected patients from others with distinct lungrelated abnormalities. The model successfully classifies images with an incredible 94.85% accuracy, highlighting its potential for biomedical applications in the real world. The model's robustness and generalizability are aided by the extensive and varied dataset, which includes four annotated classes. The VGG19 model has much to offer regarding healthcare advantages if successfully implemented. The speed and accuracy of diagnosis can be improved by automatically identifying and classifying COVID-19 instances from chest Xray images, allowing for quicker patient triage and resource allocation. Additionally, it can lighten the load on radiologists and health workers and promote efficient management tactics throughout the pandemic. This study emphasizes the potential of models based on deep learning for COVID-19 identification in digital image processing. Future studies should improve the model's effectiveness, determine if it can be used in various groups, and include more clinical data for thorough research. In the continuing COVID-19 pandemic, the proposed model may help medical practitioners identify patients more quickly and accurately, leading to better patient outcomes.

Ähnliche Arbeiten