Digital Sampling and Testing of COVID-19 to Reduce Lab Procedures Further Decreasing Carbon Emission: Review

Ananya Niraj Kumar Singh; Lacky Shiv Kumar Singh

Abstract

The COVID-19 pandemic has caused a new virus known as severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) that primarily infects the lungs and can lead to serious respiratory illnesses such as ARDS and pneumonia caused by coronavirus. Declared a pandemic by the World Health Organization (WHO), the disease belongs to the family of coronaviruses that can cause cold and flu, making it a highly contagious disease. Chest imaging tests such as CT scans, X-rays, and other radiological tests are commonly used to detect and test for COVID-19. However, manual diagnosis of these images can be a time-consuming process that generates considerable biomedical waste, thereby becoming an environmental concern. The goal of this study is to determine whether transfer learning can be used to streamline AI procedures. To achieve this, a deep learning framework using pre-trained deep convolutional neural network models is used. The result of this approach is highly accurate classification of chest radiographs. Results were visualized using pre-trained models (VGG16 and Conv2D). Experimental results showed that the accuracy obtained using VGG16 was 94% with a loss of 7%. However, in validated data, the accuracy achieved using VGG16 was 90% with a loss of 15%.

Abstract

Keywords