Harnessing AI for Accurate and Faster Results: Reducing COVID-19 RT-PCR Testing Cost Through Machine Learning

Abstract

The COVID-19 pandemic has placed unprecedented demands on healthcare systems worldwide, particularly in the realm of testing and diagnosis. RT-PCR (Reverse Transcription Polymerase Chain Reaction) tests have been the gold standard for detecting the SARS-CoV-2 virus, but they are resource-intensive, time-consuming, and costly. This research focuses on the integration of artificial intelligence (AI) into the RT-PCR testing process to enhance its efficiency, accuracy, and cost-effectiveness. Here we proposed a two-stage deep-learning based COVID-19 patient detection system to reduce the cost and time of RT-PCR testing. To complete this research, we used both symptoms and chest X-ray datasets. First, we train a CNN (VGG16) model using a chest X-ray dataset and an Artificial Neural Network (ANN) model using a symptom dataset. After that, we connected the results of these two models to another ANN model. This latter model had been trained using a customized dataset with three labels: Negative, Need RT-PCR Test, and Positive. In our findings, we saw that our model performs with 96.3% accuracy in detecting Chest X-rays, 98.0% accuracy in detecting Symptoms, and 98.5% accuracy in the final combined COVID patients detection model. The findings of this research mark a crucial step towards a more sustainable and responsive testing infrastructure during pandemics.

Ähnliche Arbeiten