Improving COVID-19 Detection: Leveraging Convolutional Neural Networks in Chest X-Ray Imaging

Abstract

The global impact of the COVID-19 pandemic has significantly disrupted healthcare systems w orldwide. Amidst challenges, there is a crucial demand for efficient me thodologies to ex pedite di sease de tection. Th is research underscores the potential of Deep Neural Networks in enhancing pandemic management over the past five years. Focusing on Artificial Intelligence (AI) application in COVID-19 detection through X-ray imaging, this research advocates using Visual Geometry Group (VGG’16), a Convolutional Neural Network (CNN) used for image classification w ith m ultiple l ayers. T hese C NNs a ct a s c lassifier-based sy stems, tr eating im ages as structured data arrays to identify and learn patterns. Quantifying the model’s effectiveness t hrough t he a ccuracy s core, t his r esearch r eveals a 0 .90% accuracy, indicating the model’s accurate detection of COVID-19 cases in X-ray images. Additionally, the study highlights a significant a chievement w ith a l ess t han 1 0% f alse p ositive r ate, c rucial f or r eliable a nd p rompt COVID-19 diagnoses in the healthcare industry. In conclusion, this research presents an AI-driven approach, utilizing VGG’16 and convolutional neural networks to enhance the efficiency an d ac curacy of CO VID-19 de tection in X-ray imaging. The high accuracy score and low false positive rate positions this methodology as a valuable contribution, offering robust pandemic management and healthcare decision-making. © 2024 SPIE