Coronaviruses, including SARS-CoV-2, are responsible for COVID-19, a highly transmissible disease that emerged in December 2019 in Wuhan, China. During the past five years, significant advancements have been made in understanding and mitigating the virus. Although the initial outbreak led to global health crises, improved vaccination strategies, antiviral treatments, and AI-driven diagnostic tools have contributed to better disease management. However, COVID-19 continues to pose risks, particularly for immuno-compromised individuals and those with pre-existing conditions. This study explores the use of deep learning for a rapid and accurate diagnosis of COVID-19, addressing ongoing challenges in healthcare infrastructure and testing accessibility. We propose an enhanced automated detection system leveraging state-of-the-art convolutional neural networks (CNNs), including updated versions of VGG16, VGG19, and ResNet50, to classify COVID-19 infections from chest radiographs and computerized tomography (CT) scans. Our results, based on an expanded dataset of over 6000 medical images, demonstrate that the optimized ResNet50 model achieves the highest classification performance, with 97.77% accuracy, 100% sensitivity, 93.33% specificity, and a 98.0% F1-score. These findings reinforce the potential of AI-assisted diagnostic tools in improving early detection and pandemic preparedness.
View on arXiv@article{dharmik2025_2503.12642,
title={ COVID 19 Diagnosis Analysis using Transfer Learning },
author={ Anjali Dharmik },
journal={arXiv preprint arXiv:2503.12642},
year={ 2025 }
}