COVID-19 and Pneumonia X-ray Classification via Custom Convolutional Neural Network

Abstract

COVID-19 has caused millions of deaths and dev-astation to the global economy. In an effort to control and limit the spread of COVID-19, a custom Convolutional Neural Network (CPnet) has been developed to aid in the diagnosis of the disease. The challenge is that the X-ray images of Pneumonia and COVID-19 are very similar. This can impact the diagnostic accuracy of Pneumonia and COVID-19 when a Convolutional Neural Network is employed. This paper will examine leading, high-accuracy models for COVID-19 detection and compare them to Pneumonia, implementing and optimizing features from these models to create CPnet. In terms of its training results, CPnet achieved 99.61% accuracy, 99.67% precision, 99.54% recall, and 99.60% F1 Score for binary classification of COVID-19 and Pneumonia. Additionally, CPnet's test results achieved 100% accuracy, 100% precision, 100% recall, and 100% F1 Score for binary classification between COVID-19 and Pneumonia. Comparable state-of-the-art models are currently outperformed by CPnet's results.

Ähnliche Arbeiten