A Novel Deep Learning model for detection of Pneumonia and Covid-19 variants from Chest X-ray images

Pneumonia is an example of a past pandemic and continues to be a serious health concern. In the USA, more than one million people are admitted in hospital with pneumonia every year, leading to about 500,000 deaths. Chest X-ray imaging is an effective and widely utilised method for diagnosing pneumonia and is essential in both healthcare and epidemiological studies. COVID-19, a viral infection initiated in Wuhan, China towards the end of 2019, quickly spread across the globe. It is caused by the Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) and has influenced millions globally. Analyzing X-ray images is regarded as the fastest and simplest methods for discovery, available at a minimal cost in many places. CT scans, on the other way, are a mere advanced imaging technique that can identify small changes in the composition of internal organs. This method uses 3-D computer technology along with X-rays for a more detailed examination. While both CT scans and X-rays provide images of internal body compositions, traditional X-ray images can sometimes occlude, making it difficult to see fine details. The proposed model outlines a framework for classifying COVID-19 variants and predicting new ones. As per the results, the proposed ResNet_Seg achieved an F1 score of 99.96%, which is higher than the CNN and other models tested. The performance of these models is assessed using datasets from SARS and MERS, resulting in accurate predictions. Future work will focus on validating these models using statistical methods. A relative analysis of deep learning models, including CNN, ResNet, and Darknet, is conducted, with performance enhancements achieved through the novel segmentation algorithm and hyperparameter fine-tuning. The results offer insights into developing more effective and reliable diagnostic methodologies for pneumonia and COVID-19 using deep learning and machine learning techniques.