One can get COVID-19 and develop severe illness at any age. The symptoms include cold or fever, cough, breathlessness or difficulty breathing, fatigue, body ache or muscle pain, headache, and so on. In the absence of vaccine, prompt isolation and medical attention are crucial for controlling and preventing the COVID-19 pandemic. Since the lack of detection kits and the daily increase in the number of cases worldwide, it is difficult to detect the presence of disease. Consequently, the goal of this paper is to create an automated tool to diagnose COVID-19 from a binary dataset of X-ray images using traditional neural network in machine learning (ML) models and pre-trained deep learning (DL) architectures. The lack of a large number of datasets was overcome by using DL-based ResNet50 and ResNet101 to extract dataset features for more effective classification.