Detection of COVID-19 Using Convolutional Neural Networks

A highly contagious illness caused by the COVID-19 pandemic is proven to havoc with people's health and well-being all over the globe. Chest radiography is one of the most crucial databases for applying detection techniques. The respiratory system is contaminated by COVID-19, which also affects the alveoli and replicates itself. Conventional approaches such as RT - PCR tests, rapid antigen tests, serological tests, etc. are generally used for the detection of COVID came out to be costly and time-consuming. There have been several suggested artificial intelligence (AI)-based models for the detection of COVID-19 in contaminated individuals' using lung ultrasound images, voice patterns, chest sounds, etc. The proposed model shows how disease cases could be identified using features and variations in chest radiography images. A deep convolutional neural network (CNN) with ResNet50 modified configuration has been used for the identification of COVID-19 cases from the chest radiography image dataset. The suggested model also depicts the comparison of the proposed technique with the Resnet 101 model. Chest radiography dataset containing Covid infected, normal, and pneumonia-infected images. Additionally, the suggested model can identify covid-19 patients' current conditions in real-time by identifying coronavirus diseases from CT scan pictures. The database has the ability to monitor detected patients and keep their databases in order to improve the training and model's accuracy. The proposed model provides approximately 96.73% of accuracy and shows explicit competency with ResNet 101 and other existing models.