Chest X-Ray-Based Covid-19 Detection Using Preprocessing Techniques and Deep Learning Algorithms

Covid-19 is a transmissible infection identified in Wuhan, China, in December 2019 where more than million people were killed. When the contamination enters body, it contacts the mucous film that lies in your mouth, eyes, and nose. The virus entering a cell makes new disease parts making use of the cell and spreads infections by attacking nearest cells. Signs could begin one to fourteen days subsequent to presenting to the infection. Majority of them suffer with pneumonia, while a few suffer with extreme side effects like shortness of breath, headache, and fast heartbeat. Medical practitioners can see indications of respiratory infections on a chest X-ray. So, to detect if a person is suffering with Covid-19, the chest images help in the early analysis and the treatment of the patient. This study utilizes multiple datasets containing images of chest X-ray and makes use of preprocessing techniques like image augmentation, resizing, Contrast-Limited Adaptive Histogram Equalization (CLAHE) and trained using Deep learning algorithms like CNN, VGG-16, VGG-19, and AlexNet. CNN model with a maximum pooling layer and a flatten layer outperformed other algorithms when trained on image augmentation and normalization images and achieved accuracy of 0.98. Finally, integrating the best suitable model with web applications like Flask helps to reduce the burden on doctors and minimize the human error in Covid-19 detection.