Multi Class Image Classification for Detection Of Diseases Using Chest X Ray Images

A Novel Coronavirus (Sars-Cov-2) struck the world in December, 2019. First Detected in Wuhan, China: this acute respiratory syndrome has spread all over the world at the present moment and has been officially declared as a global pandemic. A massive detrimental effect on global health and economy has been noticed. While researchers are continuously in search of vaccines - detection and proper diagnosis of the virus is as important to limit the spread of the virus. Chest X-Rays (CXRs) is one of the most common types of radiology examination and CXRs of the infected patients can serve as a crucial step in detection of the virus. Having a computer aided automatic diagnosis can minimize human interactions, errors, and workload and maximize efficiency. Various studies have shown that use of artificial intelligence in detection of Covid-19 patients through their CXRs is strongly optimistic. In this paper, a robust and efficient computer aided detection system has been proposed for multiclass image classification of diseases like Covid-19 and Pneumonia using the CXRs of patients. The algorithms have currently achieved desired results which can be further improved when more CXR images are available. The proposed method has outperformed current state of the art algorithms and has achieved 98.3% accuracy with a precision metric of 0.94, and can be used as a fast and reliable preliminary test for detection of the virus.