Developing a novel deep learning approach to diagnosis COVID-19 disease using lung CT-scan images

This research aims to present a suitable method for feature extraction from chest CT scan images to enhance the accuracy and speed of detecting the coronavirus. In today's world, the role of engineering in medicine has significantly expanded, and with the advancement of imaging technology and image processing, disease diagnosis has become faster, easier, and more precise. The viral infection of COVID-19, which is originated in Wuhan, China and spread worldwide, has resulted in the death of over 4.4 million people, despite the initiation of vaccination efforts. Due to the high demand for PCR kits and their severe shortage, radiographic techniques such as X-rays and CT scans can be utilized for diagnostic purposes. Rapid detection of the coronavirus in the early stages can significantly prevent mortality from this devastating disease. Given the difficulty of diagnosing this disease in the early stages, providing a method that facilitates the early diagnosis of COVID-19 is highly valuable. In the proposed method, a modified two-dimensional convolutional neural network called Residual 2D-CNN is employed for more accurate and faster detection of COVID-19 from CT scan images. The final results of this research demonstrate an approximate 97% accuracy in detecting this virus.