Image Processing Techniques for COVID-19 Detection in Chest CT

Abstract

COVID-19 virus is a disease that has spread around the world recently. Early diagnosis of the disease leads us to the opportunity to treat patients faster to reduce its spread in the community. The CT scan image is one of the routines used to diagnose the COVID-19 diseases in an efficient manner and in a faster time. Fractal Dimension which mean (fragmented or irregular) used in wide range of image processing and analysis applications to get the self- similarity of images. To classify textures, combine images, segment and compress images and to generate incredibly complex and good-looking images, the Fractal Dimension method is used. Moreover, Euler method uses features to explain the structural property caused by noise in binary images. It also describes the topological features and analyze the texture of images. In this paper, a method is proposed to obtain the features of CT scan images for COVID-19 by using a hybrid technique called Fractal Dimension Euler (FDE), which merges the two methods of image processing (Fractal Dimension method and Euler Number Method). The two algorithms aim to segment the CT-scan images for the chest to distinguish between the affected and uninfected area of the chest to detect the COVID-19. The results of the proposed approach were very useful in comparison with another approach, the FD method was applied to CT scan images for COVID- 19 using a method called box counting. After that, the Euler method was used to distinguish between foreground and background by using a threshold value. The best threshold value was (255) which achieved the finest result.