The Segmentation of Perianal Abscess Lesion Area via Gradient Operator of CT Images

Abstract

Perianal abscess is a high incidence of anorectal diseases, and simply surgical treatment could result in the removal of unnecessary tissue, and precise segmentation could solve this problem. In this work, we find a method based on the gradient of the medical image to conquer this problem. Firstly, the edge operator is used to get the gradient of the original image, and by setting the preianal area and taking the gradient again, we could get the edges in the perianal area. Then, by the symmetry of the preianal area, we could get the closed polygon of the lesion area. And by smoothing the polygon, we could finally obtain the closed curve of the lesion area. As far as we know, our method is the first to apply image gradient and mathematical methods to the field of image segmentation of the perianal abscess and provide an auxiliary calibration method for medical image data.