A hybrid edge-based technique for segmentation of renal lesions in CT images

Abstract

The entire community of medical experts uses various imaging techniques as the precursor for disease diagnosis with the assistance of computer-aided diagnosis systems. In many cases, these imaging techniques savored the status of pivotal proof of occurrence for tissue abnormalities. One of the most important steps in the analysis of tissue using medical images is the correct approximation of position, size, and shape of the lesion which plays a significant role to decrease false positives count for effective diagnosis of renal lesions. This article suggests a hybrid segmentation technique based on two methods which include spatial intuitionistic fuzzy c-means clustering (SIFCM) that integrates spatial image details and, distance regularized level-sets method for extraction of renal lesions correctly and proficiently in computed tomography (CT) images. The proposed technique works by taking an approximation of region of interest (ROI) given by Spatial IFCM clustering (SIFCM) for correct demarcation of lesions. Further, the performance of the suggested technique is tested on the considered image dataset and compared with the other state-of-the-art segmentation techniques such as thresholding, region growing, level set, fuzzy c-means clustering (FCM), active contour without edges (ACWE), geodesic active contours (GAC), spatial FCM and intuitionistic FCM. To confirm the segmentation results, the ground truth marked by the expert radiologists was considered as a gold standard for comparison. The experimental outcomes reveal that the suggested technique yields the results close to the manual delineations of experts as compared to the other considered segmentation techniques and is able to segment lesion correctly and precisely. The suggested technique attains the better lesion segmentation, even for images with low-contrast and in the presence of noise components. Furthermore, it possesses the capability to control the parameters adaptively from SIFCM clustering method.