Segmentation of the colon in magnetic resonance images

Abstract

The use of magnetic resonance images for virtual colonoscopy is a relatively new method for the prevention of colorectal cancer. Unlike computed tomography, magnetic resonance technology does not use ionizing radiations and offer a good contrast for soft tissues. However, the processing of magnetic resonance images is a very challenging issue due to noise and inhomogeneities. One of the major steps in the processing of images is called segmentation and consists of isolating the colon in images. In this work, we compare two methods for the segmentation of 2D MR images. The first method is based on thresholding algorithms and anisotropic diffusion as a preprocessing step. The second method uses B-snakes. Based on 30 images from two different datasets, we find that both B-snakes and thresholding algorithms give accurate results in94% of the cases. The segmented regions obtained with B-snakes have a more regular aspect than those obtained with thresholding algorithms.