Segmentation of Abdominal Organs on CT Images using Distance Regularized Level Set Model-A Semi Automatic Approach

In image processing and computer vision, level set algorithms are generally used for segmentation. An improved geometric active contour model is used in this paper for the segmentation of abdominal organs in abdomen CT images. The input images were preprocessed by anisotropic diffusion filter that efficiently preserve the edges. The Distance Regularized Level Set Evolution (DRLSE) is used in this paper and it doesn't require reinitialization procedure unlike the conventional level set methods. The double well potential function was used to define the distance regularized term such that the level set evolution has unique forward and backward diffusion (FAB) effect. The algorithms were developed in Matlab 2010 and tested on real time CT data sets.