A semi automatic geometric active contour model using Distance Regularized term for segmentation of abdominal organs on CT images
S. M. AnchaloBensiger, Subbiahpillai Neelakantapillai Kumar, Alfred Lenin Fred, S. Lalitha Kumari · 2016
Level set models are widely used in image processing and computer vision for segmentation. In this paper, an improved geometric active contour model is used for the segmentation of abdominal organs in abdomen CT images. The pre-processing of input images was done by anisotropic diffusion filter that effectively preserve the edges. The proposed Distance Regularized Level set Evolution (DRLSE) doesn't requires initialization procedure unlike the conventional level set methods. The distance regularized term is defined using double well potential function such that the level set evolution has unique forward and backward diffusion (FAB) diffusion effect, which is able to maintain a desired shape of the level set function. The algorithms were developed in Matlab 2010 and tested on real time CT data sets.