ResearchBib Share Your Research, Maximize Your Social Impacts
Sign for Notice Everyday Sign up >> Login

Multiphase Level Set Method for Image Segmentation in the Presence of Intensity Inhomogenity with the Help of Brain MRI Images

Journal: International Journal of Science and Research (IJSR) (Vol.4, No. 9)

Publication Date:

Authors : ; ;

Page : 276-279

Keywords : Level set; image segmentation; bias field; intensity Inhomogenity; MR image;

Source : Downloadexternal Find it from : Google Scholarexternal

Abstract

Image segmentation, in real world images is a considerable challenge in the presence of Intensity Inhomogenity. The widely used Image segmentation algorithms are region based and they depend on homogeneity of the image intensities in the region of interest. This method often fails to provide accurate segmentation results due to the non uniformity of intensities. This paper proposes a novel region based method for image segmentation, which is able to deal with intensity non uniformities in the segmentation of brain MRI images. First we derive a local intensity clustering criterion function for the image intensities, based on the model of images with intensity Inhomogenity. This local clustering criterion function is integrated with respect to the neighbourhood centre to give a global clustering criterion function. In a multiphase level set formulation this criterion function defines an energy in terms of the level set functions that represents a partition of the image domain. The multiphase level set method is mainly used to represent triple junctions in the image. A bias field which accounts for the intensity non uniformity is identified. Therefore, by minimizing the energy our method can simultaneously segment the image and estimate the bias field. For intensity non uniformity correction the above estimated bias field is used. In the presence of intensity inhomogenity our method has been validated on brain MRI images of various modalities, with desirable performance. Experimental results shows that our method is robust to contour initialization, faster and more accurate than existing methods

Last modified: 2021-06-30 21:53:24