Brain MRI Tissue Classification using Graph Cut Optimization of the Mumford-Shah Functional
Noha Youssry El-Zehiry, Adel Said Elmaghraby · 2007
In [5], we have introduced a Graph Cut Based Level Set (GCBLS) formulation that incorporates graph cuts to optimize the curve evolution energy function presented earlier by Chan and Vese. In this paper, we propose to extend our previous approach for multiphase image segmentation, since the problem of separating n classes using graph cuts is an NP complete problem, we propose to do the multiphase evolution in a sequential manner in which multipahse evolution will be implemented through a sequence of bimodal evolutions. To summarize, the novelty of this paper lies in introducing a sequential graph cut optimization technique for the multiphase image segmentation. The major advantages of this technique are; 1) The accuracy of the segmentation is improved because graph cuts solves for a global minimum rather than a local one. 2) The speed of the segmentation is much better than the speed of most of the state of the art segmentation techniques in the literature. Yet, the algorithm still preserves all the advantages of level set framework.