Improved Wells Method using MRF Radom Field and Non-local Weight
Zhongyuan Cui, Feng Wang · Advanced science and technology letters · 2014
In order to overcome the limitation of noise and intensity inhomogeneity during brain image segmentation, the MRF theory and the nonlocal information are used to construct a nonlocal Markov Random Field and to improve Wells et al. method. With our nonlocal MRF, the improved Wells method is robust to noise and intensity inhomogeneity. The experimental results demonstrate that our method is able to obtain better tissue segmentation results and simultaneously keep the image edge and slender topological structure very well.