Abstract
The lattice Boltzmann (LB) method is a mesoscopic method based on kinetic theory and statistical mechanics. The main advantage of the LB method is parallel computation, which increases the speed of calculation. In the past decade, LB methods have gradually been introduced for image processing, e.g., image segmentation. However, a major shortcoming of existing LB methods is that they can only be applied to the processing of medical images with intensity homogeneity. In practice, however, many medical images possess intensity inhomogeneity. In this study, we developed a novel LB method to integrate edge and region information for medical image segmentation. In contrast to other segmentation methods, we added edge information as a relaxing factor and used region information as a source term. The proposed method facilitates the segmentation of medical images with intensity inhomogeneity and it still allows parallel computation. Preliminary tests of the proposed method are presented in this paper.