We present a graph-cuts based method for non-rigid medical image registration on brain magnetic resonance images. In this thesis, the non-rigid image registration problem is reformulated as a discrete labeling problem. According to a voxel-to-voxel similarity measure, each voxel in the source image is assigned a displacement label, which represents a displacement vector, indicating which position in the floating image it is spatially corresponding to. A smoothness constraint based on the first derivatives is used to penalize sharp changes in the displacement labels across voxels. The image registration problem is therefore modeled by two energy terms based on intensity similarity and smoothness of the displacement field. These energy terms are submodular and can be optimized by using th...[ Read more ]

We present a graph-cuts based method for non-rigid medical image registration on brain magnetic resonance images. In this thesis, the non-rigid image registration problem is reformulated as a discrete labeling problem. According to a voxel-to-voxel similarity measure, each voxel in the source image is assigned a displacement label, which represents a displacement vector, indicating which position in the floating image it is spatially corresponding to. A smoothness constraint based on the first derivatives is used to penalize sharp changes in the displacement labels across voxels. The image registration problem is therefore modeled by two energy terms based on intensity similarity and smoothness of the displacement field. These energy terms are submodular and can be optimized by using the graph-cuts method via alpha-expansions, which is a powerful combinatorial optimization tool and capable of yielding either a global minimum or a local minimum in a strong sense. Using the realistic brain phantoms obtained from the Simulated Brain Database, we have compared the registration results of the proposed method with the two state-of-the-art image registration approaches: the free-form deformation based method and the demons based method. It is found that the proposed method is more robust to different challenging non-rigid registration cases with consistently higher registration accuracy than those two methods, and gives realistic recovered deformation fields.