Evaluation of a Landmark Curve Matching Technique in Brain Mapping based on the Level Set Method

Introduction The level set method using implicit representation has recently been developed and shown to be capable of matching all types of geometric objects. The method was evaluated in this study for its ability to perform brain surface warping subject to landmark curve constraints. This key step in brain mapping allows multiple subjects’ data to be compared after adjusting for gyral differences. Other methods that match equidistant landmark points on the curves via point matching do not allow relaxation along the curves. Moreover, it is difficult to incorporate these techniques into other methods due to the non-variational nature of point constraints. The method we use incorporates the level set method into image warping, offering a unifying approach for different types of feature-based matching. A diffeomorphic, one-to-one, and onto mapping can be generated using this approach subject to different kinds of feature-based constraints.