Robust Non-Rigid Registration of Medical Images
Chen-Lun Lin · Institutional Repositories DataBase (IRDB) · 2015
Recently, image registration process plays an important role in the analysis of medical image at medical treatment.Image registration has been widely applied at various fields such as surgical navigation, serial-image analysis, medical image fusion, etc... Image registration can be classified into two categories such as rigid and non-rigid image registration which depend on the objective of application classify.. Rigid registration is usually performed to match two rigid objects.Relatively, non-rigid registration is generally applied at to align two deformable organs for medical image analysis.The main difference between rigid and non-rigid method is transformation which is used by separately.In practice, method of global transformation such as rigid transformation is used for rigid image registration.In additions, some deformable local transformation methods such as affine transformation, B-spline transformation are applied for non-rigid registration also.This research of thesis is focused on research findings to explore for robust and achieve accurate non-rigid medical image registration and also as my contribution of this research.Non-rigid registration is a process for maximizing a spatial image correspondence of two images within constrains of a transformation model.The process of registration can be divided into four phases, which are: transformation, interpolation, criterion, and optimization.In this thesis, the author proposed three novel approaches for robust and accurate medical image registration.Two of them are focused to improve similarity metric for the phase of criterion; another is to concentrate on an advance optimization method for the purpose of to enhance optimization.In addition, a well-developed system of non-rigid image registration for assessing quality of loco regional therapy of hepatocellular carcinoma.It is also described in this research.The major contributions of this thesis are summarized as follows: