Parametric volumetric registration

Volumetric registration (VR) is one of the most studied problems in the medical image analysis field. This chapter begins by highlighting common medical image analysis problems that require VR (e.g., within-subject longitudinal registration, fusion of imaging modalities during image-guided surgery, or groupwise normalization for fMRI analysis) and describing the basic elements of the VR pipeline, such as spatial transformations, dissimilarity metrics and regularization terms. The latter part of the chapter focuses on registration approaches that parameterize spatial transformations using a limited set of coefficients, including rigid/affine transformations and B-Spline transformations. The chapter closes with a discussion of real-world considerations for implementing parametric VR.