The goal of image registration is the alignment of two or more images of the same scene or object. It is one of the most widely encountered problems in a variety of fields including medical image analysis, remote sensing, satellite imaging, optical imaging, etc.
This thesis presents a novel, unified, generic and variational framework for seamlessly integrating prior segmentation information into non-rigid registration procedures. Under this framework, in addition to the forces arising from the similarity measure in seeking a detailed correspondence, another set of forces generated by the prior segmentation contours can provide an extra guidance in assisting the alignment process towards a more meaningful, stable and noise-tolerant procedure. Local Correlation (LC) is being used as the underlying similarity measure to handle intensity variations. We present several examples on 2D/3D synthetic and real data.