Image registration with sliding motion

摘要: Abstract An in-depth introduction and thorough discussion of current approaches for medical image registration with sliding motion is presented in this chapter. Several strategies locally-adaptive regularization past, future work are described including related research from optical flow computer vision. In particular recent advances to the Demons framework discrete optimization that do no require any specific segmentation masks led substantial improvements over baseline approaches. A reduction target error respect expert landmarks visually plausible computed fields can be reached using these methods. The great clinical impact a suitable handling discontinuities highlighted directions towards advanced graph-based edge priors through supervised learning reader.