Adjoint Methods for Image Registration
Speaker: Marc Niethammer , UNC
Date: March 31 2011
Time: 11:30AM to 12:30PM
Location: 32-D507
Host: Polina Golland, CSAIL
Contact: Polina Golland, x38005, polina@mit.edu
Relevant URL: Image registration is an important component for many medical image analysis
methods. It is a means to establish spatial correspondences within or across
subjects and a fundamental building block for example for atlas-building.
Most state-of-the art registration methods require the solution of an
optimization problem. This talk will discuss adjoint solution methods. It
will introduce the basic concepts with point-based toy examples. The talk
will subsequently illustrate how the same concepts allow for simple
formulations (and the expression of optimality conditions) of standard
fluid-flow image-registration, of the registration of image time series, of
registrations between images without perfect correspondence, and of
higher-order image registration models.
See other events that are part of Biomedical Imaging and Analysis 2010/2011
See other events happening in March 2011
