Topology Aware Registration - KCL-BMEIS/Methods_JournalClub GitHub Wiki
Presented by Samuel
14th of September 2019
TopAwaRe: Topology-Aware Registration
Rune Kok Nielsen et al. (2019)
Deformable registration, or nonlinear alignment of images, is a fundamental preprocessing tool in medical imaging. State-of-the art algorithms restrict to diffeomorphisms to regularize an otherwise ill-posed problem. In particular, such models assume that a one-to-one matching exists between any pair of images. In a range of real-life-applications, however, one image may contain objects that another does not. In such cases, the one-to-one assumption is routinely accepted as unavoidable, leading to inaccurate preprocessing and, thus, inaccuracies in the subsequent analysis. We present a novel, piecewise-diffeomorphic deformation framework which models topological changes as explicitly encoded discontinuities in the deformation fields. We thus preserve the regularization properties of diffeomorphic models while locally avoiding their erroneous one-to-one assumption. The entire model is GPU-implemented, and validated on intersubject 3D registration of T1- weighted brain MRI. Qualitative and quantitative results show our ability to improve performance in pathological cases containing topological inconsistencies.
Discussion Points
- How the surface is created to do the diffeomorphic part of the framework?
- Are the ventriculus the most interesting example for this approach?
Pros
The current practice in image registration is to accept the incorrect diffeomorphic assumption as unavoidable. They believe that, through additional specialization, the model may provide a practical foundation for further improving the registration of pathological tissue.
Cons
- The framework is not symmetric. Specifically, the topology-altering model allows expansion from discontinuities in the source image but not vice-versa.
Applications to MI field
- Registration of soft tissues, namely after resections;
- Tumour resection;
- Epilepsy treatments -> surgery;
- Any registration that needs to account for a change of topology.