Project Description:
A new approach using the Beltrami Signature of a
shape to segment an object from an image with promised topology is proposed in
this paper. Given a target image $I$, a template image $J$ called the
topological prior image is deformed according to the newly proposed Beltrami
Signature, such that the topology of the segmented region is preserved as that
of the object interior in $J$. The topology preserving property of the
deformation is guaranteed by imposing only one constraint on the Beltrami
Signature, which is easy to be handled. Introducing the Beltrami Signature also
allows large deformations on the topological prior $J$, so that it can be a very
simple image, such as an image of disks, torus, disjoint disks, etc. Hence,
prior shape information of $I$ is unnecessary for the proposed model.
Additionally, the proposed model can be easily incorporated with selective
segmentation, in which landmark constraints can be imposed interactively to
facilitate any practical need (e.g. medical imaging). High accuracy and
stability of the proposed model to deal with different segmentation tasks is
validated by numerical experiments on both artificial and real images.
Publication: