Skip to content

References

Jump to bottom
Oscar Esteban edited this page Jun 13, 2013 · 4 revisions

Active contours / Active shape / Active Mesh / Levelsets

Shape priors

  • Bresson et al., A variational model of object segmentation. 2006
    • They include in the energy functional the shape model (based on PCA), the gradients and a term to asses homogeneous intensity
    • 2D
  • Chen et al., Levelsets based shape prior segmentation. 2005
    • Chan-Vese functional + shape term based on the heaviside function
    • 2D
  • Chen et al., Using prior shapes in geometric AC. 2002
    • The previous is a sequel of this one.
  • Cremers et al., Kernel density estimation and levelsets. 2006
    • Very good review of literature
    • AC+shape priors are usually expressed in terms of some shape-dissimilarity measure
    • Then this shape-dissimilarity can be solved with statistical shape priors
    • They propose for first a kernel density estimation for this.
    • 2D+t
  • Gastaud et al., Combining shape priors and statistical features for AC segmentation. 2004
    • propose a statistical distance w.r.t. prior shape.
    • 2D+t
  • Paragios, A levelsets approach for shape-driven segmentation and tracking. 2003
    • First proposal of the tracking part
    • Pixel-wise stochastic representation of the levelset ( mu, sigma)
    • 2D and 2D+t

Joint image registration and segmentation

  • Vermuri et al., Joint image registration and segmentation. 2003
    • Include a shape prior, with distance defined in terms of an affine transformation of the surface
    • This affine transformation is applied in registration, but show it only in the corpus callosum (and they don't evaluate trasformation or show the overall result).
    • Therefore, they don't have the field densification part, but it is probably the most similar thing to our work.
    • 3D
    • 9 parameters transform
  • Yezzi et al. (Zöllei included), A variational framework for segmentation and registration through AC. 2004
    • This is different, in the sense that the apply an affine transform to a moving image before computing the functional.
    • 9 parameters transform
    • No shape prior.
    • 3D
    • Give some evaluation report in terms of mean and deviation errors of registration

dMRI Segmentation

dMRI&fMRI Susceptibility distortion correction

dMRI-structural co-registration

dMRI-dMRI registration

Special issues in recognized journals

Clone this wiki locally