3D shape reconstruction based on given sparse 3D landmarks (Posterior Shape Model)

[nawazishkhan1-nk]

Reconstruct the shape model through posterior inference, given a set of sparse 3D Landmarks.

The landmarks can be given as in the following two usage scenarios:

1. There exists a set of landmarks for each subject in the shape model; Given a new shape with sparse landmarks, we need to estimate all the particle locations for it conditioned on those landmarks through posterior inference of the SSM.
2. Only a new shape is given with landmarks defined on it; This requires first estimating the corresponding particle location in the shape model for those landmarks and then deriving the posterior model.