2011 Summer Project Week normal consistency particles

• 

  Projects List

• 

  Fig1. Particle Correspondence has problems with regard to representing thin areas correctly, and thus with reconstructing them

• 

  Fig2. First attempts into incorporating normal directions into Particle Correspondence

Full Title of Project

Key Investigators

• University of North Carolina at Chapel Hill: Beatriz Paniagua, Sungkyu Jung, Martin Styner, Steve Pizer
• University of Utah: Josh Cates, Manasi Datar, Ross Whitacker

Objective

A well established method for performing statistics on an ensemble of shapes is to compare corresponding landmarks placed automatically on the individual shapes. Particle ensemble-systems, designed initially by Cates et all [1], has sometimes problems to correctly represent medical or biological shapes that have sharp features and regions of high curvature. Normal directions should be consistent in each individual particle during the optimization, that should try to minimize normal entropy. The problem of computing entropy in normals is that they are not in Euclidean space so a dimensionality reduction method is needed. Jung et al [2] created a dimension reduction and data visualization for highdimensional data method that can be applied to entropy computing in normals.

Our objective is incorporate a new entropy term into [1] based in [2] that accounts for normal consistency during landmark positioning.

Approach, Plan

1. Coding-debugging in collaboration with Josh Cates and Manasi Datar.
2. Revise the engineering details of the code with Josh Cates.
3. (hopefully) Test the code to asses the validity of the method.

Progress

References

Delivery Mechanism

This work will be delivered to the NAMIC Kit as a

1. NITRIC distribution
2. Slicer Module
   1. Built-in:
   2. Extension -- commandline:
   3. Extension -- loadable: