Software @ CIS : principal-components : Validation

manual | faq | validation | credits | changelog | feedback

Content

Overview

This Validation section provides input data, processing and visualization examples for the Vector Kernel PCA for Meshes code to ensure correctness of the resultant data. These examples are useful tests when the program is run on new environments or platforms. Examples are displayed with available visualization software (CAWorks). A sample VectorKernelPCA command is posted with the example.

Caudate Analysis Example


Input: input.tar.gz
Output: output.tar.gz
Command: VectorKernelPCA 20 6.75 PCATestSurface.vtk vectorPcaOutput PCATestSurface_alpha0_*.vtk


The input data sets are momentum measurements produced by lddmm matching of each subject in the population against the population's template caudate surface. The momenta quantify the shape difference between the subjects in the population. The example calculates the first 20 principal components of the momenta. The first principal component captures the areas of greatest shape difference over the population.

The program produces 20 vtk PolyData files (meshes) that have the principal component values at each vertex stored as PointData. The following figures show the program output as displayed in CAWorks.

First Principal Component of Momenta on a Caudate

Last Modified: