wrap.landmark: Wrap Landmark Data on Shape Space

In Riemann: Learning with Data on Riemannian Manifolds

Wrap Landmark Data on Shape Space

Description

One of the frameworks used in shape space is to represent the data as landmarks. Each shape is a point set of k points in \mathbf{R}^p where each point is a labeled object. We consider general landmarks in p=2,3,…. Note that when p > 2, it is stratified space but we assume singularities do not exist or are omitted. The wrapper takes translation and scaling out from the data to make it preshape (centered, unit-norm). Also, for convenience, orthogonal Procrustes analysis is applied with the first observation being the reference so that all the other data are rotated to match the shape of the first.

Usage

wrap.landmark(input)

Arguments

input

data matrices to be wrapped as riemdata class. Following inputs are considered, array a (k\times p\times n) array where each slice along 3rd dimension is a k-ad in \mathbf{R}^p. list a length-n list whose elements are k-ads.

Value

a named riemdata S3 object containing

data

a list of preshapes in \mathbf{R}^p.

size

size of each preshape.

name

name of the manifold of interests, "landmark"

References

\insertRef

dryden_statistical_2016Riemann

Examples

## USE 'GORILLA' DATA
data(gorilla)
riemobj = wrap.landmark(gorilla$male)

Riemann documentation built on March 18, 2022, 7:55 p.m.