createBayes: create an object of class BayesDeform to be used in...

In zarquon42b/mesheR: Meshing Operations on Triangular Meshes

create an object of class BayesDeform to be used in gaussMatch and AmbergRegister

Description

create an object of class BayesDeform to be used in gaussMatch and AmbergRegister

Usage

createBayes(
  model,
  sdmax = numeric(0),
  mahalanobis = "chisq",
  ptValueNoise = 2,
  wt = NULL,
  align = TRUE,
  shrinkfun = NULL,
  initparams = NULL
)

Arguments

model	a model of class pPCA created with package RvtkStatismo
sdmax	numeric or vector of numerics (length must be of value =< iterations as to be specified in gaussMatch or AmbergRegister. If length(sd) =n < iterations, then the first n iterations will be restricted to the according standard deviations
mahalanobis	parameter according to mahaprob in function RvtkStatismo::PredictSample from RvtkStatismo
ptValueNoise	parameter according to ptValueNoise in function RvtkStatismo::statismoConstrainModel from RvtkStatismo
wt	numeric: a weight put on the closest model configuration. E.g. if wt = 2 the deformation will be 0.33*closestPoints + 0.66*most probable model shape. This can either be a constant scalar (that might be modified in each iteration by a shrinkage function - see below) or a vector of length sdmax.
align	logical: if the target mesh is already aligned to the model, align can be set to FALSE.
shrinkfun	a two-valued function for the generation of the weight in the i-th iteration. See examples below.
initparams	initial estimation of the model parameters

Value

returns a list of class "BayesDeform"

Examples

## Not run: 
shrinkfun <- function(x,i){x <- x*0.9^i; return(x)}
# for wt=0.5 this will produce 
# 

## End(Not run)

zarquon42b/mesheR documentation built on Nov. 14, 2024, 3:30 p.m.