posteriorDeform: Fit an SSM to a target based on subsampling corresponding...

In zarquon42b/mesheR: Meshing Operations on Triangular Meshes

Fit an SSM to a target based on subsampling corresponding points

Description

Fit an SSM to a target based on subsampling corresponding points and compute the posterior mean

Usage

posteriorDeform(
  model,
  target,
  reference = NULL,
  partsample = NULL,
  samplenum = 1000,
  distance = 1e+10,
  slide = 3,
  bending = TRUE,
  ray = FALSE,
  deform = FALSE,
  Amberg = FALSE,
  rhotol = pi/2,
  modlm = NULL,
  tarlm = NULL,
  align2mod = TRUE,
  alignbymesh = FALSE,
  forceLM = FALSE,
  silent = FALSE,
  threads = 1,
  ...
)

Arguments

model	statismo shape model
target	target mesh
reference	model instance other than the mean (class mesh3d)
partsample	predetermined corresponding points on the sample mean
samplenum	integer: if partsample=NULL, this specifies the number of coordinates sampled on the model mean
distance	numeric: constrain maximum distance to mark target point as appropriate
slide	integer: if > 0 the valid correspondences on the model instance will be relaxed minimizing bending energy/procrustes distance.
bending	logical: if TRUE, the coordinates on the model instance are relaxed using bending energy, Procrustes distance otherwise
ray	logical: if TRUE, the closest point search will be performed along the normals only
deform	logical if TRUE, the posterior mean will also be deformed to the target using an elastic deformation
Amberg	if TRUE the deformation will use the function AmbergDeformSpam and tps3d otherwise
rhotol	maximal tolerated angle between normals to be considered a valid match
modlm	matrix containing 3D landmarks on the model mean (not for alignment)
tarlm	matrix containing 3D landmarks on the target surface
align2mod	logical: if TRUE, the prediction step will perform an alignment to the model using the valid correspondences.
alignbymesh	logical: if TRUE, the alignment to the SSM will be computed by the entire mesh, if FALSE only the valid correspondences are used.
forceLM	if TRUE, predfined landmarks modlm are not allowed to slide. For cases with high uncertainty, this can lead to unwanted mesh distortions.
silent	logical: supress debug output
threads	integer: number of threads to use for tps interpolation (set to 1 if using openblas, or otherwise it can become instable)
...	additional parameters passed to AmbergDeformSpam.

Value

returns a deformed version of a model instance fitted to the target

Note

Please note that it is required to align the target mesh to the model mean beforehand. This can be performed using the function icp, for example.

Examples

## Not run: 
require(RvtkStatismo)
download.file(url="https://github.com/marcelluethi/statismo-shaperegistration/raw/master/data/VSD001_femur.vtk","./VSD001_femur.vtk",method = "w")
download.file(url="https://github.com/marcelluethi/statismo-shaperegistration/raw/master/data/VSD002_femur.vtk","./VSD002_femur.vtk",method = "w")
download.file(url="https://github.com/marcelluethi/statismo-shaperegistration/raw/master/data/VSD001-lm.csv","./VSD001-lm.csv",method = "w")
download.file(url="https://github.com/marcelluethi/statismo-shaperegistration/raw/master/data/VSD002-lm.csv","./VSD002-lm.csv",method = "w")
ref <- read.vtk("VSD001_femur.vtk")
tar <- read.vtk("VSD002_femur.vtk")
ref.lm <- as.matrix(read.csv("VSD001-lm.csv",row.names=1,header = FALSE))
tar.lm <- as.matrix(read.csv("VSD002-lm.csv",row.names=1,header = FALSE))
Kernels <- SumKernels(GaussianKernel(50,50),IsoKernel(0.1,ref))
mymod <- statismoModelFromRepresenter(ref,kernel=Kernels,ncomp=100)
postDef <- posteriorDeform(mymod,tar,modlm=ref.lm,tarlm = tar.lm,samplenum = 1000)
## run a loop redoing that step using the result of the previous step as input
for (i in 1:5)
   postDef <- posteriorDeform(mymod,tar,modlm=ref.lm,tarlm = tar.lm,samplenum = 1000,reference=postDef)

## now we leave the model space for a final deform involving a TPS deform
postDefFinal <- postDef
for (i in 1:3)
    postDefFinal <- posteriorDeform(mymod,tar,modlm=ref.lm,tarlm = tar.lm,samplenum = 3000,reference=postDefFinal,deform=T,distance=3)

Morpho::meshDist(postDefFinal,tar,from=-2,to=2,tol=.5)
rgl::wire3d(tar,col="white")

## End(Not run)

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