matchedLandmarks: Convert match output to landmarks in physical space

In stnava/patchMatchR: Local patch matching with flexible transformation models in n-dimensional images

Convert match output to landmarks in physical space

Description

Convert match output to landmarks in physical space

Usage

matchedLandmarks(
  matchObject,
  referenceImage,
  movingImage,
  patchSize,
  whichK = 1
)

Arguments

matchObject	object, the output of deepPatchMatch
referenceImage	the fixed image
movingImage	the image that will be matched to the fixed image
patchSize	size of patch features
whichK	which matched point set (e.g. 1 gives the best, 2 second best and so on)

Value

output list contains fixed and matched points

Examples

## Not run: 
library( keras )
library( ANTsR )
layout( matrix(1:2,nrow=1) )
nP1 = 50
nP2 = 200
psz = 32
img <- ri( 1 ) %>% iMath( "Normalize" )  %>% resampleImage( c( 2, 2 ) )
img2 <- ri( 2 ) %>% iMath( "Normalize" )  %>% resampleImage( c( 2, 2 ) )
mask = randomMask( getMask( img ), nP1 )
mask2 = randomMask( getMask( img2 ), nP2 )
matchO = deepPatchMatch( img2, img, mask, mask2 )
mlm = matchedLandmarks( matchO, img, img2, c(psz,psz) )
ct = 0
mxct = 18
lmImage1 = img * 0
lmImage2 = img2 * 0
for ( k in 1:nrow( mlm$fixedPoints ) ) {
  if ( ct < mxct ) {
    pt1i = makePointsImage( matrix(mlm$fixedPoints[k,],ncol=2), img, radius = 2 ) * k
    pt2i = makePointsImage( matrix(mlm$movingPoints[k,],ncol=2), img2, radius = 2 ) * k
    lmImage1 = lmImage1 + pt1i
    lmImage2 = lmImage2 + pt2i
    }
  ct = ct + 1
 }

plot( img, lmImage1 )
plot( img2, lmImage2 )

## End(Not run)

stnava/patchMatchR documentation built on March 23, 2022, 6:47 a.m.