R/surgeRy.R
generateDiskPointAndSegmentationData.RdThis assumes that at least images and point sets are passed in - segmentation images are optional. The indexing of point sets will be whichSimulation by whichPoint by dimensionality (e.g. 10 x 25 x 3 for 10 simulations of point sets with 25 points in 3D.
generateDiskPointAndSegmentationData( inputImageList, pointsetList, segmentationImageList, segmentationNumbers, selector, maskIndex, smoothHeatMaps = 0, numpynames, cropping = NULL, numberOfSimulations = 16, referenceImage = NULL, transformType = "rigid", noiseModel = "additivegaussian", noiseParameters = c(0, 0.002), sdSimulatedBiasField = 5e-04, sdHistogramWarping = 5e-04, sdAffine = 0.2 )
| inputImageList | list of lists of input images to warp. The internal list sets contains one or more images (per subject) which are assumed to be mutually aligned. The outer list contains multiple subject lists which are randomly sampled to produce output image list. |
|---|---|
| pointsetList | list of matrices containing point sets where each matrix is of size n-points by dimensionality matching to input segmentation/image data (optional) |
| segmentationImageList | of segmentation images corresponding to the input image list (optional) |
| segmentationNumbers | the integer list of values in the segmentation to model |
| selector | subsets the input lists (eg to define train test splits) |
| maskIndex | the entry within the list of lists that contains a mask |
| smoothHeatMaps | numeric greater than zero will cause method to return heatmaps.
the value passed here also sets the smoothing parameter passed to |
| numpynames | the names of the numpy on disk files should include something with the word pointset. Also: the string segmentation if using segmentations and the string mask if using maskIndex and something with the word coordconv if using CC. should include something with the word heatmaps if using heatmaps. |
| cropping | a vector of size 1 plus image dimensionality where the first
parameter indicates which landmark to target and the trailing parameters define
the size of the cropping patch. e.g. |
| numberOfSimulations | number of output images/pointsets. Default = 10. |
| referenceImage | defines the spatial domain for all output images. If
the input images do not match the spatial domain of the reference image, we
internally resample the target to the reference image. This could have
unexpected consequences. Resampling to the reference domain is performed by
testing using |
| transformType | one of the following options
|
| noiseModel | one of the following options
|
| noiseParameters | 'additivegaussian': |
| sdSimulatedBiasField | Characterize the standard deviation of the amplitude. |
| sdHistogramWarping | Determines the strength of the bias field. |
| sdAffine | Determines the amount of transformation based change |
list of array
Avants BB
library( reticulate ) library( ANTsR ) library( ANTsRNet ) library( surgeRy ) image1 <- antsImageRead( getANTsRData( "r16" ) ) image2 <- antsImageRead( getANTsRData( "r27" ) ) segmentation1 <- thresholdImage( image1, "Otsu", 3 ) segmentation11 = thresholdImage( segmentation1, 1, 1 ) segmentation12 = thresholdImage( segmentation1, 2, 2 ) segmentation13 = thresholdImage( segmentation1, 3, 3 ) segmentation11[1:128,1:256]=0 segmentation12[1:256,1:180]=0 segmentation13[1:256,1:128]=0 segmentation1 = segmentation11 + segmentation12* 2 + segmentation13 * 3 segmentation2 <- thresholdImage( image2, "Otsu", 3 ) segmentation21 = thresholdImage( segmentation2, 1, 1 ) segmentation22 = thresholdImage( segmentation2, 2, 2 ) segmentation23 = thresholdImage( segmentation2, 3, 3 ) segmentation21[1:128,1:256]=0 segmentation22[1:256,1:180]=0 segmentation23[1:256,1:128]=0 segmentation2 = segmentation21 + segmentation22* 2 + segmentation23 * 3 pts1 = getCentroids( segmentation1 )[,1:2] pts2 = getCentroids( segmentation2 )[,1:2] plist = list( pts1, pts2) ilist = list( list( image1 ), list( image2 ) ) slist = list( segmentation1, segmentation2 ) npn = paste0(tempfile(), c('i.npy','pointset.npy','heatmap.npy','coordconv.npy','segmentation.npy') ) temp1 = generateDiskPointAndSegmentationData( ilist, plist, slist, segmentationNumbers = 1:3, numpynames = npn ) temp2 = generateDiskPointAndSegmentationData( ilist, plist, segmentationNumbers = 1:3, numpynames = npn, smoothHeatMaps = 3 ) temp = generateDiskPointAndSegmentationData( ilist, plist, slist, segmentationNumbers = 1:3, numpynames = npn, smoothHeatMaps = 3 ) locimg=as.antsImage( temp$images[1,,,1] ) locseg=as.antsImage( temp$segmentation[1,,,3] ) locseg2=as.antsImage( temp$segmentation[2,,,3] ) locimg2=as.antsImage( temp$images[2,,,1] ) # layout(matrix(1:4,nrow=1)) # plot(locimg,locseg) # plot(locimg,as.antsImage( temp$heatmaps[1,,,3])) # plot(locimg2,locseg2) # plot(locimg2,as.antsImage( temp$heatmaps[2,,,3])) print(getCentroids( thresholdImage(as.antsImage( temp$heatmaps[1,,,1]),0.5,1) ))#> x y z t #> [1,] 159 133 0 0#> x y z t #> [1,] 127 164 0 0#> [1] 159.3520 132.8819#> [1] 126.5328 163.7855mm = makePointsImage( temp$points[1,,], getMask( locimg ) ) mm2 = makePointsImage( temp$points[2,,], getMask( locimg2 ) ) gg1 = thresholdImage(as.antsImage( temp$heatmaps[1,,,1]),0.5,1) %>% antsCopyImageInfo2(ri(1)) gg2 = thresholdImage(as.antsImage( temp$heatmaps[1,,,2]),0.5,1) %>% antsCopyImageInfo2(ri(1)) gg3 = thresholdImage(as.antsImage( temp$heatmaps[1,,,3]),0.5,1) %>% antsCopyImageInfo2(ri(1)) # plot( locimg, mm ) # plot( locimg, gg1*1+gg2*2+gg3*3 )