This should work

DESCRIPTION

@@ -1,7 +1,7 @@
 Type: Package
 Package: LINDA
 Title: Lesion Identification with Neighborhood Data Analysis
-Version: 0.3.2
+Version: 0.4.0
 Authors@R: 
     c(
     person(given = "Dorian",
@@ -17,11 +17,14 @@ Authors@R:
 Description: A neuroimaging toolkit for automatic segmentation
     of chronic stroke lesions.
 License: Apache License | file LICENSE
+Depends: R (>= 2.10)
 Imports:
     ANTsR (>= 0.4.6),
     ANTsRCore (>= 0.6.0.1),
     randomForest,
-    magrittr
+    magrittr,
+    stats,
+    methods
 Suggests: 
     covr,
     knitr,

NAMESPACE

@@ -1,14 +1,14 @@
 # Generated by roxygen2: do not edit by hand
 
-export(LINDA.mrvnrfs.predict_chunks)
-export(LINDA.mrvnrfs_chunks)
+export(asymmetry_mask)
+export(linda_mrvnrfs.predict_chunks)
 export(linda_predict)
 export(n4_skull_strip)
 export(run_prediction)
+import(randomForest)
 importFrom(ANTsR,abpBrainExtraction)
 importFrom(ANTsR,abpN4)
 importFrom(ANTsR,composeTransformsToField)
-importFrom(ANTsR,mrvnrfs)
 importFrom(ANTsR,reflectImage)
 importFrom(ANTsR,splitMask)
 importFrom(ANTsRCore,antsApplyTransforms)
@@ -25,6 +25,10 @@ importFrom(ANTsRCore,lappend)
 importFrom(ANTsRCore,makeImage)
 importFrom(ANTsRCore,matrixToImages)
 importFrom(ANTsRCore,resampleImage)
+importFrom(ANTsRCore,resampleImageToTarget)
 importFrom(ANTsRCore,smoothImage)
 importFrom(ANTsRCore,thresholdImage)
 importFrom(magrittr,"%>%")
+importFrom(methods,formalArgs)
+importFrom(stats,median)
+importFrom(stats,predict)

R/asymmetry_mask.R

@@ -1,8 +1,20 @@
-asymmetry_mask = function(img,
-                          brain_mask,
-                          reflaxis = 0,
-                          sigma = 2,
-                          verbose = TRUE) {
+#' Asymmetry Mask
+#'
+#' @param img an \code{antsImage} of T1 image
+#' @param brain_mask an \code{antsImage} of brain mask
+#' @param reflaxis Reflection axis
+#' @param sigma Smoothing sigma
+#' @param verbose print diagnostic messages
+#'
+#' @return A list of the reflection and the mask
+#' @export
+#'
+asymmetry_mask = function(
+  img,
+  brain_mask,
+  reflaxis = 0,
+  sigma = 2,
+  verbose = TRUE) {
   # compute asymmetry mask
   print_msg("Computing asymmetry mask...", verbose = verbose)
 

R/getLesionFeatures.R

@@ -42,7 +42,6 @@ getLesionFeatures = function(
   #   kmask = bmask,
   #   verbose = verbose)$segmentation
   feats[[1]] = (conavg - kmean) %>% iMath('Normalize')
-
   # FEAT 2: gradient magnitude
   feats[[2]] = img %>% iMath('Grad') %>% iMath('Normalize')
 
@@ -60,6 +59,7 @@ getLesionFeatures = function(
     iMath(truncate, 0.01, 0.99) %>%
     iMath('Normalize')
 
+
   # FEAT 4: kmean
   feats[[4]] = antsImageClone(kmean)
 
@@ -77,5 +77,8 @@ getLesionFeatures = function(
   # FEAT 6: t1 itself
   feats[[6]] = antsImageClone(img)   # iMath(img,'Normalize')
 
+  names(feats) = c("d_controls", "d_controls_grad_mag",
+                  "n4_d_controls", "kmean", "ref_diff",
+                  "t1")
   return(feats)
 }

R/linda_predict_function.R

@@ -18,7 +18,7 @@
 #'
 #' @importFrom ANTsRCore iMath antsImageRead antsImageWrite antsRegistration
 #' @importFrom ANTsRCore resampleImage smoothImage thresholdImage antsImageClone
-#' @importFrom ANTsRCore antsApplyTransforms is.antsImage
+#' @importFrom ANTsRCore antsApplyTransforms is.antsImage resampleImageToTarget
 #' @importFrom ANTsR abpN4 abpBrainExtraction reflectImage
 #' @importFrom ANTsR composeTransformsToField splitMask
 #' @importFrom magrittr %>%
@@ -290,52 +290,66 @@ linda_predict = function(
 
   seg = out3$segmentation
 
-  L$segmentation = seg
+  seg_file = file.path(outdir, 'Prediction3_template.nii.gz')
+  L$segmentation = seg_file
 
-  antsImageWrite(seg, file.path(outdir, 'Prediction3_template.nii.gz'))
+  antsImageWrite(seg, seg_file)
 
   print_msg("Saving 3rd final prediction in native space...",
             verbose = verbose)
+  segnative_file = file.path(outdir, 'Prediction3_native.nii.gz')
   segnative = out3$segmentation_native
 
-  antsImageWrite(segnative,
-                 file.path(outdir, 'Prediction3_native.nii.gz'))
+  L$segmentation_native = segnative_file
+  antsImageWrite(segnative, segnative_file)
 
   graded_map = out3$multi_res_seg$probs[[1]][[4]]
 
+  grad_file = file.path(outdir, 'Prediction3_graded_map.nii.gz')
+  antsImageWrite(graded_map, grad_file)
+
   # save graded map
   probles = resampleImage(graded_map,
                           dim(tempbrain),
                           useVoxels = 1,
                           interpType = 0)
+  # probles = resampleImageToTarget(
+  #   image = graded_map,
+  #   target = tempbrain,
+  #   interpType = "nearestNeighbor")
+
+
+
   problesnative = antsApplyTransforms(
     fixed = simg,
     moving = probles,
     transformlist = reg3$fwdtransforms,
     interpolator = 'Linear',
     verbose = verbose > 1
   )
+  probles_file = file.path(outdir,
+                            'Prediction3_probability_template.nii.gz')
+  L$lesion_probability_template = probles_file
   print_msg("Saving probabilistic prediction in template space...",
             verbose = verbose)
-  antsImageWrite(probles,
-                 file.path(outdir,
-                           'Prediction3_probability_template.nii.gz'))
+  antsImageWrite(probles, probles_file)
 
+  problesnative_file = file.path(
+    outdir, 'Prediction3_probability_native.nii.gz')
+  L$lesion_probability_native = problesnative_file
   print_msg("Saving probabilistic prediction in native space...",
             verbose = verbose)
-  antsImageWrite(problesnative,
-                 file.path(outdir,
-                           'Prediction3_probability_native.nii.gz'))
-
+  antsImageWrite(problesnative, problesnative_file)
 
 
   # save in MNI coordinates
   print_msg("Transferring data in MNI (ch2) space...",
             verbose = verbose)
 
-  warppenn = file.path(outdir, 'Reg3_sub_to_template_warp.nii.gz')
-  affpenn = file.path(outdir, 'Reg3_sub_to_template_affine.mat')
-
+  # warppenn = file.path(outdir, 'Reg3_sub_to_template_warp.nii.gz')
+  warppenn = reg_to_temp_warp
+  # affpenn = file.path(outdir, 'Reg3_sub_to_template_affine.mat')
+  affpenn = reg_to_temp_aff
 
   mni = system.file("extdata", "pennTemplate", "ch2.nii.gz",
                     package = "LINDA",
@@ -391,11 +405,15 @@ linda_predict = function(
 
   print_msg("Saving subject in MNI (ch2) space...",
             verbose = verbose)
-  antsImageWrite(submni, file.path(outdir, 'Subject_in_MNI.nii.gz'))
+  t1_template = file.path(outdir, 'Subject_in_MNI.nii.gz')
+  antsImageWrite(submni, t1_template)
+
+  L$t1_template = t1_template
 
   print_msg("Saving lesion in MNI (ch2) space...",
             verbose = verbose)
-  antsImageWrite(lesmni, file.path(outdir, 'Lesion_in_MNI.nii.gz'))
-  return(lesmni)
-
+  lesion_template = file.path(outdir, 'Lesion_in_MNI.nii.gz')
+  L$lesion_template = lesion_template
+  antsImageWrite(lesmni, lesion_template)
+  return(L)
 }