""" Post-process Mindboggle-101 volume images for distribution, using Mindboggle, FreeSurfer, and FSL tools. This is modified from the original code_postprocess_Mindboggle101.py to (1) only generate DKT31 (not DKT25) labeling protocol data (2) use the existing T1 data and transforms (don't generate new ones): - Convert label volume from FreeSurfer to original space x Extract brain by masking with manual cortical and automated subcortical labels - Remove non-DKT31 (non-cortical) labels - Affine register T1-weighted brain to MNI152 brain - Transfer whole-head images with affine transform - Transfer labeled images with affine transform (nearest-neighbor interpolation) Authors: Arno Klein . arno@mindboggle.info . www.binarybottle.com (c) 2013-2019 Mindbogglers (www.mindboggle.info), under Apache License Version 2.0 """ import os # Paths, template, and label conversion files mb101_path = os.path.join('/Users', 'arno.klein', 'Data', 'Mindboggle101') mb_info_path = os.path.join(mb101_path, 'docs') template = os.path.join(mb101_path, 'MNI152_T1_1mm_brain.nii.gz') # Loop through subjects list_file = os.path.join(mb_info_path, 'mindboggle101_list.txt') fid = open(list_file, 'r') subjects = fid.readlines() subjects = [''.join(x.split()) for x in subjects] def keep_volume_labels(input_file, labels_to_keep, output_file='', second_file=''): """ Keep only given labels in an image volume (or use to mask second volume). Parameters ---------- input_file : string labeled nibabel-readable (e.g., nifti) file labels_to_keep : list of integers labels to keep output_file : string output file name second_file : string second nibabel-readable file (keep/erase voxels in this file instead) Returns ------- output_file : string output file name Examples -------- >>> # Remove right hemisphere labels >>> import os >>> from mindboggle.guts.relabel import keep_volume_labels >>> from mindboggle.mio.labels import DKTprotocol >>> from mindboggle.mio.fetch_data import prep_tests >>> urls, fetch_data = prep_tests() >>> input_file = fetch_data(urls['freesurfer_labels'], '', '.nii.gz') >>> second_file = '' >>> labels_to_keep = list(range(1000, 1036)) >>> output_file = 'keep_volume_labels.nii.gz' >>> output_file = keep_volume_labels(input_file, labels_to_keep, ... output_file, second_file) View nifti file (skip test): >>> from mindboggle.mio.plots import plot_volumes >>> plot_volumes(output_file) # doctest: +SKIP """ import os import numpy as np import nibabel as nb # ------------------------------------------------------------------------ # Load labeled image volume and extract data as 1-D array: # ------------------------------------------------------------------------ vol = nb.load(input_file) xfm = vol.get_affine() data = vol.get_data().ravel() # ------------------------------------------------------------------------ # If second file specified, erase voxels whose corresponding # voxels in the input_file have labels not in labels_to_keep: # ------------------------------------------------------------------------ if second_file: # Load second image volume and extract data as 1-D array: vol = nb.load(second_file) xfm = vol.get_affine() new_data = vol.get_data().ravel() if not output_file: output_file = os.path.join(os.getcwd(), os.path.basename(second_file)) # ------------------------------------------------------------------------ # If second file not specified, remove labels not in labels_to_keep: # ------------------------------------------------------------------------ else: new_data = data.copy() if not output_file: output_file = os.path.join(os.getcwd(), os.path.basename(input_file)) # ------------------------------------------------------------------------ # Erase voxels as specified above: # ------------------------------------------------------------------------ ulabels = np.unique(data) for label in ulabels: label = int(label) if label not in labels_to_keep: new_data[np.where(data == label)[0]] = 0 # ------------------------------------------------------------------------ # Reshape to original dimensions: # ------------------------------------------------------------------------ new_data = np.reshape(new_data, vol.shape) # ------------------------------------------------------------------------ # Save relabeled file: # ------------------------------------------------------------------------ img = nb.Nifti1Image(new_data, xfm) img.to_filename(output_file) if not os.path.exists(output_file): raise IOError("keep_volume_labels() did not create " + output_file + ".") return output_file for subject in subjects: print(">>> Process subject: {0}...".format(subject)) subject_path = os.path.join(mb101_path, 'subjects', subject, 'mri') # Identify original files full_labels_orig = os.path.join(subject_path, 'aparcNMMjt+aseg.nii.gz') head = os.path.join(subject_path, 't1weighted.nii.gz') brain = os.path.join(subject_path, 't1weighted_brain.nii.gz') # Name all output files full_labels = os.path.join(subject_path, 'labels.DKT31.manual+aseg.nii.gz') DKT31_labels = os.path.join(subject_path, 'labels.DKT31.manual.nii.gz') xfm_matrix = os.path.join(subject_path, 't1weighted_brain.MNI152.affine.txt') xfm_brain = os.path.join(subject_path, 't1weighted_brain.MNI152.nii.gz') xfm_head = os.path.join(subject_path, 't1weighted.MNI152.nii.gz') xfm_DKT31 = os.path.join(subject_path, 'labels.DKT31.manual.MNI152.nii.gz') xfm_DKT31aseg = os.path.join(subject_path, 'labels.DKT31.manual+aseg.MNI152.nii.gz') # Remove old labels and affine-transformed files rm_files = [x for x in os.listdir(subject_path) if 'labels.' in x or '.MNI152.' in x] for rm_file in rm_files: os.remove(os.path.join(subject_path, rm_file)) # Convert label volume from FreeSurfer to original space print("Convert label volume from FreeSurfer to original space...") cmd = ' '.join(['mri_vol2vol --nearest --mov', full_labels_orig, '--targ', head, '--regheader --o', full_labels]) print(cmd); os.system(cmd) # Affine register T1-weighted brain to MNI152 brain using FSL's flirt print("Affine register T1-weighted brain to MNI152 brain using FSL's flirt...") cmd = ' '.join(['flirt', '-in', brain, '-ref', template, '-out', xfm_brain, '-omat', xfm_matrix]) print(cmd); os.system(cmd) # Transfer whole-head images with affine transform using FSL's flirt print("Apply affine transform to whole-head using FSL's flirt...") cmd = ' '.join(['flirt', '-in', head, '-ref', template, '-applyxfm -init', xfm_matrix, '-out', xfm_head]) print(cmd); os.system(cmd) # Transfer DKT31- plus FreeSurfer-aseg-labeled images with affine transform (nearest-neighbor interpolation) print("Apply affine transform to labeled images (with nearest neighbor interpolation)...") cmd = ' '.join(['flirt', '-in', full_labels, '-ref', template, '-applyxfm -init', xfm_matrix, '-interp nearestneighbour -out', xfm_DKT31aseg]) print(cmd); os.system(cmd) # Remove all but DKT31 (cortical) labels print("Remove non-DKT31 (cortical) labels...") DKT31_numbers = [2, 3] + list(range(5, 32)) + [34, 35] labels_to_keep = [1000 + x for x in DKT31_numbers] labels_to_keep.extend([2000 + x for x in DKT31_numbers]) output_file = keep_volume_labels(full_labels, labels_to_keep, output_file=DKT31_labels, second_file='') # Transfer DKT31-labeled images with affine transform (nearest-neighbor interpolation) cmd = ' '.join(['flirt', '-in', DKT31_labels, '-ref', template, '-applyxfm -init', xfm_matrix, '-interp nearestneighbour -out', xfm_DKT31]) print(cmd); os.system(cmd) # Compress subject directory #subject_path2 = os.path.join(mb101_path, 'subjects', subject) #cmd = ' '.join(['tar cvfz', subject_path2+'.tar.gz', subject_path2]) #print(cmd); os.system(cmd)