# predict dir
root_dir = '/media/louis/Volume/ProgramWorkResult/MICCAI_2017_6_12/Result_Eigenfaces_KNN/RegPatchedImage'
if OrganPatch == True:
result_dir_pre = root_dir + '/'+organ + '/PCA_predict'
elif LocalPatach == True:
result_dir_pre = root_dir + '/'+organ + 'LocalPatch'+ '/PCA_predict'
else:
raise ValueError('please choose the patch pattern')
print 'Results will be stored in:', result_dir_pre
if not os.path.exists(result_dir_pre):
subprocess.call('mkdir ' + '-p ' + result_dir_pre, shell=True)
#--------------- --------------------------------------------------------------
im_fns = Dpp.readTxtIntoList(result_dir_pre + '/' + 'FileList.txt')
predict_list = Dpp.ImageModificationAfterTransform(im_fns, result_dir_pre, constraintPredict)
# ground truth dir
root_dir = '/media/louis/Volume/ProgramWorkResult/MICCAI_2017_6_12/Result_Eigenfaces_KNN/RegPatchedImage'
if OrganPatch == True:
result_dir_true = root_dir + '/'+organ + '/PCA_test'
elif LocalPatach == True:
result_dir_true = root_dir + '/'+organ + 'LocalPatch'+ '/PCA_test'
else:
raise ValueError('please choose the patch pattern')
if not os.path.exists(result_dir_true):
subprocess.call('mkdir ' + '-p ' + result_dir_true, shell=True)
print 'Results will be stored in:', result_dir
if not os.path.exists(result_dir):
subprocess.call('mkdir ' + '-p ' + result_dir, shell=True)
outputPrefix = os.path.join(result_dir, 'JLF_predict/')
if not os.path.exists(outputPrefix):
subprocess.call('mkdir ' + '-p ' + outputPrefix, shell=True)
# For reproducibility: save all parameters into the result dir
currentPyFile = os.path.realpath(__file__)
os.system('cp ' + currentPyFile + ' ' + result_dir)
#%%
X_data_dir = root_image_dir + '/' + organ + '_' + regIfLinear + '_Imagepatch'
Y_data_dir = root_label_dir + '/' + organ + '_' + regIfLinear + '_Labelpatch'
X_data_list = Dpp.readTxtIntoList(X_data_dir + '/FileList.txt')
Y_data_list = Dpp.readTxtIntoList(Y_data_dir + '/FileList.txt')
reference_im_fn = X_data_list[0]
# lamada: the tunning paramter that weights between the low-rank component and the sparse component
global lamda
lamda = 0.7
tol = 1e-07
# train data and test data
data_num = len(X_data_list)
PcNo = data_num - 1
MonteTime = 1
best_k = 30
#%%
#%%
root_dir = '/media/data/louis/ProgramWorkResult/VisercialMAS_ANT'
Input_Root_dir = root_dir + '/Patched_Image'
Image_Root_Dir = Input_Root_dir + '/GC_Volumes_patch_Elastix/' + organ + '_ImageMask'
Label_Root_Dir = Input_Root_dir + '/GC_label_patch_Elastix/' + organ + '_LabelMask'
Output_Root_dir = '/media/louis/Volume/ProgramWorkResult/MICCAI_2017_6_12/OrganReg_Elastix'
Image_Result_Dir = Output_Root_dir + '/Volume_OrganReg' + '/' + organ
Label_Result_Dir = Output_Root_dir + '/Label_OrganReg' + '/' + organ
# Make reference
reference_dir = root_dir + '/Full_Image/GC_Volumes_adjustment' + '/referenceImage'
selName = "10000128_1_CTce_ThAb"
DirFile = Dpp.readTxtIntoList(Image_Root_Dir + '/FileList.txt')
for dir_i in DirFile:
if dir_i.find(selName) != -1:
referenceImageMask = dir_i
break
#%%
# Registration
if IFNormalization != False:
reference_im_fn = reference_dir + '/reference_WithNorm' + '.nrrd'
else:
reference_im_fn = reference_dir + '/reference_WithoutNorm' + '.nrrd'
Imagereginput_dir = Image_Root_Dir
images = Dpp.readTxtIntoList(Imagereginput_dir + '/InputImageList.txt')
iamges_mask = Dpp.readTxtIntoList(Imagereginput_dir + '/FileList.txt')
ImageRegOutput_dir = Image_Result_Dir
reference_im_fn = reference_dir + '/reference_WithoutNorm' + '.nrrd'
#------------------------------------------------------------------------------
Regmode = 'linear'
#Regmode = 'nonlinear'
#organ = '40358_muscle_body_of_left_rectus_abdominis'
#organ = '187_gallbladder'
#organ = '170_pancreas'
#organ = '30325_left_adrenal_gland'
#organ = '30324_right_adrenal_gland'
organ = '29193_first_lumbar_vertebra'
#'''
# %%
# adjust the original Label data.
data_folder = '/media/louis/Volume/ResearchData/visceral_used/GC_Labels/' + organ
im_fns = Dpp.readTxtIntoList(data_folder + '/FileList.txt')
searchName = "10000129_1_CTce_ThAb"
for dir_i in im_fns:
if dir_i.find(searchName) != -1:
im_fns.remove(dir_i)
break
label_dir = root_label_dir + '/' + organ
sizeadjust_label_dir = label_dir + '/sizeadjustment'
if not os.path.exists(sizeadjust_label_dir):
subprocess.call('mkdir ' + '-p ' + sizeadjust_label_dir, shell=True)
cuttingresult_dir = label_dir + '/cuttingresult'
if not os.path.exists(cuttingresult_dir):
subprocess.call('mkdir ' + '-p ' + cuttingresult_dir, shell=True)
# show image information
import SimpleITK as sitk # SimpleITK to load images
import sys
sys.path.insert(0, '../')
import DataPreProcessing.Func_data_preprocessing as Dpp
#%%
data_folder = '/media/louis/Volume/ProgramWorkResult/MICCAI_2017_6_12/Full_Image/GC_Volumes_adjustment/cuttingresult'
InputImageList = Dpp.readTxtIntoList(data_folder + '/FileList.txt')
for i in xrange(len(InputImageList)):
image = sitk.ReadImage(InputImageList[i])
image_array = sitk.GetArrayFromImage(image) # get numpy array
print 'Image No. %d' % (i)
print 'dim z,x,y :' + str(image_array.shape)
print 'Origin:' + str(image.GetOrigin())
print 'sapcing:' + str(image.GetSpacing())
import os
import subprocess
sys.path.insert(0, '../')
import DataPreProcessing.Func_data_preprocessing as Dpp
regIfLinear = "Linear"
#regIfLinear = "nonlinear"
organ = '187_gallbladder'
#organ = '170_pancreas'
#organ = '30325_left_adrenal_gland'
#organ = '30324_right_adrenal_gland'
#organ = '29193_first_lumbar_vertebra'
root_dir = '/media/louis/Volume/ProgramWorkResult/VisercialMAS_ANT'
label_data_dir = root_dir + '/Full_Image/GC_label_adjustment' + '/'+organ +\
'/Regtrans_' + regIfLinear +'_ANT'
result_dir_lab = root_dir + '/Full_Image/GC_label_adjustment' + '/'+organ +\
'/Modification_Regtrans_' + regIfLinear + '_ANT'
print 'Results will be stored in:', result_dir_lab
if not os.path.exists(result_dir_lab):
subprocess.call('mkdir ' + '-p ' + result_dir_lab, shell=True)
#--------------------------------------------------------------------------------
im_fns_label = Dpp.readTxtIntoList(label_data_dir + '/FileList.txt')
result_lab_list = Dpp.ImageModificationAfterTransform(im_fns_label,
result_dir_lab)
Dpp.WriteListtoFile(result_lab_list, result_dir_lab + '/FileList.txt')