def AugmentDS3D(self,Augentries):
args=self.arggen(Augentries)
# trimgs,labimgs,data_path,num
config['data_path'],config['fext'],num,cr=args[0],args[1],int(args[2]),float(args[3])
fext=config['fext']
data_path=config['data_path']
fold3D='3D'
self.Savefiles(data_path,fext,'trlab')
imhdf5=open_hdf5_file(config['image_hdf5_path'])
trimgs=np.squeeze(imhdf5.root.data,axis=3)
imhdf5.close()
labhdf5=open_hdf5_file(config['label_hdf5_path'])
labimgs=np.squeeze(labhdf5.root.truth,axis=3)
labhdf5.close()
nLabels=np.max(labimgs)
print('estimated number of lables:',nLabels)
print('stack shape',trimgs.shape)
(nd,nr,nc)=trimgs.shape
print(trimgs.shape)
patch_size=64
origin_row = np.random.randint(0, nr-patch_size, num)
origin_col = np.random.randint(0, nc-patch_size, num)
origin_dep = np.random.randint(0, nd-patch_size, num)
origins=np.array((origin_dep,origin_row,origin_col))
trimgs_patches=extract_3D_patches(trimgs,patch_size,origins)
labs_patches=extract_3D_patches(labimgs,patch_size,origins)
if nLabels>1:
for i in range(trimgs_patches.shape[0]):
trimgs_patches[i],labs_patches[i]= transform_3Dpatch(trimgs_patches[i],labs_patches[i])
if nLabels==1:
for i in range(trimgs_patches.shape[0]):
trimgs_patches[i],labs_patches[i]= transform_3Dpatch(trimgs_patches[i],labs_patches[i],scale_deviation=0.15)
print('patches shape',trimgs_patches.shape)
if fold3D in sorted(os.listdir(os.path.join(data_path,image_p))):
shutil.rmtree(os.path.join(data_path,image_p,fold3D), ignore_errors=True)
shutil.rmtree(os.path.join(data_path,label_p,fold3D), ignore_errors=True)
os.makedirs(os.path.join(data_path,image_p,fold3D))
os.makedirs(os.path.join(data_path,label_p,fold3D))
for i in range(trimgs_patches.shape[0]):
os.makedirs(os.path.join(data_path,image_p,fold3D,str(i)))
os.makedirs(os.path.join(data_path,label_p,fold3D,str(i)))
for k in range(trimgs_patches.shape[1]):
self.save_tif(data_path,os.path.join(image_p,fold3D,str(i)),'img',trimgs_patches[i,k],k,fext)
self.save_tif(data_path,os.path.join(label_p,fold3D,str(i)),'img',labs_patches[i,k],k,fext)
imglabpatches=np.array((trimgs_patches,labs_patches))
self.Savefiles3D(data_path,imglabpatches,fext)
return
def Predict3D(self,TMentries):
args=self.arggen(TMentries)
data_path,mName,epchs,vspl,nCl,lr,bs=args[0],args[2],int(args[3]),float(args[4]),int(args[5]),float(args[6]),int(args[7])
if nCl>1:
nLab= nCl +1
else:
nLab=nCl
print('-'*30)
print('Loading and preprocessing test data...')
print('-'*30)
test_open=open_hdf5_file(os.path.join(data_path,config['test_npy']))
imgs_test_tot = np.squeeze(test_open.root.test,axis=3)
test_open.close()
img_depth, img_rows, img_cols = imgs_test_tot.shape
print('-'*30)
print('Loading saved weights...')
print('-'*30)
print(max(glob.glob(os.path.join(data_path,model_p,'*.hdf5')),key=os.path.getctime))
latest=max(glob.glob(os.path.join(data_path,model_p,'*.hdf5')),key=os.path.getctime)
pred_dir = os.path.join(data_path,'preds_'+mName+'_'+latest.split('weights',1)[1])
CPU=False
GPU=True
num_cores = os.cpu_count()-2
if GPU:
num_GPU = 1
num_CPU = 1
if CPU:
num_CPU = 1
num_GPU = 0
configK = tf.ConfigProto(intra_op_parallelism_threads=num_cores,\
inter_op_parallelism_threads=num_cores, allow_soft_placement=True,\
device_count = {'CPU' : num_CPU, 'GPU' : num_GPU})
session = tf.Session(config=configK)
K.set_session(session)
if not GPU:
vol_test = preprocess3D(imgs_test_tot.reshape(1,img_depth, img_rows, img_cols),img_depth, img_rows, img_cols)
vol_test = vol_test.astype('float32')
mean = np.mean(vol_test) # mean for data centering
std = np.std(vol_test) # std for data normalization
vol_test -= mean
vol_test /= std
voltmp=vol_test.reshape(img_depth, img_rows, img_cols,1)
model= getattr(Nmodels3D,mName)(nLab,voltmp.shape,latest)
print('-'*30)
print('Predicting labels on test data...')
print('-'*30)
imgs_mask_test=model.predict(vol_test, verbose=1)
if not os.path.exists(pred_dir):
os.mkdir(pred_dir)
if nLab>1:
imgs_mask_test = imgs_mask_test.reshape(( img_depth, img_rows, img_cols, nLab ) )
vol_mask= imgs_mask_test.argmax( axis=3 ).astype(np.uint8)
else:
vol_mask = np.squeeze((imgs_mask_test * 255.).astype(np.uint8),axis=4)
else:
p_size=64
imgs_test_totshft=np.roll(np.roll(np.roll(imgs_test_tot,p_size//2,axis=2),p_size//2,axis=1),p_size//2,axis=0)
volmask1=self.pred_vol(imgs_test_tot,nLab,mName,p_size,latest)
volmask2=self.pred_vol(imgs_test_totshft,nLab,mName,p_size,latest)
volmask2shft=np.roll(np.roll(np.roll(volmask2,-p_size//2,axis=2),-p_size//2,axis=1),-p_size//2,axis=0)
vol_mask=volmask1+volmask2shft
vol_mask[vol_mask>0.1]=255
# vol_mask=volmask2shft
if not os.path.exists(pred_dir):
os.mkdir(pred_dir)
for i in range(vol_mask.shape[0]):
imsave(os.path.join(pred_dir, str(i) + '_pred.tif'), vol_mask[i].astype(np.uint8))
i+=1
def Predict(self,TMentries):
args=self.arggen(TMentries)
data_path,mName,epchs,vspl,nCl,lr,bs=args[0],args[2],int(args[3]),float(args[4]),int(args[5]),float(args[6]),int(args[7])
if nCl>1:
nLab= nCl +1
else:
nLab=nCl
print('-'*30)
print('Loading and preprocessing test data...')
print('-'*30)
test_open=open_hdf5_file(os.path.join(data_path,config['test_npy']))
imgs_test_tot = np.squeeze(test_open.root.test,axis=3)
test_open.close()
imgs_id_test_tot=np.load(os.path.join(data_path,test_id_npy))
img_rows, img_cols = imgs_test_tot[0].shape
print('-'*30)
print('Loading saved weights...')
print('-'*30)
print(max(glob.glob(os.path.join(data_path,model_p,'*.hdf5')),key=os.path.getctime))
latest=max(glob.glob(os.path.join(data_path,model_p,'*.hdf5')),key=os.path.getctime)
imgtmp=imgs_test_tot[0].reshape(img_rows, img_cols,1)
model= getattr(Nmodels,mName)(nLab,imgtmp.shape,latest,lr)
i=0
for i in range(len(imgs_test_tot)):
print('image',i+1,'of',len(imgs_test_tot))
imgs_test = preprocess(imgs_test_tot[i].reshape(1,img_rows, img_cols),img_rows, img_cols)
imgs_id_test=imgs_id_test_tot[i]
print('processing image',imgs_id_test)
imgs_test = imgs_test.astype('float32')
mean = np.mean(imgs_test) # mean for data centering
std = np.std(imgs_test) # std for data normalization
imgs_test -= mean
imgs_test /= std
# model.load_weights(os.path.join(data_path,model_p,'weights.h5'))
# print(len(imgs_test))
# op_shape=model.output_shape
# t=list(op_shape)
# t[0]=3
# op_shape=tuple(t)
print('-'*30)
print('Predicting labels on test data...')
print('-'*30)
imgs_mask_test=model.predict(imgs_test, verbose=1)
print('-' * 30)
print('Saving predicted labels to files...')
print('-' * 30)
pred_dir = os.path.join(data_path,'preds_'+mName+'_'+latest.split('weights',1)[1])
if not os.path.exists(pred_dir):
os.mkdir(pred_dir)
if nLab>1:
imgs_mask_test = imgs_mask_test.reshape(( img_rows, img_cols , nLab ) )
imgs_mask_test= imgs_mask_test.argmax( axis=2 ).astype(np.uint8)
# imagep = (image * 255.).astype(np.uint8)
else:
imgs_mask_test = (imgs_mask_test * 255.).astype(np.uint8)
imsave(os.path.join(pred_dir, str(imgs_id_test) + '_pred.tif'), imgs_mask_test)
def TrainModel3D(self,TMentries):
args=self.arggen(TMentries)
data_path,mName,epchs,vspl,nCl,lr,bs=args[0],args[2],int(args[3]),float(args[4]),int(args[5]),float(args[6]),int(args[7])
if nCl>1:
nLab= nCl +1
else:
nLab=nCl
config['data_path']=data_path
config['image_path']=os.path.join(data_path,image_p)
config['label_path']=os.path.join(data_path,label_p)
config['image_hdf5_path']=os.path.join(config['data_path'],config['trimg_npy'])
config['label_hdf5_path']=os.path.join(config['data_path'],config['trlab_npy'])
print('-'*30)
print('Loading and preprocessing train data...')
print('-'*30)
imgsopen=open_hdf5_file(config['image_hdf5_path'])
imgs_train = np.squeeze(imgsopen.root.data,axis=4)
imgsopen.close()
labopen=open_hdf5_file(config['label_hdf5_path'])
imgs_mask_train = np.squeeze(labopen.root.truth,axis=4)
labopen.close()
(img_depth, img_rows, img_cols) = (imgs_train[0].shape)
imgs_train2 = []
for patch in imgs_train:
patchtmp=preprocess(patch, img_rows, img_cols)
imgs_train2.append(patchtmp)
imgs_mask_train2 = []
for patch in imgs_mask_train:
patchtmp=preprocess(patch, img_rows, img_cols)
imgs_mask_train2.append(patchtmp)
imgs_train = np.array(imgs_train2).astype('float32')
mean = np.mean(imgs_train) # mean for data centering
std = np.std(imgs_train) # std for data normalization
imgs_train -= mean
imgs_train /= std
# imgs_mask_train = imgs_mask_train.astype('float32')
# imgs_mask_train /= 255. # scale masks to [0, 1]
# imgs_mask_train *= (nCl) # generates the labes as integers
imgs_mask_train = np.array(imgs_mask_train2).astype('uint8')
print('Size of the training data:',imgs_mask_train.shape)
if np.max(imgs_mask_train) != nCl:
print('Warning: the number of classes does not match the intesities of the label images')
if nLab>1:
imgs_mask_train = getSegmentationArr(imgs_mask_train , nLab)
# global imgs_mask_train2
# imgs_mask_train2 = np.copy(imgs_mask_train)
else:
imgs_mask_train[imgs_mask_train > 0.5] = 1
imgs_mask_train[imgs_mask_train <= 0.5] = 0
print(imgs_mask_train.shape)
print('-'*30)
print('Creating and compiling model...')
print('-'*30)
# model = get_unet(imgs_train[0].shape)
# model = Model4(imgs_train[0].shape)
if os.path.exists(os.path.join(data_path,model_p)) and len([x for x in os.listdir(os.path.join(data_path,model_p)) if ('.hdf5') in x])>0:
print('loading weights and compiling the model')
latest=max(glob.glob(os.path.join(data_path,model_p,'*.hdf5')),key=os.path.getctime)
model= getattr(Nmodels3D,mName)(nLab,imgs_train[0].shape,latest,lr)
else:
if not os.path.exists(os.path.join(data_path,model_p)):
os.makedirs(os.path.join(data_path,model_p))
model= getattr(Nmodels3D,mName)(nLab,imgs_train[0].shape,'',lr)
# model_checkpoint = ModelCheckpoint('weights.h5', monitor='val_loss', save_best_only=True)
model_checkpoint = ModelCheckpoint(os.path.join(data_path,model_p,mName+'weights.ep{epoch:02d}-il{loss:.3f}-vl{val_loss:.3f}.hdf5'), monitor='loss',verbose=1, save_best_only=True)
print('-'*30)
print('Fitting model...')
print('-'*30)
# model.fit(imgs_train, imgs_mask_train, batch_size=34, nb_epoch=20, verbose=1, shuffle=True,
# validation_split=0.2,
# callbacks=[model_checkpoint])
model.fit(imgs_train, imgs_mask_train, batch_size=bs, epochs=epchs, verbose=1,
validation_split=vspl,
shuffle=True,
callbacks=[model_checkpoint])
return
def AugmentDS(self,Augentries):
datagen = ImageDataGenerator(
# featurewise_center=False,
# featurewise_std_normalization=False,
# samplewise_center=False,
# samplewise_std_normalization=False,
# zca_whitening=True,
# rescale=None,
rotation_range=3,
width_shift_range=0.08,
height_shift_range=0.08,
shear_range=0.07,
zoom_range=0.07,
horizontal_flip=True,
vertical_flip=True,
fill_mode='constant',
cval=0.
)
args=self.arggen(Augentries)
# trimgs,labimgs,data_path,num
config['data_path'],config['fext'],num,cr=args[0],args[1],int(args[2]),float(args[3])
fext=config['fext']
data_path=config['data_path']
self.Savefiles(data_path,fext,'trlab')
imhdf5=open_hdf5_file(config['image_hdf5_path'])
trimgs=np.squeeze(imhdf5.root.data,axis=3)
imhdf5.close()
labhdf5=open_hdf5_file(config['label_hdf5_path'])
labimgs=np.squeeze(labhdf5.root.truth,axis=3)
labhdf5.close()
nLabels=np.max(labimgs)
print('estimated number of lables:',nLabels)
if nLabels>1:
labimgstmp=[]
for i in range(1,nLabels+1):
labimgstmp.append(np.ma.masked_not_equal(labimgs,i).filled(0)/i)
labimgstmp=np.array(labimgstmp)
imgshape=trimgs[0].shape
print(imgshape)
print('-'*30)
print('Augmenting train and labels dataset: ',num,'replica per image...')
print('-'*30)
# seed = np.random.randint(10000)
seed=np.random.randint(10000,size=2*len(trimgs)*num)
if tmpf in sorted(os.listdir(config['image_path'])):
shutil.rmtree(os.path.join(config['image_path'],tmpf), ignore_errors=True)
shutil.rmtree(os.path.join(config['label_path'],tmpf), ignore_errors=True)
os.makedirs(os.path.join(config['image_path'],tmpf))
os.makedirs(os.path.join(config['label_path'],tmpf))
global batchdata
batchdata=[]
j=0
for x in trimgs:
x[x==0]=1
x = x.reshape((1,) + x.shape+(1,))
# the .flow() command below generates batches of randomly transformed images
# and saves the results to the `preview/` directory
i = 0
for batch in datagen.flow(x, batch_size=1,seed=seed[j]):
self.save_tif(data_path,os.path.join(image_p,tmpf),'img',batch[0,:,:,0].astype('uint8'),seed[i+j*2*num],fext)
i += 1
if i >= 2*num:
break # otherwise the generator would loop indefinitely
j +=1
if nLabels>1:
for k in range(1,nLabels+1):
os.makedirs(os.path.join(config['label_path'],tmpf,str(k)))
j=0
for y in labimgstmp[k-1]:
y = y.reshape((1,) + y.shape+(1,))
i = 0
for batch in datagen.flow(y, batch_size=1,seed=seed[j]):
self.save_tif(data_path,os.path.join(label_p,tmpf,str(k)),'img',batch[0,:,:,0].astype('uint8'),seed[i+j*2*num],fext)
batchdata.append(batch[0,:,:,0])
i += 1
if i >= 2*num:
break # otherwise the generator would loop indefinitely
j +=1
imglist=[f for f in sorted(os.listdir(os.path.join(config['image_path'],tmpf))) if fext in f]
for n in range(len(imglist)):
tmp=sum(read_image(os.path.join(config['label_path'],tmpf,str(k),imglist[n]))*k for k in range(1,nLabels+1))
self.save_tif(data_path,os.path.join(label_p,tmpf),'img',tmp.astype('uint8'),imglist[n].split('.')[0][-4:],fext)
for k in range(1,nLabels+1):
shutil.rmtree(os.path.join(config['label_path'],tmpf,str(k)), ignore_errors=True)
else:
j=0
for y in labimgs:
y = y.reshape((1,) + y.shape+(1,))
i = 0
for batch in datagen.flow(y, batch_size=1,seed=seed[j]):
self.save_tif(data_path,os.path.join(label_p,tmpf),'img',batch[0,:,:,0].astype('uint8'),seed[i+j*2*num],fext)
batchdata.append(batch[0,:,:,0])
i += 1
if i >= 2*num:
break # otherwise the generator would loop indefinitely
j +=1
self.Savefiles(data_path,fext,'trlab',subtask='augtmp')
# create_train_data(data_path,os.path.join(image_p,tmpf),os.path.join(label_p,tmpf),fext)
imhdf5=open_hdf5_file(config['image_hdf5_path'])
tmptr=np.squeeze(imhdf5.root.data,axis=3)
imhdf5.close()
labhdf5=open_hdf5_file(config['label_hdf5_path'])
tmplab=np.squeeze(labhdf5.root.truth,axis=3)
labhdf5.close()
print(imgshape,cr)
lencrop=int(((imgshape[0]*cr)//16)*16),int(((imgshape[1]*cr)//16)*16)
print(lencrop)
# delta=imgshape[0]-lencrop[0],imgshape[1]-lencrop[1]
# print(delta)
seltr=[]
sellab=[]
j=0
for i,img in enumerate(tmptr):
tmpres=crop_no_black(tmptr[i],tmplab[i],lencrop)
if tmpres is not None:
seltr.append(tmpres[0])
sellab.append(tmpres[1])
j += 1
if j > len(trimgs)*(num+1):
break
seltr=np.array(seltr)
sellab=np.array(sellab)
print(seltr.shape,sellab.shape)
if selfold in sorted(os.listdir(os.path.join(data_path,image_p))):
shutil.rmtree(os.path.join(data_path,image_p,selfold), ignore_errors=True)
shutil.rmtree(os.path.join(data_path,label_p,selfold), ignore_errors=True)
os.makedirs(os.path.join(data_path,image_p,selfold))
os.makedirs(os.path.join(data_path,label_p,selfold))
for i in range(len(seltr)):
self.save_tif(data_path,os.path.join(image_p,selfold),'img',seltr[i],i,fext)
self.save_tif(data_path,os.path.join(label_p,selfold),'img',sellab[i],i,fext)
# create_train_data(data_path,image_p,label_p,fext)
if tmpf in sorted(os.listdir(os.path.join(data_path,image_p))):
shutil.rmtree(os.path.join(data_path,image_p,tmpf), ignore_errors=True)
shutil.rmtree(os.path.join(data_path,label_p,tmpf), ignore_errors=True)
self.Savefiles(data_path,fext,'trlab',subtask='augm')
print('Done')
return
def Savefiles(self,data_path,fext,task,subtask=None):
if task =='trlab':
config['image_path']=os.path.join(data_path,image_p)
config['label_path']=os.path.join(data_path,label_p)
outim=config['image_hdf5_path']=os.path.join(config['data_path'],config['trimg_npy'])
outlab=config['label_hdf5_path']=os.path.join(config['data_path'],config['trlab_npy'])
if subtask == None:
if selfold in sorted(os.listdir(config['image_path'])):
print('Augmented data found. Saving augmented data instead of original ones')
imgs=fetch_data_1dir(os.path.join(config['image_path'],selfold),fext)
lbls=fetch_data_1dir(os.path.join(config['label_path'],selfold),fext)
# create_train_data(data_path,os.path.join(image_p,selfold),os.path.join(label_p,selfold),fext)
if aug3Dfol in sorted(os.listdir(config['image_path'])):
print('3D Augmented data found. Saving augmented data instead of original ones')
imgs=list()
lbls=list()
imgstmp=fetch_augdata_subdirs(os.path.join(config['image_path'],aug3Dfol),fext)
lblstmp=fetch_augdata_subdirs(os.path.join(config['label_path'],aug3Dfol),fext)
for volume in imgstmp:
imgs.append(volume)
for volume in lblstmp:
lbls.append(volume)
else:
if datasubf3D in sorted(os.listdir(config['image_path'])):
imgs=fetch_data_subdirs(os.path.join(config['image_path'],datasubf3D),fext)
lbls=fetch_data_subdirs(os.path.join(config['label_path'],datasubf3D),fext)
else:
imgs=fetch_data_1dir(config['image_path'],fext)
lbls=fetch_data_1dir(config['label_path'],fext)
if subtask=='augm':
if datasubf3D in sorted(os.listdir(config['image_path'])):
imgs=list()
lbls=list()
if aug3Dfol in sorted(os.listdir(config['image_path'])):
shutil.rmtree(os.path.join(config['image_path'],aug3Dfol), ignore_errors=True)
imgs=fetch_data_subdirs(os.path.join(config['image_path'],datasubf3D),fext)
if aug3Dfol in sorted(os.listdir(config['label_path'])):
shutil.rmtree(os.path.join(config['label_path'],aug3Dfol), ignore_errors=True)
lbls=fetch_data_subdirs(os.path.join(config['label_path'],datasubf3D),fext)
else:
imgs=fetch_data_1dir(os.path.join(config['image_path'],selfold),fext)
lbls=fetch_data_1dir(os.path.join(config['label_path'],selfold),fext)
if subtask=='augtmp':
imgs=fetch_data_1dir(os.path.join(config['image_path'],tmpf),fext)
lbls=fetch_data_1dir(os.path.join(config['label_path'],tmpf),fext)
# create_train_data(data_path,image_p,label_p,fext)
if datasubf3D in sorted(os.listdir(config['image_path'])):
print('3D data found')
config['img_comp']=write_data_to_file(imgs,outim,'data',is3D=True)
if datasubf3D in sorted(os.listdir(config['label_path'])):
config['lab_comp']=write_data_to_file(lbls,outlab,'truth',is3D=True)
else:
config['img_comp']=write_data_to_file(imgs,outim,'data')
config['lab_comp']=write_data_to_file(lbls,outlab,'truth')
print('Training images hdf5 file written to:',outim)
print('Training labels hdf5 file written to:',outlab)
imgsopen=open_hdf5_file(config['image_hdf5_path'])
imgs_train = np.squeeze(imgsopen.root.data,axis=(-1))
print('hdf image data shape:',imgs_train.shape)
imgsopen.close()
if task =='test':
config['test_path']=os.path.join(data_path,test_p)
outtest=os.path.join(config['data_path'],config['test_npy'])
test=fetch_data_1dir(config['test_path'],fext)
config['test_comp']=write_data_to_file(test,outtest,'test')
print('Test images hdf5 file written to:',outtest)
imgs_id = np.ndarray((len(test), ), dtype=np.int32)
i=0
for image_name in test:
img_id = int(image_name.split('.')[0][-4:].lstrip('0'))
imgs_id[i] = img_id
i += 1
np.save(os.path.join(data_path,'imgs_id_test.npy'), imgs_id)
def AugmentDS3D(self,Augentries):
args=self.arggen(Augentries)
# trimgs,labimgs,data_path,num
config['data_path'],config['fext'],num,cr=args[0],args[1],int(args[2]),float(args[3])
fext=config['fext']
data_path=config['data_path']
fold3D=config['3Daugm']
self.Savefiles(data_path,fext,'trlab',subtask='augm')
imhdf5=open_hdf5_file(config['image_hdf5_path'])
trimgs=np.squeeze(imhdf5.root.data,axis=(-1))
imhdf5.close()
labhdf5=open_hdf5_file(config['label_hdf5_path'])
labimgs=np.squeeze(labhdf5.root.truth,axis=(-1))
labhdf5.close()
nLabels=np.max(labimgs)
print('estimated number of lables:',nLabels)
print('stack shape',trimgs.shape)
(nvols,nd,nr,nc)=trimgs.shape
patch_size=64
origin_row = np.random.randint(0, nr-patch_size, num*nvols)
origin_col = np.random.randint(0, nc-patch_size, num*nvols)
origin_dep = np.random.randint(0, nd-patch_size, num*nvols)
origins=[]
for i in range(nvols):
volorigins=np.array((origin_dep[i*num:(i+1)*num],origin_row[i*num:(i+1)*num],origin_col[i*num:(i+1)*num]))
origins.append(volorigins)
origins=np.array(origins)
trimgs_patches=[]
labs_patches=[]
for i in range(nvols):
trimgs_patches.append(extract_3D_patches(trimgs[i],patch_size,origins[i]))
labs_patches.append(extract_3D_patches(labimgs[i],patch_size,origins[i]))
trimgs_patches=np.array(trimgs_patches)
labs_patches=np.array(labs_patches)
if nLabels>1:
for i in range(trimgs_patches.shape[0]):
for j in range(trimgs_patches.shape[1]):
# print(i,j)
# plt.imshow(trimgs_patches[i,j][0])
# plt.show()
trimgs_patches[i,j],labs_patches[i,j]= transform_3Dpatch(trimgs_patches[i,j],labs_patches[i,j])
# print('Transformed')
# plt.imshow(trimgs_patches[i,j][0])
# plt.show()
# print('-'*30)
if nLabels==1:
for i in range(trimgs_patches.shape[0]):
for j in range(trimgs_patches.shape[1]):
# trimgs_patches[i,j],labs_patches[i,j]= transform_3Dpatch(trimgs_patches[i,j],labs_patches[i,j],scale_deviation=0.15)
trimgs_patches[i,j],labs_patches[i,j]= transform_3Dpatch(trimgs_patches[i,j],labs_patches[i,j])
print('final training patch shape',trimgs_patches.shape)
if fold3D in sorted(os.listdir(os.path.join(data_path,image_p))):
shutil.rmtree(os.path.join(data_path,image_p,fold3D), ignore_errors=True)
shutil.rmtree(os.path.join(data_path,label_p,fold3D), ignore_errors=True)
os.makedirs(os.path.join(data_path,image_p,fold3D))
os.makedirs(os.path.join(data_path,label_p,fold3D))
for i in range(trimgs_patches.shape[0]):
os.makedirs(os.path.join(data_path,image_p,fold3D,str(i)))
os.makedirs(os.path.join(data_path,label_p,fold3D,str(i)))
for k in range(trimgs_patches.shape[1]):
os.makedirs(os.path.join(data_path,image_p,fold3D,str(i),str(k)))
os.makedirs(os.path.join(data_path,label_p,fold3D,str(i),str(k)))
for j in range(trimgs_patches.shape[2]):
self.save_tif(data_path,os.path.join(image_p,fold3D,str(i),str(k)),'img',trimgs_patches[i,k][j],j,fext)
self.save_tif(data_path,os.path.join(label_p,fold3D,str(i),str(k)),'img',labs_patches[i,k][j],j,fext)
# imglabpatches=np.array((trimgs_patches,labs_patches))
self.Savefiles(data_path,fext,'trlab')
# self.Savefiles3D(data_path,imglabpatches,fext)
return