def predict_all_xmls(args, dirs, wsi, iteration):
# reshape regions calc
downsample = int(args.downsampleRateLR**.5)
region_size = int(args.boxSizeLR * (downsample))
step = int(region_size * (1 - args.overlap_percentLR))
# figure out the number of classes
annotatedXMLs = glob(args.base_dir + '/' + args.project +
dirs['training_data_dir'] + str(iteration - 1) +
'/*.xml')
classes = []
for xml in annotatedXMLs:
classes.append(get_num_classes(xml))
classNum = max(classes)
# chop wsi
fileID, test_num_steps, slide = chop_suey(wsi, dirs, downsample,
region_size, step, args)
dirs['fileID'] = fileID
print('Chop SUEY!\n')
# call DeepLab for prediction (Low resolution)
print('finding Glom locations ...\n')
make_folder(dirs['outDir'] + fileID + dirs['img_save_dir'] + 'prediction')
modeldir = args.base_dir + '/' + args.project + dirs['modeldir'] + str(
iteration) + '/LR'
test_steps_LR = get_all_test_steps(modeldir)
for idx, test_step in enumerate(test_steps_LR):
# test all saved models in current iteration
modeldir = args.base_dir + '/' + args.project + dirs['modeldir'] + str(
iteration) + '/LR'
test_data_list = fileID + '_images' + '.txt'
print("\033[1;32;40m" + "starting prediction using model: \n\t" +
modeldir + str(test_step) + "\033[0;37;40m" + "\n\n")
call([
'python3', args.base_dir + '/Codes/Deeplab_network/main.py',
'--option', 'predict', '--test_data_list',
dirs['outDir'] + fileID + dirs['txt_save_dir'] + test_data_list,
'--out_dir', dirs['outDir'] + fileID + dirs['img_save_dir'],
'--test_step',
str(test_step), '--test_num_steps',
str(test_num_steps), '--modeldir', modeldir, '--data_dir',
dirs['outDir'] + fileID + dirs['img_save_dir'], '--num_classes',
str(classNum), '--gpu', '0'
])
# un chop
print('\nreconstructing wsi map ...\n')
wsiMask = un_suey(dirs=dirs, args=args)
# save hotspots
if dirs['save_outputs'] == True:
make_folder(dirs['outDir'] + fileID + dirs['mask_dir'])
print('saving to: ' + dirs['outDir'] + fileID + dirs['mask_dir'] +
fileID + '.png')
with warnings.catch_warnings():
warnings.simplefilter("ignore")
imsave(
dirs['outDir'] + fileID + dirs['mask_dir'] + fileID +
'.png', wsiMask)
# find glom locations in reconstructed map
print('\ninterpreting prediction map ...')
test_num_steps, labeledArray, label_offsets = find_suey(
wsiMask, dirs, downsample, args, wsi)
print('\n\nthe cropped regions have been saved to: ' + dirs['outDir'] +
fileID + dirs['img_save_dir'] + fileID + dirs['crop_dir'])
# call DeepLab to predict Glom boundaries (High resolution)
print('\ngetting Glom boundaries ...\n')
make_folder(dirs['outDir'] + fileID + dirs['final_output_dir'] +
'prediction')
test_data_list = fileID + '_crops.txt'
modeldir = args.base_dir + '/' + args.project + dirs['modeldir'] + str(
iteration) + '/HR'
test_steps_HR = get_all_test_steps(modeldir)
if idx > len(test_steps_HR):
test_step = test_steps_HR[-1]
else:
test_step = test_steps_HR[idx]
print("\033[1;32;40m" + "starting prediction using model: \n\t" +
modeldir + str(test_step) + "\033[0;37;40m" + "\n\n")
call([
'python3', args.base_dir + '/Codes/Deeplab_network/main.py',
'--option', 'predict', '--test_data_list',
dirs['outDir'] + fileID + dirs['txt_save_dir'] + test_data_list,
'--out_dir', dirs['outDir'] + fileID + dirs['final_output_dir'],
'--test_step',
str(test_step), '--test_num_steps',
str(test_num_steps), '--modeldir', modeldir, '--data_dir',
dirs['outDir'] + fileID + dirs['img_save_dir'], '--num_classes',
str(classNum), '--gpu', '0'
])
print('\nsaving final glom images ...')
print('\nworking on:')
crop_suey(label_offsets, dirs, args, classNum, iteration, idx)
# clean up
if dirs['save_outputs'] == False:
print('cleaning up')
rmtree(dirs['outDir'] + fileID)
def predict_xml(args, dirs, wsi, iteration):
# reshape regions calc
downsample = int(args.downsampleRateLR**.5)
downsample_HR = int(args.downsampleRateHR**.5)
region_size = int(args.boxSizeLR*(downsample))
step = int(region_size*(1-args.overlap_percentLR))
# figure out the number of classes
if args.classNum == 0:
annotatedXMLs=glob(args.base_dir + '/' + args.project + dirs['training_data_dir'] + str(iteration-1) + '/*.xml')
classes = []
for xml in annotatedXMLs:
classes.append(get_num_classes(xml))
print(classes)
print(args.classNum)
classNum_LR = max(classes)
classNum_HR = max(classes)
else:
classNum_LR = args.classNum
if args.classNum_HR != 0:
classNum_HR = args.classNum_HR
else:
classNum_HR = classNum_LR
# chop wsi
fileID, test_num_steps, slide = chop_suey(wsi, dirs, downsample, region_size, step, args)
dirs['fileID'] = fileID
print('Chop SUEY!\n')
# call DeepLab for prediction (Low resolution)
print('finding Glom locations ...\n')
make_folder(dirs['outDir'] + fileID + dirs['img_save_dir'] + 'prediction')
test_data_list = fileID + '_images' + '.txt'
modeldir = args.base_dir + '/' + args.project + dirs['modeldir'] + str(iteration) + '/LR'
test_step = get_test_step(modeldir)
print("\033[1;32;40m"+"starting prediction using model: \n\t" + modeldir + str(test_step) + "\033[0;37;40m"+"\n\n")
call(['python3', args.base_dir+'/Codes/Deeplab_network/main.py',
'--option', 'predict',
'--test_data_list', dirs['outDir']+fileID+dirs['txt_save_dir']+test_data_list,
'--out_dir', dirs['outDir']+fileID+dirs['img_save_dir'],
'--test_step', str(test_step),
'--test_num_steps', str(test_num_steps),
'--modeldir', modeldir,
'--data_dir', dirs['outDir']+fileID+dirs['img_save_dir'],
'--num_classes', str(classNum_LR),
'--gpu', str(args.gpu)])
# un chop
print('\nreconstructing wsi map ...\n')
wsiMask = un_suey(dirs=dirs, args=args)
# save hotspots
if dirs['save_outputs'] == True:
make_folder(dirs['outDir'] + fileID + dirs['mask_dir'])
print('saving to: ' + dirs['outDir'] + fileID + dirs['mask_dir'] + fileID + '.png')
with warnings.catch_warnings():
warnings.simplefilter("ignore")
imsave(dirs['outDir'] + fileID + dirs['mask_dir'] + fileID + '.png', wsiMask/np.max(wsiMask))
# find glom locations in reconstructed map
print('\ninterpreting prediction map ...')
test_num_steps, labeledArray, label_offsets = find_suey(wsiMask, dirs, downsample, args, wsi)
print('\n\nthe cropped regions have been saved to: ' + dirs['outDir'] + fileID + dirs['img_save_dir'] + fileID + dirs['crop_dir'])
# call DeepLab to predict Glom boundaries (High resolution)
print('\ngetting Glom boundaries ...\n')
make_folder(dirs['outDir'] + fileID + dirs['final_output_dir'] + 'prediction')
test_data_list = fileID + '_crops.txt'
modeldir = args.base_dir + '/' + args.project + dirs['modeldir'] + str(iteration) + '/HR'
test_step = get_test_step(modeldir)
print("\033[1;32;40m"+"starting prediction using model: \n\t" + modeldir + str(test_step) + "\033[0;37;40m"+"\n\n")
call(['python3', args.base_dir+'/Codes/Deeplab_network/main.py',
'--option', 'predict',
'--test_data_list', dirs['outDir']+fileID+dirs['txt_save_dir']+test_data_list,
'--out_dir', dirs['outDir']+fileID+dirs['final_output_dir'],
'--test_step', str(test_step),
'--test_num_steps', str(test_num_steps),
'--modeldir', modeldir,
'--data_dir', dirs['outDir']+fileID+dirs['img_save_dir'],
'--num_classes', str(classNum_HR),
'--gpu', str(args.gpu)])
print('\nsaving final glom images ...')
print('\nstitching high resolution WSI mask:')
wsiMask_HR=unstitch_HR(dirs=dirs,args=args,wsi=wsi)
wsidims=wsiMask_HR.shape
d1=int(wsidims[0]*(1/args.approx_downsample))
d2=int(wsidims[1]*(1/args.approx_downsample))
if args.approx_downsample!=1:
print('\nDownsampling high resolution mask for prediction smoothing...')
wsiMask_HR=resize(wsiMask_HR,(d1,d2),order=0,anti_aliasing=False,preserve_range=True)
#mask_display=resize(wsiMask_HR,(d1,d2))
#plt.imshow(mask_display/np.max(wsiMask))
#plt.show()
print('\nGenerating XML annotations from WSI mask...')
make_folder(dirs['outDir'] + dirs['fileID'] + dirs['final_output_dir'] + dirs['final_boundary_image_dir'])
crop_suey(wsiMask_HR,label_offsets, dirs, args, classNum_HR, downsample_HR)
# clean up
if dirs['save_outputs'] == False:
print('\ncleaning up')
rmtree(dirs['outDir']+fileID)
def predict_xml(args, dirs, wsi, iteration):
# reshape regions calc
downsample = int(args.downsampleRateHR**.5)
region_size = int(args.boxSizeHR * (downsample))
step = int(region_size * (1 - args.overlap_percentHR))
# figure out the number of classes
if args.classNum == 0:
annotatedXMLs = glob(args.base_dir + '/' + args.project +
dirs['training_data_dir'] + str(iteration - 1) +
'/*.xml')
classes = []
for xml in annotatedXMLs:
classes.append(get_num_classes(xml))
classNum = max(classes)
else:
classNum = args.classNum
# chop wsi
fileID, test_num_steps = chop_suey(wsi, dirs, downsample, region_size,
step, args)
dirs['fileID'] = fileID
print('Chop SUEY!\n')
# call DeepLab for prediction
print('Segmenting tissue ...\n')
make_folder(dirs['outDir'] + fileID + dirs['img_save_dir'] + 'prediction')
test_data_list = fileID + '_images' + '.txt'
modeldir = args.base_dir + '/' + args.project + dirs['modeldir'] + str(
iteration) + '/HR'
test_step = get_test_step(modeldir)
print("\033[1;32;40m" + "starting prediction using model: \n\t" +
modeldir + '/' + str(test_step) + "\033[0;37;40m" + "\n\n")
call([
'python3', args.base_dir + '/Codes/Deeplab_network/main.py',
'--option', 'predict', '--test_data_list',
dirs['outDir'] + fileID + dirs['txt_save_dir'] + test_data_list,
'--out_dir', dirs['outDir'] + fileID + dirs['img_save_dir'],
'--test_step',
str(test_step), '--test_num_steps',
str(test_num_steps), '--modeldir', modeldir, '--data_dir',
dirs['outDir'] + fileID + dirs['img_save_dir'], '--num_classes',
str(classNum), '--gpu',
str(args.gpu)
])
# un chop
print('\nreconstructing wsi map ...\n')
wsiMask = un_suey(dirs=dirs, args=args)
# save hotspots
if dirs['save_outputs'] == True:
make_folder(dirs['outDir'] + fileID + dirs['mask_dir'])
print('saving to: ' + dirs['outDir'] + fileID + dirs['mask_dir'] +
fileID + '.png')
with warnings.catch_warnings():
warnings.simplefilter("ignore")
imsave(
dirs['outDir'] + fileID + dirs['mask_dir'] + fileID + '.png',
wsiMask)
print('\n\nStarting XML construction: ')
xml_suey(wsiMask=wsiMask,
dirs=dirs,
args=args,
classNum=classNum,
downsample=downsample)
# clean up
if dirs['save_outputs'] == False:
print('cleaning up')
rmtree(dirs['outDir'] + fileID)
def predict_xml(args, dirs, wsi, iteration):
# reshape regions calc
downsample = int(args.downsampleRateHR**.5)
region_size = int(args.boxSizeHR*(downsample))
step = int(region_size*(1-args.overlap_percentHR))
# figure out the number of classes
if args.classNum == 0:
annotatedXMLs=glob(args.base_dir + '/' + args.project + dirs['training_data_dir'] + str(iteration-1) + '/*.xml')
classes = []
for xml in annotatedXMLs:
classes.append(get_num_classes(xml))
classNum = max(classes)
else:
classNum = args.classNum
# chop wsi
if args.chop_data == 'True':
# chop wsi
fileID, test_num_steps = chop_suey(wsi, dirs, downsample, region_size, step, args)
dirs['fileID'] = fileID
print('Chop SUEY!\n')
else:
basename = os.path.splitext(wsi)[0]
if wsi.split('.')[-1] != 'tif':
slide=getWsi(wsi)
# get image dimensions
dim_x, dim_y=slide.dimensions
else:
im = Image.open(wsi)
dim_x, dim_y=im.size
fileID=basename.split('/')
dirs['fileID'] = fileID=fileID[len(fileID)-1]
test_num_steps = file_len(dirs['outDir'] + fileID + dirs['txt_save_dir'] + fileID + '_images' + ".txt")
# call DeepLab for prediction
print('Segmenting tissue ...\n')
make_folder(dirs['outDir'] + fileID + dirs['img_save_dir'] + 'prediction')
test_data_list = fileID + '_images' + '.txt'
modeldir = args.base_dir + '/' + args.project + dirs['modeldir'] + str(iteration) + '/HR'
test_step = get_test_step(modeldir)
print("\033[1;32;40m"+"starting prediction using model: \n\t" + modeldir + '/' + str(test_step) + "\033[0;37;40m"+"\n\n")
call(['python3.5', args.base_dir+'/Codes/Deeplab_network/main.py',
'--option', 'predict',
'--test_data_list', dirs['outDir']+fileID+dirs['txt_save_dir']+test_data_list,
'--out_dir', dirs['outDir']+fileID+dirs['img_save_dir'],
'--test_step', str(test_step),
'--test_num_steps', str(test_num_steps),
'--modeldir', modeldir,
'--data_dir', dirs['outDir']+fileID+dirs['img_save_dir'],
'--num_classes', str(classNum),
'--gpu', str(args.gpu),
'--encoder_name',args.encoder_name])
# un chop
print('\nreconstructing wsi map ...\n')
wsiMask = un_suey(dirs=dirs, args=args)
# save hotspots
if dirs['save_outputs'] == True:
#reduce the resolution of the image
wsidims=wsiMask.shape
wsiMask_save=resize(wsiMask,(int(wsidims[0]/4),int(wsidims[1]/4)),order=0,preserve_range=True)
make_folder(dirs['outDir'] + fileID + dirs['mask_dir'])
print('saving to: ' + dirs['outDir'] + fileID + dirs['mask_dir'] + fileID + '.png')
with warnings.catch_warnings():
warnings.simplefilter("ignore")
imsave(dirs['outDir'] + fileID + dirs['mask_dir'] + fileID + '.png', wsiMask_save)
print('\n\nStarting XML construction: ')
xml_suey(wsiMask=wsiMask, dirs=dirs, args=args, classNum=classNum, downsample=downsample)
# clean up
if dirs['save_outputs'] == False:
print('cleaning up')
rmtree(dirs['outDir']+fileID)