#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Wed Oct 31 09:05:57 2018
@author: Josefine
"""
import tensorflow as tf
import numpy as np
import nibabel as nib
import glob
import re
from skimage.transform import resize
imgDim = 128
##############################################################################
### Data functions ######
##############################################################################
def natural_sort(l):
convert = lambda text: int(text) if text.isdigit() else text.lower()
alphanum_key = lambda key: [ convert(c) for c in re.split('([0-9]+)', key) ]
return sorted(l, key = alphanum_key)
def create_image(filename_img,direction):
images = []
a = nib.load(filename_img)
a = a.get_data()
# Normalize:
a2 = np.clip(a,-1000,1000)
a3 = np.interp(a2, (a2.min(), a2.max()), (-1, +1))
# Reshape:
img = np.zeros([512,512,512])+np.min(a3)
index1 = int(np.ceil((512-a.shape[2])/2))
index2 = int(512-np.floor((512-a.shape[2])/2))
img[:,:,index1:index2] = a3
im = resize(img,(imgDim,imgDim,imgDim),order=0)
if direction == 'sag':
for i in range(im.shape[0]):
images.append((im[i,:,:]))
if direction == 'cor':
for i in range(im.shape[1]):
images.append((im[:,i,:]))
if direction == 'axial':
for i in range(im.shape[2]):
images.append((im[:,:,i]))
images = np.asarray(images)
images = images.reshape(-1, imgDim,imgDim,1)
return images
# Load test data
filelist_test = natural_sort(glob.glob('WHS/ct_train_test/ct_test/*_image.nii.gz')) # list of file names
# Load train data for segmentation network
filelist_train = natural_sort(glob.glob('WHS/Augment_data/*_image.nii')) # list of file names
##############################################################################
### Reload network and predict ######
##############################################################################
print("====================== LOAD SAGITTAL NETWORK: ===========================")
# Doing predictions with the model
tf.reset_default_graph()
new_saver = tf.train.import_meta_graph('WHS/Results/region/model_sag/model.ckpt.meta')
with tf.Session() as sess:
new_saver.restore(sess, tf.train.latest_checkpoint('WHS/Results/region/model_sag/'))
graph = tf.get_default_graph()
x = graph.get_tensor_by_name("x_train:0")
op_to_restore = graph.get_tensor_by_name("output/Softmax:0") #ME
#
# for i in range(len(filelist_test)):
# print('Processing test image', (i+1),'out of',(np.max(range(len(filelist_test)))+1))
# # Find renderings corresponding to the given name
# prob_maps = []
# x_test = create_image(filelist_test[i],'sag')
# for k in range(x_test.shape[0]):
# x_test_image = np.expand_dims(x_test[k,:,:,:], axis=0)
# y_output = sess.run(tf.nn.softmax(op_to_restore), feed_dict={x: x_test_image,'Placeholder:0':1.0})
# prob_maps.append(y_output[0,:,:,:])
# np.savez('WHS/Results/Predictions/region/test_prob_maps_sag_{}'.format(i),prob_maps=prob_maps)
# print("================ DONE WITH TEST PREDICTIONS! ==================")
for i in range(30,len(filelist_train)):
print('Processing test image', (i+1),'out of',(np.max(range(len(filelist_train)))+1))
# Find renderings corresponding to the given name
prob_maps = []
x_test = create_image(filelist_train[i],'sag')
for k in range(x_test.shape[0]):
x_test_image = np.expand_dims(x_test[k,:,:,:], axis=0)
y_output = sess.run(tf.nn.softmax(op_to_restore), feed_dict={x: x_test_image,'Placeholder:0':1.0})
prob_maps.append(y_output[0,:,:,:])
np.savez('WHS/Results/Predictions/region/train_prob_maps_sag_{}'.format(i),prob_maps=prob_maps)
print("================ DONE WITH TRAIN PREDICTIONS! ==================")
print("================ DONE WITH AXIAL PREDICTIONS! ==================")
Datasets
Models
