mrunetpred_gts.py 15 KB
Newer Older
sjromuel's avatar
d  
sjromuel committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
import tkinter as tk
import os
import matplotlib
from tkinter import filedialog
from skimage import transform
import SimpleITK as sitk
import argparse
#import os
#import pydot
#from graphviz import Digraph
#import shutil
#from tensorflow.keras import layers, models
#from utils.dataLoader import *
from utils.other_functions import *


def main():

    root = tk.Tk()
    root.withdraw()
sjromuel's avatar
d    
sjromuel committed
24
    file_path = filedialog.askopenfilename(initialdir="saves/mr_unet_cv_ctthresh")
sjromuel's avatar
d  
sjromuel committed
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
    file_path = file_path[:-9]
    print(file_path)

    ##################### U-Net #####################
    if 'Unet' in file_path or 'Cluster' in file_path or 'class' in file_path:
        from nets.Unet import run_test_patient, show_test_patient_pred, dice_loss, Unet
        print("UNet_model selected")

        ### read out files ###
        weights = np.load(file_path+"model.npy", allow_pickle=True)
        [test_patients,
         val_patients,
         number_patients,
         img_path,
         shrink_data,
         newSize,
         lr,
         batch_size,
         num_epochs,
         e,
         augment,
         save_path,
         gt_type,
         filter_multiplier] = np.load(file_path+"params.npy", allow_pickle=True)
        #  autoencoder_model__e100_switchclass1024_nohiddenclusternet needs to comment out gt_type and val_patients
        print('Training Parameters:')
        print('-----------------')
        print('Number of Patients: ', number_patients)
        print('Number of epochs: ', num_epochs)
        print('Test Patient number: ', test_patients)
        print('Image Size: ', newSize)
        print('Filter Multiplier: ', filter_multiplier)
        print('Data Augmentation: ', augment)
        print('Learning rate: ', lr)
        print('Image Path: ', img_path)
        print('Save Path: ', save_path)
        print('GT Type:', gt_type)

        ### Load test patient
        if gt_type == "thresh" or gt_type == "ctthresh_gt":
            img_path = "data/npy_thresh/"
        else:
            img_path = "data/npy/"
        full_list = os.listdir(img_path)
        seg_list = os.listdir("data/npy/")

        X_img_list = []
        GT_img_list = []
        ytrue_img_list = []
        # thresh_img_list = []
        if "mr" in save_path:
            for elem in full_list:
                if elem.endswith("T1.gipl.npy") and (elem.startswith('P' + str(test_patients[0]).zfill(2)) or elem.startswith('P' + str(test_patients[1]).zfill(2))):
                    X_img_list.append(elem)

                elif elem.endswith("250_mr_ctthresh_gt.gipl.npy") and (elem.startswith('P' + str(test_patients[0]).zfill(2)) or elem.startswith('P' + str(test_patients[1]).zfill(2))):
                    GT_img_list.append(elem)
            for elem in seg_list:
                if elem.endswith("segmr.gipl.npy") and (elem.startswith('P' + str(test_patients[0]).zfill(2)) or elem.startswith('P' + str(test_patients[1]).zfill(2))):
                    ytrue_img_list.append(elem)


        list.sort(X_img_list)
        list.sort(GT_img_list)
        list.sort(ytrue_img_list)
        print("Input Image List", X_img_list)
        print("GT Image List", GT_img_list)
        print("True Segmentation Image List", ytrue_img_list)


        img_savepath = file_path + "imgresults/"
        print(img_savepath)
        if not os.path.exists(img_savepath):
            os.makedirs(img_savepath)

        for j in range(2):
            X_img_npys = np.load(img_path + X_img_list[j])
            GT_img_npys = np.load(img_path + GT_img_list[j])
            ytrue_img_npys = np.load(img_path + ytrue_img_list[j])

            print("Input shape: ", np.shape(X_img_npys))
            print("GT shape: ", np.shape(GT_img_npys))
            print("True Segm shape: ", np.shape(ytrue_img_npys))



            X_img_npys = transform.resize(X_img_npys, (X_img_npys.shape[0], newSize[0], newSize[1]), order=0,
                                          preserve_range=True, mode='constant', anti_aliasing=False,
                                          anti_aliasing_sigma=None)
            GT_img_npys = transform.resize(GT_img_npys, (GT_img_npys.shape[0], newSize[0], newSize[1]), order=0,
                                           preserve_range=True, mode='constant', anti_aliasing=False,
                                           anti_aliasing_sigma=None)
            ytrue_img_npys = transform.resize(ytrue_img_npys, (ytrue_img_npys.shape[0], newSize[0], newSize[1]),
                                              order=0,
                                              preserve_range=True, mode='constant', anti_aliasing=False,
                                              anti_aliasing_sigma=None)
            X_test = np.reshape(X_img_npys, (X_img_npys.shape[0], X_img_npys.shape[1], X_img_npys.shape[2], 1))
            GT_test = np.reshape(GT_img_npys, (GT_img_npys.shape[0], GT_img_npys.shape[1], GT_img_npys.shape[2], 1))
            ytrue = np.reshape(ytrue_img_npys,
                               (ytrue_img_npys.shape[0], ytrue_img_npys.shape[1], ytrue_img_npys.shape[2], 1))


            test_dataset = tf.data.Dataset.from_tensor_slices((X_test, ytrue))
            test_dataset = test_dataset.batch(batch_size=1)
            print(test_patients)
            TP_num = test_patients[j]

            ###################################################################################

            detailed_images = False
            npys3d          = True

            ###################################################################################

            test_loss = []
            test_loss_hdd = []
            test_loss_hdd2 = []
            counter=0
            y_pred3d = []
            for features in test_dataset:
                image, y_true = features
                y_true = onehotencode(y_true)
                y_pred = Unet(image, weights, filter_multiplier, training=False)

                if npys3d:
sjromuel's avatar
d    
sjromuel committed
150
                    y_predflip = np.flipud(y_pred)
sjromuel's avatar
d  
sjromuel committed
151
                    if y_pred3d == []:
sjromuel's avatar
d    
sjromuel committed
152
                        y_pred3d = y_predflip[:,:,:,0].numpy()
sjromuel's avatar
d  
sjromuel committed
153
                    else:
sjromuel's avatar
d    
sjromuel committed
154
                        y_pred3d =  np.append(y_predflip[:,:,:,0].numpy(), y_pred3d, axis=0)
sjromuel's avatar
d  
sjromuel committed
155
156
                    #print(np.shape(y_pred3d))

sjromuel's avatar
d    
sjromuel committed
157

sjromuel's avatar
d  
sjromuel committed
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179


                loss = dice_loss(y_pred, y_true)
                loss = tf.make_ndarray(tf.make_tensor_proto(loss))
                test_loss.append(loss)
                #print(test_loss)
                try:
                    y_true_np = np.squeeze(y_true[0, :, :, 0].numpy() > 0.5)
                    y_true_np = y_true_np.astype(np.float_)
                    pred_np = np.squeeze(y_pred[0, :, :, 0].numpy() > 0.5)
                    pred_np = pred_np.astype(np.float_)
                    hausdorff_distance_filter = sitk.HausdorffDistanceImageFilter()
                    hausdorff_distance_filter.Execute(sitk.GetImageFromArray(y_true_np), sitk.GetImageFromArray(pred_np))
                    test_loss_hdd.append(hausdorff_distance_filter.GetHausdorffDistance())

                    hausdorff_distance_filter2 = sitk.HausdorffDistanceImageFilter()
                    hausdorff_distance_filter2.Execute(sitk.GetImageFromArray(pred_np), sitk.GetImageFromArray(y_true_np))
                    test_loss_hdd2.append(hausdorff_distance_filter2.GetHausdorffDistance())
                except:
                    pass


sjromuel's avatar
d    
sjromuel committed
180

sjromuel's avatar
d  
sjromuel committed
181
182

            if npys3d:
sjromuel's avatar
d    
sjromuel committed
183
                y_pred3d = np.flip(y_pred3d)
sjromuel's avatar
d  
sjromuel committed
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
                #np.save(img_savepath + "P" + str(TP_num).zfill(2) + "_pred3D", y_pred3d)
                np.save("data/npy_thresh/" + "P" + str(TP_num).zfill(2) + "_mr_ctunetpred.gipl", y_pred3d)
                print("ypred shape: ", np.shape(y_pred3d))
            #plt.show()
            #print(test_loss)
            print("TestLoss Mean for P", test_patients[j], ": ", np.mean(test_loss))

            #print(test_loss_hdd)
            print("Hausdorff-Distance for P", test_patients[j],":", np.mean(test_loss_hdd))
            print("Hausdorff-Distance2 for P", test_patients[j], ":", np.mean(test_loss_hdd2))



        #####
        '''X_img_npys = np.load(img_path + X_img_list[0])
        GT_img_npys = np.load(img_path + GT_img_list[0])
        ytrue_img_npys = np.load("../data/npy/" + ytrue_img_list[0])
        for i in range(len(X_img_list)-1):
            X_img_npys = np.append(X_img_npys, np.load(img_path + X_img_list[i+1]), axis=0)
            GT_img_npys = np.append(GT_img_npys, np.load(img_path + GT_img_list[i+1]), axis=0)
            ytrue_img_npys = np.append(ytrue_img_npys, np.load("../data/npy/" + ytrue_img_list[i+1]), axis=0)


        X_img_npys = transform.resize(X_img_npys, (X_img_npys.shape[0], newSize[0], newSize[1]), order=0,
                                      preserve_range=True, mode='constant', anti_aliasing=False, anti_aliasing_sigma=None)
        GT_img_npys = transform.resize(GT_img_npys, (GT_img_npys.shape[0], newSize[0], newSize[1]), order=0,
                                      preserve_range=True, mode='constant', anti_aliasing=False, anti_aliasing_sigma=None)
        ytrue_img_npys = transform.resize(ytrue_img_npys, (ytrue_img_npys.shape[0], newSize[0], newSize[1]), order=0,
                                      preserve_range=True, mode='constant', anti_aliasing=False, anti_aliasing_sigma=None)
        X_test = np.reshape(X_img_npys,(X_img_npys.shape[0], X_img_npys.shape[1], X_img_npys.shape[2], 1))
        GT_test = np.reshape(GT_img_npys, (GT_img_npys.shape[0], GT_img_npys.shape[1], GT_img_npys.shape[2], 1))
        ytrue = np.reshape(ytrue_img_npys, (ytrue_img_npys.shape[0], ytrue_img_npys.shape[1], ytrue_img_npys.shape[2], 1))

        ### recreate test-dataset ###
        test_dataset = tf.data.Dataset.from_tensor_slices((X_test, GT_test))
        test_dataset = test_dataset.batch(batch_size)

        print("x_test shape: " + str(X_test.shape))

        ### test model and show plots ###
        test_patient_pred = run_test_patient(test_dataset, weights, filter_multiplier)
        print("test_patient_pred shape: " + str(test_patient_pred.shape))
        # show_test_patient_pred(test_patient_pred, test_dataset)

        print(GT_test.shape)
        ### compute loss for each slice ###


        test_loss = []
        test_loss_hdd = []
        for i in range(test_patient_pred.shape[0]):
            y_pred = tf.convert_to_tensor(test_patient_pred[i])
            y_true = tf.convert_to_tensor(ytrue[i])
            #print(tf.shape(y_pred), tf.shape(y_true))
            y_true = onehotencode(tf.reshape(y_true, (1, 512, 512, 1)), autoencoder=True)
            y_pred = tf.reshape(y_pred, (1, 512, 512, 2))
            loss = dice_loss(y_pred, y_true)
            loss = tf.make_ndarray(tf.make_tensor_proto(loss))
            test_loss.append(loss)

            gt_np = np.squeeze(y_true[0, :, :, 0].numpy() > 0)
            gt_np = gt_np.astype(np.float_)
            pred_np = np.squeeze(y_pred[0, :, :, 0].numpy() > 0)
            pred_np = pred_np.astype(np.float_)
            hausdorff_distance_filter = sitk.HausdorffDistanceImageFilter()
            hausdorff_distance_filter.Execute(sitk.GetImageFromArray(gt_np), sitk.GetImageFromArray(pred_np))
            test_loss_hdd.append(hausdorff_distance_filter.GetHausdorffDistance())

        print(test_loss)
        print("TestLoss Mean: ", np.mean(test_loss))

        print(test_loss_hdd)
        print("Hausdorff-Distance: ", np.mean(test_loss_hdd))
        #show_loss_plot(len(test_loss), test_loss)



        print("test_patient_pred shape: " + str(test_patient_pred.shape))'''
        #show_test_patient_pred(test_patient_pred, test_dataset)

    ################## Autoencoder ##################

    elif 'AE' in file_path:
        # from nets.AE_3conv32 import run_test_patient, show_test_patient_pred, dice_loss
        # from nets.AE_3conv128 import run_test_patient, show_test_patient_pred, dice_loss
        # from nets.AE_3conv1024 import run_test_patient, show_test_patient_pred, dice_loss
        # from nets.AE_4conv2048 import run_test_patient, show_test_patient_pred, dice_loss
        from nets.AE_ulike import run_test_patient, show_test_patient_pred, dice_loss

        print("Autoencoder_model selected")
        ### read out files ###
        weights = np.load(file_path + "model.npy", allow_pickle=True)
        [test_patients,
         val_patients,
         number_patients,
         img_path,
         shrink_data,
         newSize,
         lr,
         batch_size,
         num_epochs,
         e,
         augment,
         save_path,
         gt_type,
         filter_multiplier] = np.load(file_path + "params.npy", allow_pickle=True)
        print('Training Parameters:')
        print('-----------------')
        print('Number of Patients: ', number_patients)
        print('Number of epochs: ', num_epochs)
        print('Test Patient numbers: ', test_patients)
        # print('Validation Patient numbers: ', val_patients)
        print('Image Size: ', newSize)
        print('Data Augmentation: ', augment)
        print('Learning rate: ', lr)
        print('Image Path: ', img_path)
        print('Save Path: ', save_path)
        print('GT Type: ', gt_type)

        ### Load test patient
        if gt_type == "thresh":
            img_path = "data/npy_thresh/"
        else:
            img_path = "data/npy/"

        full_list = os.listdir(img_path)
        X_img_list = []
        print(test_patients)
        for elem in full_list:
            if elem.endswith(gt_type+".gipl.npy") and (elem.startswith('P' + str(test_patients[0]).zfill(2)) or elem.startswith('P' + str(test_patients[1]).zfill(2))):
                X_img_list.append(elem)
                if gt_type == "thresh":
                    X_img_list.append(elem)
                    X_img_list.append(elem)
        print(X_img_list)
        test_patients_npys = np.append(np.load(img_path + X_img_list[0]), np.load(img_path + X_img_list[1]), axis=0)
        x_test = np.reshape(test_patients_npys, (test_patients_npys.shape[0], test_patients_npys.shape[1], test_patients_npys.shape[2], 1))

        ### recreate test-dataset ###
        test_dataset = tf.data.Dataset.from_tensor_slices((x_test, x_test))
        test_dataset = test_dataset.batch(batch_size=1)

        ### test model and show plots ###
        test_patient_pred = run_test_patient(test_dataset, weights)

        test_loss = []
        for i in range(test_patient_pred.shape[0]):
            y_pred = tf.convert_to_tensor(test_patient_pred[i])
            y_true = tf.convert_to_tensor(x_test[i])
            y_pred = tf.reshape(y_pred, (1, 512, 512, 2))
            y_true = onehotencode(tf.reshape(y_true, (1, 512, 512, 1)), autoencoder=True)
            loss = dice_loss(y_pred, y_true)
            loss = tf.make_ndarray(tf.make_tensor_proto(loss))
            test_loss.append(loss)

        print('Test losses: ', test_loss)
        print('Test loss-mean: ', np.mean(test_loss))

        print("test_patient_pred shape: " + str(test_patient_pred.shape))
        show_test_patient_pred(test_patient_pred, test_dataset)


def show_loss_plot(slice, loss_list):
    x_axis = list(range(1, slice+1, 1))
    plt.plot(x_axis, loss_list)
    plt.xlabel('Slice Number')
    plt.ylabel('Loss')
    plt.show()


if __name__ == "__main__":
    main()