Newer
Older
# -*- coding: utf-8 -*-
"""
Created on Wed Feb 21 16:43:17 2018
This script is utilized for creating the data
data['image'] = image [1, Number_of_Image]
data['label'] = label [1, Number_of_Image]
data['edge'] = Edge [1, Number_of_Image]
data['boundingbox'] = Bounding_Box [1, Number_of_Image, Maximum_Num_of_Instances(30), 4](x_min,x_max,y_min,y_max)
@author: s161488
"""
import os
import numpy as np
import scipy
from scipy import misc
from skimage.morphology import dilation, disk
from scipy import ndimage
import matplotlib.pyplot as plt
21
22
23
24
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
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
# path_mom = "/Users/bo/Documents/Exp_Data/"
path_use = path_mom + 'Data'
if not os.path.exists(path_use):
os.makedirs(path_use)
def get_filename_train_list():
path = path_mom + '/gland_data/'
all_file = os.listdir(path)
train_label_filename = [path + filename for filename in sorted(all_file)
if filename.startswith("train") & filename.endswith("_anno.bmp")]
train_image_filename = [filename.strip().split("_anno")[0] + ".bmp" for filename in train_label_filename]
print("Total number of training images:", np.size(train_image_filename))
return train_image_filename, train_label_filename
def get_filename_test_list():
path = path_mom + '/gland_data/'
all_file = os.listdir(path)
test_a_label_filename = [path + filename for filename in sorted(all_file)
if filename.startswith("testA") & filename.endswith("_anno.bmp")]
test_a_image_filename = [filename.strip().split("_anno")[0] + ".bmp" for filename in test_a_label_filename]
test_b_label_filename = [path + filename for filename in sorted(all_file)
if filename.startswith("testB") & filename.endswith("_anno.bmp")]
test_b_image_filename = [filename.strip().split("_anno")[0] + ".bmp" for filename in test_b_label_filename]
print("Total number of testA image:", np.size(test_a_image_filename))
print("Total number of testB image:", np.size(test_b_image_filename))
return test_a_image_filename, test_a_label_filename, test_b_image_filename, test_b_label_filename
def read_gland_training_data(im_list, la_list):
"""This function is utilized to read the image and label
"""
images = []
labels = []
images_index = []
index = 0
for im_filename, la_filename in zip(im_list, la_list):
im = scipy.misc.imread(im_filename)
im_index = int(im_filename.strip().split("train_")[1].strip().split(".bmp")[0])
la = scipy.misc.imread(la_filename)
images_index.append(im_index)
images.append(im)
labels.append(la)
index = index + 1
print('%d Gland Training images are loaded' % index)
return images, labels, images_index
def read_gland_test_data(im_list, la_list, name):
"""This function is utilized to read the image and label
name: "testA_" or "testB_"
"""
images = []
labels = []
images_index = []
index = 0
for im_filename, la_filename in zip(im_list, la_list):
im = scipy.misc.imread(im_filename)
im_index = int(im_filename.strip().split(name)[1].strip().split(".bmp")[0])
la = scipy.misc.imread(la_filename)
images_index.append(im_index)
images.append(im)
labels.append(la)
index = index + 1
print('%6.1f Gland test %s images are loaded' % (index, name))
return images, labels, images_index
def extract_edge(label):
"""This function is utilized to extract the edge from the ground truth
Args:
label: The ground truth of all images
shape [Number_of_image, Image_Height, Image_Width,1]
Returns:
The edge feature map. If the pixel belongs to edge, then the label is set to be one.
If the pixel doesn't belong to edge, then the label is set to be zero.
shape: [Number_of_image, Image_height, Image_Width,1]
The requirement for this function is scipy!
"""
selem = disk(3)
edge_feat = []
for la_sep in label:
sx = ndimage.sobel(la_sep, axis=0, mode='constant')
sy = ndimage.sobel(la_sep, axis=1, mode='constant')
sob = np.hypot(sx, sy)
row = (np.reshape(sob, -1) > 0) * 1
edge_sep = np.reshape(row, np.shape(sob))
edge_sep = dilation(edge_sep, selem)
edge_feat.append(edge_sep.astype('int64'))
return edge_feat
def extract_benign_malignant(test_name):
path = path_mom + '/gland_data/Grade.csv'
fd = open(path)
index = []
class_index = []
for i in fd:
i = i.strip().split("\r")
for j in i:
j = j.strip().split(",")
train_index = [k for k in j if test_name in k]
if train_index:
index.append(int(train_index[0].split("_")[-1]))
for k in j:
if k.endswith("ign"):
class_index.append(1)
if k.endswith("nant"):
class_index.append(2)
return index, class_index
def transfer_data_to_dict():
"""This function is utilized to save the original image in a dictionary
Return:
data['image'] = image [1, Number_of_Image*4] 85*4
data['label'] = label [1, Number_of_Image*4]
data['edge'] = Edge [1, Number_of_Image*4]
Requirements:
from collections import defaultdict
"""
from collections import defaultdict
tr_im, tr_la = get_filename_train_list()
image, label, image_index = read_gland_training_data(tr_im, tr_la)
im_ind, class_index = extract_benign_malignant('train')
cla_ind_fin = []
for i in range(np.shape(image_index)[0]):
cla_ind_fin.append(class_index[int(np.where(np.array(im_ind) == image_index[i])[0])])
edge = extract_edge(label)
data = defaultdict(list)
data['image'] = image
data['label'] = label
data['edge'] = edge
data['ImageIndex'] = image_index
data['ClassIndex'] = cla_ind_fin
filename = path_mom + "/Data/glanddata.npy"
if os.path.isfile(filename):
print("Remove the existing data file", os.remove(filename))
print("Saving the data in path:", filename.split(".")[0])
else:
print("Oh, this is the first time of creating this file")
171
172
173
174
175
176
177
178
179
180
181
182
183
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
np.save(filename.split(".")[0], data)
# The standardeivation for the boudnign box ([ 158.73026619, 241.80872181, 159.79253117, 240.67909876])
def transfer_data_to_dict_test():
"""This function is utilized to save the test image in a dictionary
Return:
data['image'] = image [1, Number_of_Image*4] 85*4
data['label'] = label [1, Number_of_Image*4]
data['edge'] = Edge [1, Number_of_Image*4]
Requirements:
from collections import defaultdict
"""
from collections import defaultdict
te_a_im, te_a_la, te_b_im, te_b_la = get_filename_test_list()
imagea, labela, image_indexa = read_gland_test_data(te_a_im, te_a_la, "testA_")
im_inda, class_indexa = extract_benign_malignant("testA")
cla_ind_fina = []
for i in range(np.shape(image_indexa)[0]):
cla_ind_fina.append(class_indexa[int(np.where(np.array(im_inda) == image_indexa[i])[0])])
imageb, labelb, image_indexb = read_gland_test_data(te_b_im, te_b_la, "testB_")
im_indb, class_indexb = extract_benign_malignant("testB")
cla_ind_finb = []
for i in range(np.shape(image_indexb)[0]):
cla_ind_finb.append(class_indexb[int(np.where(np.array(im_indb) == image_indexb[i])[0])])
image_benign = []
label_benign = []
image_index_benign = []
image_mali = []
label_mali = []
image_index_mali = []
for index, class_single in enumerate(cla_ind_fina):
if class_single == 1:
image_benign.append(imagea[index])
label_benign.append(labela[index])
image_index_benign.append(image_indexa[index])
else:
image_mali.append(imagea[index])
label_mali.append(labela[index])
image_index_mali.append(image_indexa[index])
for index, class_single in enumerate(cla_ind_finb):
if class_single == 1:
image_benign.append(imageb[index])
label_benign.append(labelb[index])
image_index_benign.append(image_indexb[index])
else:
image_mali.append(imageb[index])
label_mali.append(labelb[index])
image_index_mali.append(image_indexb[index])
edge_benign = extract_edge(label_benign)
edge_mali = extract_edge(label_mali)
cla_ind_benign = np.repeat(1, 37)
cla_ind_mali = np.repeat(2, 43)
data = defaultdict(list)
data['image'] = image_benign
data['label'] = label_benign
data['edge'] = edge_benign
data['ImageIndex'] = image_index_benign
data['ClassIndex'] = cla_ind_benign
filename = path_mom + "/Data/glanddata_test_benign.npy"
if os.path.isfile(filename):
print("Remove the existing data file", os.remove(filename))
print("Saving the data in path:", filename.split(".")[0])
else:
print("Saving the data in path:", filename.split(".")[0])
np.save(filename.split(".")[0], data)
data1 = defaultdict(list)
data1['image'] = image_mali
data1['label'] = label_mali
data1['edge'] = edge_mali
data1['ImageIndex'] = image_index_mali
data1['ClassIndex'] = cla_ind_mali
filename = path_mom + "/Data/glanddata_test_mali.npy"
if os.path.isfile(filename):
print("Remove the existing data file", os.remove(filename))
print("Saving the data in path:", filename.split(".")[0])
else:
print("Creating GlaS test data (malignant)")
print("Saving the data in path:", filename.split(".")[0])