-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathmerge_isomaps.py
executable file
·253 lines (197 loc) · 9.43 KB
/
merge_isomaps.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
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
170
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
239
240
241
242
243
244
245
246
247
248
249
250
251
252
#!/usr/bin/env python3.5
import sys, os
import numpy as np
import cv2
import glob
ISOMAP_SIZE = 256
isomap_paths = glob.glob('/user/HS204/m09113/my_project_folder/IJB_A/multi_iter75_reg30/540/*isomap.png')
isomap_merged = None
def merge( isomaps):
merge = np.zeros([ISOMAP_SIZE, ISOMAP_SIZE, 4], dtype='uint32')
threshould_angle = 90 #between 0 (facing camera) and 90 (edge to camera)
threshould_alpha = (-255.0 / 90.0) * threshould_angle + 255.0
confidence_maps = np.zeros([ISOMAP_SIZE, ISOMAP_SIZE, len(isomaps)], dtype='uint32')
for i in range(len(isomaps)):
confidence_maps[:,:,i] = isomaps[i][:,:,3].astype(dtype='uint32')
rule_mask = confidence_maps[:,:,i] < threshould_alpha
#confidence_maps[:,:,i] = np.exp(confidence_maps[:,:,i]/50)
confidence_maps[rule_mask, i] = 0
#confidence_sum = confidence1+confidence2
confidence_sum = np.sum(confidence_maps,2)
merge[:,:,3] = confidence_sum
conf_over_0 = merge[:,:,3] > 0
#merge[conf_over_0,:3] = (isomap1[conf_over_0,:3]*confidence1[conf_over_0,None]+isomap2[conf_over_0,:3]*confidence2[conf_over_0,None])/(confidence1[conf_over_0,None]+confidence2[conf_over_0,None])
for i in range(len(isomaps)):
merge[:,:,:3] += (isomaps[i][:,:,:3]/confidence_sum[:,:,None]*confidence_maps[:,:,i,None]).astype(dtype='uint32')
#merge[:,:,:3] = isomaps[0][:,:,:3]/confidence_sum[:,:,None]*confidence_maps[:,:,0,None]+isomaps[1][:,:,:3]/confidence_sum[:,:,None]*confidence_maps[:,:,1,None]#confidence2[:,:,None]
conf_over_255 = merge[:,:,3] > 255
merge[conf_over_255,3] = 255
return merge.astype(dtype="uint8")
def merge_sm_with_tf(isomap_lists, confidence_lists, output_list):
import tensorflow as tf
import cnn_tf_graphs
from shutil import copyfile
#zipped_input = zip(isomap_lists, confidence_lists, output_list)
#zipped_input.sort(key=lambda x: len(x[0]))
#isomap_lists, confidence_lists, output_list = zip(*zipped_input)
sorted_idx_list = sorted(range(len(isomap_lists)), key=lambda x: len(isomap_lists[x]))
#print (sorted_idx_list)
isomap_lists = [isomap_lists[i] for i in sorted_idx_list]
confidence_lists = [confidence_lists[i] for i in sorted_idx_list]
output_list = [output_list[i] for i in sorted_idx_list]
#print ('HHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH')
#for i in range(len(isomap_lists)):
# print (isomap_lists[i])
# print (confidence_lists[i])
# print (output_list[i])
#isomap_lists.sort(key=len)
merge_legth = -1
sess = None
for j, isomap_list in enumerate(isomap_lists):
with tf.Graph().as_default():
if len(isomap_list) == 0:
continue
elif len(isomap_list) ==1:
copyfile(isomap_list[0],output_list[j])
else:
if len(isomap_list) != merge_legth:
if sess:
sess.close()
placeholders = []
outpath = tf.placeholder(tf.string)
for i in range(len(isomap_list)):
colour = tf.placeholder(tf.float32, shape=(1, ISOMAP_SIZE, ISOMAP_SIZE, 3))
conf = tf.placeholder(tf.float32, shape=(1, ISOMAP_SIZE, ISOMAP_SIZE, 1))
placeholders.append([colour, conf])
merged = tf.squeeze(cnn_tf_graphs.merge_isomaps_softmax(placeholders))
merged_uint8 = tf.cast(merged, tf.uint8)
encoded = tf.image.encode_png(merged_uint8)
write_file_op = tf.write_file(outpath, encoded)
merge_legth = len(isomap_list)
sess = tf.Session()
print ('merging',merge_legth,'images (max',len(isomap_lists[-1]),') idx',j,'of',len(isomap_lists))
feed_dict = {}
for i in range(len(isomap_list)):
feed_dict[placeholders[i][0]] = np.expand_dims(cv2.imread(isomap_list[i], cv2.IMREAD_UNCHANGED)[:,:,:3].astype(np.float32)[:,:,::-1], axis=0)
feed_dict[placeholders[i][1]] = np.expand_dims(np.load(confidence_lists[j][i]).astype(np.float32), axis=0)
feed_dict[outpath] = output_list[j]
sess.run(write_file_op, feed_dict=feed_dict)
def write_cnn_confidences(saved_model_path, images, output_paths):
import tensorflow as tf
import tf_utils
import cnn_tf_graphs
with tf.device('/gpu:0'):
with tf.Graph().as_default():
image_path_tensor = tf.placeholder(tf.string)
image_tf = tf_utils.single_input_image(image_path_tensor, os.path.dirname(saved_model_path)+'/../db_input/total_mean.png' ,png_with_alpha=True, image_size=256)
image_tf = tf.expand_dims(image_tf,0)
# Build a Graph that computes the logits predictions from the inference model.
with tf.variable_scope("confidence_estimation") as scope:
confidence_map = cnn_tf_graphs.confidence_cnn13(image_tf, input_size=256)
saver = tf.train.Saver()
config = tf.ConfigProto( allow_soft_placement=False, log_device_placement=False)
config.gpu_options.allow_growth = True
with tf.Session(config=config) as sess:
print('restore model')
saver.restore(sess, saved_model_path)
print ('restoring done')
#print('we have',db_loader.num_examples_eval, 'images to evaluate')
for idx, image_path in enumerate(images):
if idx%1000==0:
print (idx,'of',len(images))
confidence_pic = sess.run(confidence_map, feed_dict={image_path_tensor: image_path})
confidence_output_path = output_paths[idx]
np.save(confidence_output_path, confidence_pic[0,...])
def calc_isomap_coverage(isomap):
visible = isomap[:,:,3]>0
pixels = isomap.shape[0]*isomap.shape[1]
return np.sum(visible)/pixels
def isomap_playground():
isomaps =[]
for i in range(len(isomap_paths)):
isomaps.append(cv2.imread(isomap_paths[i], cv2.IMREAD_UNCHANGED))
old_isomap_merged = np.zeros([ISOMAP_SIZE, ISOMAP_SIZE, 4], dtype='uint8')
all_isomaps_merged = merge(isomaps)
show_isomap('all_isomaps_merged', all_isomaps_merged)
#cv2.waitKey()
#cv2.destroyAllWindows()
#exit()
for i in range(len(isomaps)):
new_isomap_merged = merge([old_isomap_merged, isomaps[i]])
#blurryness = cv2.Laplacian(isomaps[i], cv2.CV_64F).var()
blurryness_map = cv2.Laplacian(isomaps[i], cv2.CV_64F)
blurryness_map[np.logical_or(blurryness_map<-700, blurryness_map>700)]=0 #try to filter out the edges
blurryness = blurryness_map.var()
#show_isomap('laplac',cv2.Laplacian(isomaps[i], cv2.CV_8U))
#print ('max', np.max(cv2.Laplacian(isomaps[i], cv2.CV_64F)), 'min', np.min(cv2.Laplacian(isomaps[i], cv2.CV_64F)))
coverage = calc_isomap_coverage(isomaps[i])
print(isomap_paths[i]," isomap coverage:",coverage,"blur detection:",blurryness, "overall score", coverage*coverage*blurryness)
show_isomap('new isomap', isomaps[i])
show_isomap('merge', new_isomap_merged)
cv2.waitKey()
old_isomap_merged = new_isomap_merged
#cv2.imwrite('/user/HS204/m09113/Desktop/merge_test.png', isomap_merged)
#cv2.waitKey()
#cv2.destroyAllWindows()
def show_isomap(window, isomap):
#isomap_copy = isomap.copy()
background = np.zeros([ISOMAP_SIZE, ISOMAP_SIZE, 4], dtype='uint8')
background[:,:,3]=10
mask = np.array([[int(x/8) %2==int(y/8) %2 for x in range(isomap.shape[0])] for y in range(isomap.shape[1])])
#mask = np.array([[int(x/8) %2==0 for x in range(isomap.shape[0])] for y in range(isomap.shape[1])])
background[mask,:3]=[200,200,200]
mask = np.invert(mask)
background[mask,:3]=[150,150,150]
cv2.imshow(window, merge([background,isomap]))
def merge_isomaps_pg():
isomap_paths = ['/user/HS204/m09113/my_project_folder/Boris/new_isomaps/image-00058.isomap.png', '/user/HS204/m09113/my_project_folder/Boris/new_isomaps/image-00456.isomap.png']
isomaps =[]
for i in range(len(isomap_paths)):
isomaps.append(cv2.imread(isomap_paths[i], cv2.IMREAD_UNCHANGED))
isomap_merged = merge(isomaps)
cv2.imwrite('/user/HS204/m09113/my_project_folder/Boris/new_isomaps/merged.png', isomap_merged[:,:,:3])
def pseudocolor(val, minval, maxval):
# from here: http://stackoverflow.com/questions/10901085/range-values-to-pseudocolor
import colorsys
# convert val in range minval..maxval to the range 0..120 degrees which
# correspond to the colors red..green in the HSV colorspace
h = (float(val-minval) / (maxval-minval)) * 240
#h=240-h
# convert hsv color (h,1,1) to its rgb equivalent
# note: the hsv_to_rgb() function expects h to be in the range 0..1 not 0..360
r, g, b = colorsys.hsv_to_rgb(h/360, 1., 1.)
r*=255
g*=255
b*=255
return r, g, b
def color_isomap_alpha(isomap):
colored = np.zeros([ISOMAP_SIZE, ISOMAP_SIZE, 3], dtype='uint8')
for x in range(colored.shape[0]):
for y in range(colored.shape[1]):
#print (isomap[x,y,3])
#print (pseudocolor(isomap[x,y,3],0, 255)*255)
colored[x,y,:] = pseudocolor(isomap[x,y,3],0, 255)
#print (colored[x,y,:])
return colored
def color_alpha_only(confidence, minval = None, maxval = None):
#print (confidence.shape)
if not maxval:
maxval = np.max(confidence)
if not minval:
minval = 0
colored = np.zeros([confidence.shape[0], confidence.shape[1], 3], dtype='uint8')
for x in range(colored.shape[0]):
for y in range(colored.shape[1]):
#print (isomap[x,y,3])
#print (pseudocolor(isomap[x,y,3],0, 255)*255)
colored[x,y,:] = pseudocolor(confidence[x,y,0],minval, maxval)
#print (colored[x,y,:])
return colored
if __name__ == "__main__":
#isomap_playground()
#show_isomap(cv2.imread(isomap_paths[0], cv2.IMREAD_UNCHANGED))
merge_isomaps_pg()
#cv2.imwrite('/user/HS204/m09113/Desktop/Boris/new_isomaps/image-00456.isomap_colored.png',color_isomap_alpha(cv2.imread('/user/HS204/m09113/Desktop/Boris/new_isomaps/image-00456.isomap.png', cv2.IMREAD_UNCHANGED)))
#cv2.imwrite('/user/HS204/m09113/my_project_folder/Boris/new_isomaps/image-00058_conf01_colored.png',color_alpha_only(np.load('/user/HS204/m09113/my_project_folder/Boris/new_isomaps/image-00058_cnn_conf01.npy')))
#cv2.waitKey()
#cv2.destroyAllWindows()