forked from Joker316701882/Deep-Image-Matting
-
Notifications
You must be signed in to change notification settings - Fork 0
/
matting.py
209 lines (180 loc) · 9.01 KB
/
matting.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
import tensorflow as tf
import numpy as np
import random
from scipy import misc,ndimage
import copy
import itertools
import os
from sys import getrefcount
import gc
trimap_kernel = [val for val in range(20,40)]
g_mean = np.array(([126.88,120.24,112.19])).reshape([1,1,3])
hard_samples = [
1,4,8,11,13,15,16,19,28,42,43,44,46,65,68,69,70,81,91,92,94,101,104,
118,137,145,152,155,156,176,187,189,191,193,198,203,208,212,215,
216,221,233,239,243,254,264,265,267,272,278,279,281,288,290,291,292,
293,298,300,301,302,309,316,320,325,337,345,346,347,369,370,374,381,
386,402,416,432,443,450,451,454,456,457,459,464,487,490,499,502,513,
514,552,555,558,559,577,580,587,593,602,608,609,613,632,634,639,640,
649,663,688,710,717,718,723,729,736,740,741,745,757,769,775,778,785,
788,805,808,815,820,834,839,840,845,846,848,860,861,864,868,870,872,
877,885,889,892,894,895
]
def unpool(pool, ind, ksize=[1, 2, 2, 1], scope='unpool'):
with tf.variable_scope(scope):
input_shape = pool.get_shape().as_list()
output_shape = (input_shape[0], input_shape[1] * ksize[1], input_shape[2] * ksize[2], input_shape[3])
flat_input_size = np.prod(input_shape)
flat_output_shape = [output_shape[0], output_shape[1] * output_shape[2] * output_shape[3]]
pool_ = tf.reshape(pool, [flat_input_size])
batch_range = tf.reshape(tf.range(output_shape[0], dtype=ind.dtype), shape=[input_shape[0], 1, 1, 1])
b = tf.ones_like(ind) * batch_range
b = tf.reshape(b, [flat_input_size, 1])
ind_ = tf.reshape(ind, [flat_input_size, 1])
ind_ = tf.concat([b, ind_], 1)
ret = tf.scatter_nd(ind_, pool_, shape=flat_output_shape)
ret = tf.reshape(ret, output_shape)
return ret
def UR_center(trimap):
target = np.where(trimap==128)
index = random.choice([i for i in range(len(target[0]))])
return np.array(target)[:,index][:2]
def load_path(alpha,eps,BG,hard_mode = False):
folders = os.listdir(alpha)
common_paths = []
if hard_mode:
for folder in folders:
if int(folder) in hard_samples:
images = os.listdir(os.path.join(alpha,folder))
common_paths.extend([os.path.join(folder,image) for image in images])
else:
for folder in folders:
#if int(folder)==137:
images = os.listdir(os.path.join(alpha,folder))
common_paths.extend([os.path.join(folder,image) for image in images])
print(common_paths)
alphas_abspath = [os.path.join(alpha,common_path) for common_path in common_paths]
epses_abspath = [os.path.join(eps,common_path) for common_path in common_paths]
BGs_abspath = [os.path.join(BG,common_path)[:-3] + 'jpg' for common_path in common_paths]
return np.array(alphas_abspath),np.array(epses_abspath),np.array(BGs_abspath)
def load_data(batch_alpha_paths,batch_eps_paths,batch_BG_paths):
batch_size = batch_alpha_paths.shape[0]
train_batch = []
images_without_mean_reduction = []
for i in range(batch_size):
alpha = misc.imread(batch_alpha_paths[i],'L').astype(np.float32)
eps = misc.imread(batch_eps_paths[i]).astype(np.float32)
BG = misc.imread(batch_BG_paths[i]).astype(np.float32)
batch_i,raw_RGB = preprocessing_single(alpha, BG, eps,batch_alpha_paths[i])
train_batch.append(batch_i)
images_without_mean_reduction.append(raw_RGB)
train_batch = np.stack(train_batch)
return train_batch[:,:,:,:3],np.expand_dims(train_batch[:,:,:,3],3),np.expand_dims(train_batch[:,:,:,4],3),train_batch[:,:,:,5:8],train_batch[:,:,:,8:],images_without_mean_reduction
def generate_trimap(trimap,alpha):
k_size = random.choice(trimap_kernel)
trimap[np.where((ndimage.grey_dilation(alpha[:,:,0],size=(k_size,k_size)) - ndimage.grey_erosion(alpha[:,:,0],size=(k_size,k_size)))!=0)] = 128
#trimap[np.where((ndimage.grey_dilation(alpha[:,:,0],size=(k_size,k_size)) - alpha[:,:,0]!=0))] = 128
return trimap
def preprocessing_single(alpha, BG, eps,name,image_size=320):
alpha = np.expand_dims(alpha,2)
trimap = np.copy(alpha)
trimap = generate_trimap(trimap,alpha)
train_data = np.zeros([image_size,image_size,8])
crop_size = random.choice([320,480,620])
# crop_size = 320
flip = random.choice([0,1])
i_UR_center = UR_center(trimap)
#i_UR_center = [int(alpha.shape[0]/2),int(alpha.shape[1]/2)]
train_pre = np.concatenate([trimap,alpha,BG,eps],2)
if crop_size == 320:
h_start_index = i_UR_center[0] - 159
w_start_index = i_UR_center[1] - 159
tmp = train_pre[h_start_index:h_start_index+320, w_start_index:w_start_index+320, :]
if flip:
tmp = tmp[:,::-1,:]
tmp[:,:,1] = tmp[:,:,1] / 255.0
tmp[:,:,5:] = np.expand_dims(tmp[:,:,1],2) * tmp[:,:,5:] # here replace eps with FG
raw_RGB = np.expand_dims(tmp[:,:,1],2) * tmp[:,:,5:] + np.expand_dims((1. - tmp[:,:,1]),2) * tmp[:,:,2:5]
reduced_RGB = raw_RGB - g_mean
tmp = np.concatenate([reduced_RGB,tmp],2)
train_data = tmp
if crop_size == 480:
h_start_index = i_UR_center[0] - 239
w_start_index = i_UR_center[1] - 239
tmp = train_pre[h_start_index:h_start_index+480, w_start_index:w_start_index+480, :]
if flip:
tmp = tmp[:,::-1,:]
tmp1 = np.zeros([image_size,image_size,8]).astype(np.float32)
tmp1[:,:,0] = misc.imresize(tmp[:,:,0].astype(np.uint8),[image_size,image_size],interp = 'nearest',mode='L').astype(np.float32)
tmp1[:,:,1] = misc.imresize(tmp[:,:,1].astype(np.uint8),[image_size,image_size]).astype(np.float32) / 255.0
tmp1[:,:,2:5] = misc.imresize(tmp[:,:,2:5].astype(np.uint8),[image_size,image_size,3]).astype(np.float32)
tmp1[:,:,5:] = misc.imresize(tmp[:,:,5:].astype(np.uint8),[image_size,image_size,3]).astype(np.float32)
tmp1[:,:,5:] = np.expand_dims(tmp1[:,:,1],2) * tmp1[:,:,5:] # here replace eps with FG
raw_RGB = np.expand_dims(tmp1[:,:,1],2) * tmp1[:,:,5:] + np.expand_dims((1. - tmp1[:,:,1]),2) * tmp1[:,:,2:5]
reduced_RGB = raw_RGB - g_mean
tmp1 = np.concatenate([reduced_RGB,tmp1],2)
train_data = tmp1
if crop_size == 620:
h_start_index = i_UR_center[0] - 309
#boundary security
if h_start_index<0:
h_start_index = 0
w_start_index = i_UR_center[1] - 309
if w_start_index<0:
w_start_index = 0
tmp = train_pre[h_start_index:h_start_index+620, w_start_index:w_start_index+620, :]
if flip:
tmp = tmp[:,::-1,:]
tmp1 = np.zeros([image_size,image_size,8]).astype(np.float32)
tmp1[:,:,0] = misc.imresize(tmp[:,:,0].astype(np.uint8),[image_size,image_size],interp = 'nearest',mode='L').astype(np.float32)
tmp1[:,:,1] = misc.imresize(tmp[:,:,1].astype(np.uint8),[image_size,image_size]).astype(np.float32) / 255.0
tmp1[:,:,2:5] = misc.imresize(tmp[:,:,2:5].astype(np.uint8),[image_size,image_size,3]).astype(np.float32)
tmp1[:,:,5:] = misc.imresize(tmp[:,:,5:].astype(np.uint8),[image_size,image_size,3]).astype(np.float32)
tmp1[:,:,5:] = np.expand_dims(tmp1[:,:,1],2) * tmp1[:,:,5:] # here replace eps with FG
raw_RGB = np.expand_dims(tmp1[:,:,1],2) * tmp1[:,:,5:] + np.expand_dims((1. - tmp1[:,:,1]),2) * tmp1[:,:,2:5]
reduced_RGB = raw_RGB - g_mean
tmp1 = np.concatenate([reduced_RGB,tmp1],2)
train_data = tmp1
train_data = train_data.astype(np.float32)
# misc.imsave('./train_alpha.png',train_data[:,:,4])
return train_data,raw_RGB
def load_alphamatting_data(test_alpha):
rgb_path = os.path.join(test_alpha,'rgb')
trimap_path = os.path.join(test_alpha,'trimap')
alpha_path = os.path.join(test_alpha,'alpha')
images = os.listdir(trimap_path)
test_num = len(images)
all_shape = []
rgb_batch = []
tri_batch = []
alp_batch = []
for i in range(test_num):
rgb = misc.imread(os.path.join(rgb_path,images[i]))
trimap = misc.imread(os.path.join(trimap_path,images[i]),'L')
alpha = misc.imread(os.path.join(alpha_path,images[i]),'L')/255.0
all_shape.append(trimap.shape)
rgb_batch.append(misc.imresize(rgb,[320,320,3])-g_mean)
trimap = misc.imresize(trimap,[320,320],interp = 'nearest').astype(np.float32)
tri_batch.append(np.expand_dims(trimap,2))
alp_batch.append(alpha)
return np.array(rgb_batch),np.array(tri_batch),np.array(alp_batch),all_shape,images
def load_validation_data(vali_root):
alpha_dir = os.path.join(vali_root,'alpha')
RGB_dir = os.path.join(vali_root,'RGB')
images = os.listdir(alpha_dir)
test_num = len(images)
all_shape = []
rgb_batch = []
tri_batch = []
alp_batch = []
for i in range(test_num):
rgb = misc.imread(os.path.join(RGB_dir,images[i]))
alpha = misc.imread(os.path.join(alpha_dir,images[i]),'L')
trimap = generate_trimap(np.expand_dims(np.copy(alpha),2),np.expand_dims(alpha,2))[:,:,0]
alpha = alpha / 255.0
all_shape.append(trimap.shape)
rgb_batch.append(misc.imresize(rgb,[320,320,3])-g_mean)
trimap = misc.imresize(trimap,[320,320],interp = 'nearest').astype(np.float32)
tri_batch.append(np.expand_dims(trimap,2))
alp_batch.append(alpha)
return np.array(rgb_batch),np.array(tri_batch),np.array(alp_batch),all_shape,images