-
Notifications
You must be signed in to change notification settings - Fork 25
/
Copy pathtest
191 lines (142 loc) · 6.89 KB
/
test
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
#!/usr/bin/python3
import argparse
import sys
import os
import torchvision.transforms as transforms
from torchvision.utils import save_image
from torch.utils.data import DataLoader
from torch.autograd import Variable
import torch
from PIL import Image
import numpy as np
from models_guided import Generator_F2S, Generator_S2F
from utils import mask_generator
from utils import QueueMask
parser = argparse.ArgumentParser()
parser.add_argument('--batchSize', type=int, default=1, help='size of the batches')
parser.add_argument('--dataroot', type=str, default='datasets/horse2zebra/', help='root directory of the dataset')
parser.add_argument('--input_nc', type=int, default=3, help='number of channels of input data')
parser.add_argument('--output_nc', type=int, default=3, help='number of channels of output data')
parser.add_argument('--size', type=int, default=400, help='size of the data (squared assumed)')
parser.add_argument('--cuda', action='store_true', help='use GPU computation')
parser.add_argument('--n_cpu', type=int, default=8, help='number of cpu threads to use during batch generation')
parser.add_argument('--generator_A2B', type=str, default='output/netG_A2B.pth', help='A2B generator checkpoint file')
parser.add_argument('--generator_B2A', type=str, default='output/netG_B2A.pth', help='B2A generator checkpoint file')
opt = parser.parse_args()
### ISTD
# opt.dataroot_A = '/home/xwhu/dataset/ISTD/test/test_A'
# opt.dataroot_B = '/home/xwhu/dataset/ISTD/test/test_C'
#
# opt.im_suf_A = '.png'
# opt.im_suf_B = '.png'
### SRD
# opt.dataroot_A = '/home/xwhu/dataset/SRD/test_data/shadow'
# opt.dataroot_B = '/home/xwhu/dataset/SRD/test_data/shadow_free'
#
# opt.im_suf_A = '.jpg'
# opt.im_suf_B = '.jpg'
### USR
opt.dataroot_A = '/home/xwhu/dataset/shadow_USR/shadow_test'
opt.dataroot_B = '/home/xwhu/dataset/shadow_USR/shadow_free'
opt.im_suf_A = '.jpg'
opt.im_suf_B = '.jpg'
if torch.cuda.is_available():
opt.cuda = True
device = torch.device('cuda:0')
print(opt)
###### Definition of variables ######
# Networks
netG_A2B = Generator_S2F(opt.input_nc, opt.output_nc)
netG_B2A = Generator_F2S(opt.output_nc, opt.input_nc)
if opt.cuda:
netG_A2B.to(device)
netG_B2A.to(device)
# Load state dicts
netG_A2B.load_state_dict(torch.load(opt.generator_A2B))
netG_B2A.load_state_dict(torch.load(opt.generator_B2A))
# Set model's test mode
netG_A2B.eval()
netG_B2A.eval()
# Inputs & targets memory allocation
Tensor = torch.cuda.FloatTensor if opt.cuda else torch.Tensor
input_A = Tensor(opt.batchSize, opt.input_nc, opt.size, opt.size)
input_B = Tensor(opt.batchSize, opt.output_nc, opt.size, opt.size)
# Dataset loader
img_transform = transforms.Compose([
transforms.Resize((int(opt.size),int(opt.size)), Image.BICUBIC),
transforms.ToTensor(),
transforms.Normalize((0.5,0.5,0.5), (0.5,0.5,0.5))
])
#dataloader = DataLoader(ImageDataset(opt.dataroot, transforms_=transforms_, mode='test'),
# batch_size=opt.batchSize, shuffle=False, num_workers=opt.n_cpu)
###################################
to_pil = transforms.ToPILImage()
###### Testing######
# Create output dirs if they don't exist
if not os.path.exists('output/A'):
os.makedirs('output/A')
if not os.path.exists('output/B'):
os.makedirs('output/B')
if not os.path.exists('output/mask'):
os.makedirs('output/mask')
# if not os.path.exists('output/recovered_shadow'):
# os.makedirs('output/recovered_shadow')
# if not os.path.exists('output/same_A'):
# os.makedirs('output/same_A')
# if not os.path.exists('output/recovered_shadow_free'):
# os.makedirs('output/recovered_shadow_free')
# if not os.path.exists('output/same_B'):
# os.makedirs('output/same_B')
##################################### A to B // shadow to shadow-free
gt_list = [os.path.splitext(f)[0] for f in os.listdir(opt.dataroot_A) if f.endswith(opt.im_suf_A)]
mask_queue = QueueMask(gt_list.__len__())
mask_non_shadow = Variable(Tensor(1, 1, opt.size, opt.size).fill_(-1.0), requires_grad=False)
for idx, img_name in enumerate(gt_list):
print('predicting: %d / %d' % (idx + 1, len(gt_list)))
# Set model input
img = Image.open(os.path.join(opt.dataroot_A, img_name + opt.im_suf_A)).convert('RGB')
w, h = img.size
img_var = (img_transform(img).unsqueeze(0)).to(device)
# Generate output
temp_B = netG_A2B(img_var)
fake_B = 0.5*(temp_B.data + 1.0)
mask_queue.insert(mask_generator(img_var, temp_B))
fake_B = np.array(transforms.Resize((h, w))(to_pil(fake_B.data.squeeze(0).cpu())))
Image.fromarray(fake_B).save('output/B/%s' % img_name + opt.im_suf_A)
mask_last = mask_queue.last_item()
# recovered_A = netG_B2A(temp_B, mask_last)
# recovered_shadow_image = 0.5 * (recovered_A.data + 1.0)
# recovered_shadow_image = np.array(transforms.Resize((h, w))(to_pil(recovered_shadow_image.data.squeeze(0).cpu())))
# Image.fromarray(recovered_shadow_image).save('output/recovered_shadow/%s' % img_name + opt.im_suf_A)
# same_A = netG_B2A(img_var, mask_non_shadow)
# same_A_image = 0.5 * (same_A.data + 1.0)
# same_A_image = np.array(transforms.Resize((h, w))(to_pil(same_A_image.data.squeeze(0).cpu())))
# Image.fromarray(same_A_image).save('output/same_A/%s' % img_name + opt.im_suf_A)
print('Generated images %04d of %04d' % (idx+1, len(gt_list)))
##################################### B to A
gt_list = [os.path.splitext(f)[0] for f in os.listdir(opt.dataroot_B) if f.endswith(opt.im_suf_B)]
for idx, img_name in enumerate(gt_list):
print('predicting: %d / %d' % (idx + 1, len(gt_list)))
# Set model input
img = Image.open(os.path.join(opt.dataroot_B, img_name + opt.im_suf_B)).convert('RGB')
w, h = img.size
img_var = (img_transform(img).unsqueeze(0)).to(device)
# Generate output
#fake_B = 0.5 * (netG_A2B(img_var).data + 1.0)
mask = mask_queue.rand_item()
mask_cpu = np.array(transforms.Resize((h, w))(to_pil(((mask.data + 1 ) * 0.5).squeeze(0).cpu())))
temp_A = netG_B2A(img_var, mask)
fake_A = 0.5*(temp_A.data + 1.0)
fake_A = np.array(transforms.Resize((h, w))(to_pil(fake_A.data.squeeze(0).cpu())))
# Save image files
Image.fromarray(fake_A).save('output/A/%s' % img_name + opt.im_suf_B)
Image.fromarray(mask_cpu).save('output/mask/%s' % img_name + opt.im_suf_B)
# recovered_B = netG_A2B(temp_A)
# recovered_shadow_free_image = 0.5 * (recovered_B.data + 1.0)
# recovered_shadow_free_image = np.array(transforms.Resize((h, w))(to_pil(recovered_shadow_free_image.data.squeeze(0).cpu())))
# Image.fromarray(recovered_shadow_free_image).save('output/recovered_shadow_free/%s' % img_name + opt.im_suf_A)
# same_B = netG_A2B(img_var)
# same_B_image = 0.5 * (same_B.data + 1.0)
# same_B_image = np.array(transforms.Resize((h, w))(to_pil(same_B_image.data.squeeze(0).cpu())))
# Image.fromarray(same_B_image).save('output/same_B/%s' % img_name + opt.im_suf_A)
print('Generated images %04d of %04d' % (idx + 1, len(gt_list)))