-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathtest_imageMF.py
163 lines (130 loc) · 5.85 KB
/
test_imageMF.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
# test phase
import torch
import os
from torch.autograd import Variable
from net import GhostFusion_net
import utils
from utils import gradient
from scipy.misc import imread, imsave, imresize
from args_fusion import args
import numpy as np
import time
import cv2
def load_model(path, input_nc, output_nc):
nest_model = GhostFusion_net(input_nc, output_nc)
nest_model.load_state_dict(torch.load(path))
para = sum([np.prod(list(p.size())) for p in nest_model.parameters()])
type_size = 4
print('Model {} : params: {:4f}M'.format(nest_model._get_name(), para * type_size / 1000 / 1000))
nest_model.eval()
#nest_model.cuda()
return nest_model
def _generate_fusion_image(model, strategy_type, img1, img2):
# encoder
en_r = model.encoder(img1)
en_v = model.encoder(img2)
f = model.fusion(en_r, en_v, strategy_type=strategy_type)
img_fusion = model.decoder(f);
return img_fusion[0]
#input order shold follow the order of "irBase,viBase,irDetail,viDetail" far, near, far, near
def _generate_fusion_mf(model,imgBase1,imgBase2,imgDetail1,imgDetail2):
#imsave('./gradient_perception/imgDetail1.png',(imgDetail1[0,0,:,:]*15).numpy());
#imsave('./gradient_perception/imgDetail2.png',(imgDetail2[0,0,:,:]*15).numpy());
imgGradDetail1 = gradient(imgDetail1);
imgGradDetail2 = gradient(imgDetail2);
#print(imgGradDetail1);
#imsave('./gradient_perception/imgGradDetail1.png',(imgGradDetail1[0,0,:,:]*100).numpy());
#imsave('./gradient_perception/imgGradDetail2.png',(imgGradDetail2[0,0,:,:]*100).numpy());
shape = imgBase1.shape;
imgGradDetail1 = torch.abs(imgGradDetail1);
imgGradDetail2 = torch.abs(imgGradDetail2);
en_GradDetail1 = model.encoder(imgGradDetail1);
en_GradDetail2 = model.encoder(imgGradDetail2);
focusMap1 = model.fusion(en_GradDetail1,en_GradDetail2,strategy_type='AGL1')[0];
ones = torch.ones(1,1,shape[2],shape[3]);
if (args.cuda):
ones = ones.cuda(args.device);
focusMap2 = ones-focusMap1;
fBase = imgBase1*focusMap1+imgBase2*focusMap2;
fDetail = imgDetail1*focusMap1+imgDetail2*focusMap2;
return fBase,fDetail
def run_demo(model, irBase_path,irDetail_path, visBase_path, visDetail_path, output_path_root, index, BS,DS, mode):
irBase_img = utils.get_test_images(irBase_path, height=None, width=None, mode=mode)
irDetail_img = utils.get_test_images(irDetail_path, height=None, width=None, mode=mode)
visBase_img = utils.get_test_images(visBase_path, height=None, width=None, mode=mode)
visDetail_img = utils.get_test_images(visDetail_path, height=None, width=None, mode=mode)
# dim = img_ir.shape
if args.cuda:
irBase_img = irBase_img.cuda(args.device)
irDetail_img = irDetail_img.cuda(args.device)
visBase_img = visBase_img.cuda(args.device)
visDetail_img = visDetail_img.cuda(args.device)
model = model.cuda(args.device);
irBase_img = Variable(irBase_img, requires_grad=False)
irDetail_img = Variable(irDetail_img, requires_grad=False)
visBase_img = Variable(visBase_img, requires_grad=False)
visDetail_img = Variable(visDetail_img, requires_grad=False)
#MF task
if (BS=='AGL1' and DS=='AGL1'):
fusedBase,fusedDetail = _generate_fusion_mf(model,irBase_img,visBase_img,irDetail_img,visDetail_img);
else:
#strategy_type_list = strategy_type_list = ['AVG', 'L1','SC','MAX','AGL1']
#Base L1
fusedBase = _generate_fusion_image(model, BS, irBase_img, visBase_img)
fusedDetail = _generate_fusion_image(model, DS, irDetail_img, visDetail_img)
fusedBase = fusedBase[0].cpu();
fusedBase = fusedBase.squeeze().squeeze();
fusedBase = fusedBase.numpy();
fusedBase = fusedBase*255;
file_name = 'fuseBase'+str(index) + '.png'
output_path = output_path_root + file_name
imsave(output_path,fusedBase);
#Detail max
fusedDetail = fusedDetail[0].cpu();
fusedDetail = fusedDetail.squeeze().squeeze();
fusedDetail = fusedDetail.numpy();
fusedDetail = fusedDetail*255;
fusedDetail = fusedDetail - np.mean(fusedBase);
file_name = 'fuseDetail'+str(index) + '.png'
output_path = output_path_root + file_name
imsave(output_path,fusedDetail);
#finalFuseResult
fusedFinalResult = fusedDetail + fusedBase;
############################ multi outputs ##############################################
file_name = 'fuseMF'+str(index) + '.png'
output_path = output_path_root + file_name
imsave(output_path,fusedFinalResult);
print(output_path)
def main():
test_path = "sample_input/"
fusion_type = 'auto' # auto, fusion_layer, fusion_all
strategy_type_list = ['AVG', 'L1','SC','MAX','AGL1'] # addition, attention_weight, attention_enhance, adain_fusion, channel_fusion, saliency_mask
BS = strategy_type_list[4];
DS = strategy_type_list[4];
output_path = './outputs/';
if os.path.exists(output_path) is False:
os.mkdir(output_path)
# in_c = 3 for RGB images; in_c = 1 for gray images
in_c = 1
if in_c == 1:
out_c = in_c
mode = 'L'
model_path = args.model_path_gray
else:
out_c = in_c
mode = 'RGB'
model_path = args.model_path_rgb
with torch.no_grad():
print('SSIM weight ----- ' + args.ssim_path[2])
ssim_weight_str = args.ssim_path[2]
model = load_model(model_path, in_c, out_c)
for i in range(1):
index = i + 1
irBase_path = test_path + 'MF_IRBase.png'
irDetail_path = test_path + 'MF_IRDetail.png'
visBase_path = test_path + 'MF_VISBase.png'
visDetail_path = test_path + 'MF_VISDetail.png'
run_demo(model,irBase_path,irDetail_path, visBase_path, visDetail_path, output_path, index, BS, DS, mode)
print('Done......')
if __name__ == '__main__':
main()