-
Notifications
You must be signed in to change notification settings - Fork 14
/
Copy pathIQAmodel.py
115 lines (105 loc) · 4.34 KB
/
IQAmodel.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
import torch
import torch.nn as nn
import torch.nn.functional as F
from torchvision import models
import numpy as np
def SPSP(x, P=1, method='avg'):
batch_size = x.size(0)
map_size = x.size()[-2:]
pool_features = []
for p in range(1, P+1):
pool_size = [np.int(d / p) for d in map_size]
if method == 'maxmin':
M = F.max_pool2d(x, pool_size)
m = -F.max_pool2d(-x, pool_size)
pool_features.append(torch.cat((M, m), 1).view(batch_size, -1)) # max & min pooling
elif method == 'max':
M = F.max_pool2d(x, pool_size)
pool_features.append(M.view(batch_size, -1)) # max pooling
elif method == 'min':
m = -F.max_pool2d(-x, pool_size)
pool_features.append(m.view(batch_size, -1)) # min pooling
elif method == 'avg':
a = F.avg_pool2d(x, pool_size)
pool_features.append(a.view(batch_size, -1)) # average pooling
else:
m1 = F.avg_pool2d(x, pool_size)
rm2 = torch.sqrt(F.relu(F.avg_pool2d(torch.pow(x, 2), pool_size) - torch.pow(m1, 2)))
if method == 'std':
pool_features.append(rm2.view(batch_size, -1)) # std pooling
else:
pool_features.append(torch.cat((m1, rm2), 1).view(batch_size, -1)) # statistical pooling: mean & std
return torch.cat(pool_features, dim=1)
class IQAModel(nn.Module):
def __init__(self, arch='resnext101_32x8d', pool='avg', use_bn_end=False, P6=1, P7=1):
super(IQAModel, self).__init__()
self.pool = pool
self.use_bn_end = use_bn_end
if pool in ['max', 'min', 'avg', 'std']:
c = 1
else:
c = 2
self.P6 = P6 #
self.P7 = P7 #
features = list(models.__dict__[arch](pretrained=True).children())[:-2]
if arch == 'alexnet':
in_features = [256, 256]
self.id1 = 9
self.id2 = 12
features = features[0]
elif arch == 'vgg16':
in_features = [512, 512]
self.id1 = 23
self.id2 = 30
features = features[0]
elif 'res' in arch:
self.id1 = 6
self.id2 = 7
if arch == 'resnet18' or arch == 'resnet34':
in_features = [256, 512]
else:
in_features = [1024, 2048]
else:
print('The arch is not implemented!')
self.features = nn.Sequential(*features)
self.dr6 = nn.Sequential(nn.Linear(in_features[0] * c * sum([p * p for p in range(1, self.P6+1)]), 1024),
nn.BatchNorm1d(1024),
nn.Linear(1024, 256),
nn.BatchNorm1d(256),
nn.Linear(256, 64),
nn.BatchNorm1d(64), nn.ReLU())
self.dr7 = nn.Sequential(nn.Linear(in_features[1] * c * sum([p * p for p in range(1, self.P7+1)]), 1024),
nn.BatchNorm1d(1024),
nn.Linear(1024, 256),
nn.BatchNorm1d(256),
nn.Linear(256, 64),
nn.BatchNorm1d(64), nn.ReLU())
if self.use_bn_end:
self.regr6 = nn.Sequential(nn.Linear(64, 1), nn.BatchNorm1d(1))
self.regr7 = nn.Sequential(nn.Linear(64, 1), nn.BatchNorm1d(1))
self.regression = nn.Sequential(nn.Linear(64 * 2, 1), nn.BatchNorm1d(1))
else:
self.regr6 = nn.Linear(64, 1)
self.regr7 = nn.Linear(64, 1)
self.regression = nn.Linear(64 * 2, 1)
def extract_features(self, x):
f, pq = [], []
for ii, model in enumerate(self.features):
x = model(x)
if ii == self.id1:
x6 = SPSP(x, P=self.P6, method=self.pool)
x6 = self.dr6(x6)
f.append(x6)
pq.append(self.regr6(x6))
if ii == self.id2:
x7 = SPSP(x, P=self.P7, method=self.pool)
x7 = self.dr7(x7)
f.append(x7)
pq.append(self.regr7(x7))
f = torch.cat(f, dim=1)
return f, pq
def forward(self, x):
f, pq = self.extract_features(x)
s = self.regression(f)
pq.append(s)
return pq