-
Notifications
You must be signed in to change notification settings - Fork 99
/
grad_loss.py
158 lines (134 loc) · 4.87 KB
/
grad_loss.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
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.autograd import Variable
import numpy as np
np.set_printoptions(threshold=np.nan)
from math import exp
import cv2
# import matplotlib.pyplot as plt
def sobel(window_size):
assert(window_size%2!=0)
ind=window_size/2
matx=[]
maty=[]
for j in range(-ind,ind+1):
row=[]
for i in range(-ind,ind+1):
if (i*i+j*j)==0:
gx_ij=0
else:
gx_ij=i/float(i*i+j*j)
row.append(gx_ij)
matx.append(row)
for j in range(-ind,ind+1):
row=[]
for i in range(-ind,ind+1):
if (i*i+j*j)==0:
gy_ij=0
else:
gy_ij=j/float(i*i+j*j)
row.append(gy_ij)
maty.append(row)
# matx=[[-3, 0,+3],
# [-10, 0 ,+10],
# [-3, 0,+3]]
# maty=[[-3, -10,-3],
# [0, 0 ,0],
# [3, 10,3]]
if window_size==3:
mult=2
elif window_size==5:
mult=20
elif window_size==7:
mult=780
matx=np.array(matx)*mult
maty=np.array(maty)*mult
return torch.Tensor(matx), torch.Tensor(maty)
def create_window(window_size, channel):
windowx,windowy = sobel(window_size)
windowx,windowy= windowx.unsqueeze(0).unsqueeze(0), windowy.unsqueeze(0).unsqueeze(0)
windowx = torch.Tensor(windowx.expand(channel,1,window_size,window_size))
windowy = torch.Tensor(windowy.expand(channel,1,window_size,window_size))
# print windowx
#print windowy
return windowx,windowy
def gradient(img, windowx, windowy, window_size, padding, channel):
if channel > 1 : # do convolutions on each channel separately and then concatenate
gradx=torch.ones(img.shape)
grady=torch.ones(img.shape)
if img.is_cuda:
gradx=gradx.cuda(img.get_device())
grady=grady.cuda(img.get_device())
# print(gradx[:,0,:,:].shape)
for i in range(channel):
gradx[:,i,:,:]=F.conv2d(img[:,i,:,:].unsqueeze(1), windowx, padding=padding,groups=1).squeeze(1) #fix the padding according to the kernel size
grady[:,i,:,:]=F.conv2d(img[:,i,:,:].unsqueeze(1), windowy, padding=padding,groups=1).squeeze(1)
else:
gradx = F.conv2d(img, windowx, padding=padding,groups=1)
grady = F.conv2d(img, windowy, padding=padding,groups=1)
return gradx, grady
class Gradloss(torch.nn.Module):
def __init__(self, window_size = 3, padding= 1):
super(Gradloss, self).__init__()
self.window_size = window_size
self.padding= padding
self.channel = 1 # out channel
self.windowx,self.windowy = create_window(window_size, self.channel)
def forward(self, pred,label):
(batch_size, channel, _, _) = pred.size()
if pred.is_cuda:
self.windowx = self.windowx.cuda(pred.get_device())
self.windowx = self.windowx.type_as(pred)
self.windowy = self.windowy.cuda(pred.get_device())
self.windowy = self.windowy.type_as(pred)
pred_gradx,pred_grad_y=gradient(pred,self.windowx,self.windowy,self.window_size, self.padding,channel)
label_gradx,label_grad_y=gradient(label,self.windowx,self.windowy,self.window_size,self.padding, channel)
#label_grad=torch.sqrt((label_gradx*label_gradx) + (label_grad_y*label_grad_y))
#w=((label_grad[:,0,:,:]>=1)&(label_grad[:,1,:,:]>=1)&(label_grad[:,2,:,:]>=1)).float()*0.7
#msk=((label[:,0,:,:]!=0)&(label[:,1,:,:]!=0)&(label[:,2,:,:]!=1)).float()
#w+=msk*0.2
#w+=(1-msk)*0.1
# w=w.expand_as(pred)
l1_loss=nn.L1Loss()
#l2_loss=nn.MSELoss()
grad_loss=l1_loss(pred_gradx,label_gradx)+l1_loss(pred_grad_y,label_grad_y)
# w_grad_loss=(label-pred)**2
# w_grad_loss=w*w_grad_loss
# w_grad_loss=torch.mean(w_grad_loss)
return grad_loss#, w_grad_loss
# # For testing
# if __name__ == '__main__':
# img1_path="1_1_2-cp_Page_0654-XKI0001.exr"
# img2_path="1_1_1-tc_Page_065-YGB0001.exr"
# img1=cv2.imread(img1_path, cv2.IMREAD_ANYCOLOR | cv2.IMREAD_ANYDEPTH)
# # print(img1.shape)
# img2=cv2.imread(img2_path, cv2.IMREAD_ANYCOLOR | cv2.IMREAD_ANYDEPTH)
# # OpenCV sobel gradient for to check correctness
# sobelx1 = cv2.Sobel(img1,cv2.CV_64F,1,0,ksize=5)
# sobely1 = cv2.Sobel(img1,cv2.CV_64F,0,1,ksize=5)
# sobelx2 = cv2.Sobel(img2,cv2.CV_64F,1,0,ksize=5)
# sobely2 = cv2.Sobel(img2,cv2.CV_64F,0,1,ksize=5)
# img1=np.array(img1,dtype=np.float).transpose(2,0,1)
# img2=np.array(img2,dtype=np.float).transpose(2,0,1)
# img1=torch.from_numpy(img1).float().unsqueeze(0)
# img2=torch.from_numpy(img2).float().unsqueeze(0)
# gradloss=Gradloss(window_size=5)
# same_gloss,label_gradx,label_grady=gradloss(img1,img1)
# gradx1=np.array(label_gradx[0]).transpose(1,2,0)
# grady1=np.array(label_grady[0]).transpose(1,2,0)
# diff_gloss,label_gradx,label_grady=gradloss(img1,img2)
# gradx2=np.array(label_gradx[0]).transpose(1,2,0)
# grady2=np.array(label_grady[0]).transpose(1,2,0)
# f, axarr = plt.subplots(2, 4)
# axarr[0][0].imshow(sobelx1)
# axarr[0][1].imshow(sobely1)
# axarr[0][2].imshow(sobelx2)
# axarr[0][3].imshow(sobely2)
# axarr[1][0].imshow(gradx1)
# axarr[1][1].imshow(grady1)
# axarr[1][2].imshow(gradx2)
# axarr[1][3].imshow(grady2)
# plt.show()
# print(same_gloss.item())
# print(diff_gloss.item())