/
visu_filters_deconv.py
162 lines (111 loc) · 3.26 KB
/
visu_filters_deconv.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
# -*- coding: utf-8 -*-
"""
Created on Tue Jul 26 15:27:25 2016
@author: rflamary
"""
import sys
import numpy as np
import scipy as sp
import scipy.signal
import scipy.io as spio
import deconv
import matplotlib.pylab as pl
import dsutils
import theano
def get_fname(method,n,npsf,sigma,img):
return 'res/{}_{}x{}_PSF{}_sigma{:1.3f}_{}.mat'.format(method,n,n,npsf,sigma,img)
def get_fname_all(method,n,npsf,sigma):
return 'res/{}_{}x{}_PSF{}_sigma{:1.3f}_all.mat'.format(method,n,n,npsf,sigma)
#%% load image
I0=deconv.load_fits_image('M51a')
I0=I0/I0.max()
#%% generat
i=2
cr=32
lst_img=['M31','Hoag','M51a','M81','M101','M104']
#lst_img=['M31']
nb_img=len(lst_img)
def sel(I):
return I[300:-cr-100,300:-cr-100]
img_txt=lst_img[i]
method='none'
I0=deconv.load_fits_image(img_txt)
n=1024
iv=1200;jv=1200
I0=I0[iv:iv+n,jv:jv+n]
npsf=64
sigma=0.01
fname=get_fname('none',n,npsf,sigma,img_txt)
data=spio.loadmat(fname)
Inoise=data['Irec']
#%% get PSF
npsf=64
nr=5
#%% deconvnn
fname='models/32x32_10x10-64_6x6-16_5x5-1_PSF64_sigma0.010_M51a'
model=dsutils.load_model(fname)
model.compile(optimizer='SGD', loss='mse')
sz=32
szp=14
deconv.tic()
I_dcnnn=dsutils.apply_model(Inoise,model,sz,szp)
deconv.toc()
#%% visu last layer
szp2=18
Ip=dsutils.im2patch(Inoise,sz,szp2)
convout1_f = theano.function([model.get_input_at(0)], model.layers[1].get_output_at(0),allow_input_downcast=True)
Ip2=convout1_f(Ip)
I_layer1=dsutils.patch2im(Ip2[:,0:1,:,:],Inoise.shape,sz,szp2)
#
#pl.figure("last layer")
#
#for i in range(16):
# pl.subplot(4,4,i+1)
# pl.imshow(dsutils.patch2im(Ip2[:,i:i+1,:,:],Inoise.shape,sz,szp2),interpolation='nearest')
Il2_1=dsutils.patch2im(Ip2[:,15:16,:,:],Inoise.shape,sz,szp2)
Il2_2=dsutils.patch2im(Ip2[:,2:3,:,:],Inoise.shape,sz,szp2)
Il2_3=dsutils.patch2im(Ip2[:,7:8,:,:],Inoise.shape,sz,szp2)
#%% visu first layer
szp2=23
Ip=dsutils.im2patch(Inoise,sz,szp2)
convout0_f = theano.function([model.get_input_at(0)], model.layers[0].get_output_at(0),allow_input_downcast=True)
Ip2=convout0_f(Ip)
I_layer1=dsutils.patch2im(Ip2[:,0:1,:,:],Inoise.shape,sz,szp2)
#pl.figure("first layer")
#
#for i in range(64):
# pl.subplot(8,8,i+1)
# pl.imshow(dsutils.patch2im(Ip2[:,i:i+1,:,:],Inoise.shape,sz,szp2),interpolation='nearest')
Il1_1=dsutils.patch2im(Ip2[:,4:5,:,:],Inoise.shape,sz,szp2)
Il1_2=dsutils.patch2im(Ip2[:,2:3,:,:],Inoise.shape,sz,szp2)
Il1_3=dsutils.patch2im(Ip2[:,33:34,:,:],Inoise.shape,sz,szp2)
#%%
yt=1
fs=10
pl.figure(1)
pl.subplot(3,3,1)
pl.imshow(sel(Il1_1),cmap='gray')
pl.title('Layer 1 output 1',fontsize=fs,y=yt)
pl.axis("off")
pl.subplot(3,3,2)
pl.imshow(sel(Il1_2),cmap='gray')
pl.title('Layer 1 output 2',fontsize=fs,y=yt)
pl.axis("off")
pl.subplot(3,3,3)
pl.imshow(sel(Il1_3),cmap='gray')
pl.title('Layer 1 output 3',fontsize=fs,y=yt)
pl.axis("off")
pl.subplot(3,3,4)
pl.imshow(sel(Il2_1),cmap='gray')
pl.title('Layer 2 output 1',fontsize=fs,y=yt)
pl.axis("off")
pl.subplot(3,3,5)
pl.imshow(sel(Il2_2),cmap='gray')
pl.title('Layer 2 output 2',fontsize=fs,y=yt)
pl.axis("off")
pl.subplot(3,3,6)
pl.imshow(sel(Il2_3),cmap='gray')
pl.title('Layer 2 output 3',fontsize=fs,y=yt)
pl.axis("off")
pl.subplots_adjust(wspace=-.5,hspace=0.3)
pl.savefig('imgs/images_features.png',dpi=300,bbox_inches='tight',pad_inches=.01)