-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathstep1-pre.py
196 lines (161 loc) · 7.71 KB
/
step1-pre.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
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
192
193
194
195
196
# -*- coding: utf-8 -*-
"""
Created on Tong Xu 21 22:31:24 2020
@author: Administrator
"""
import os
import glob
import h5py
import random
import matplotlib.pyplot as plt
import imageio
from PIL import Image # for loading images as YCbCr format
import scipy.misc
import scipy.ndimage
import numpy as np
import scipy.misc as smi
import tensorflow as tf
def rgb2ycbcr(img, only_y=True): #自己重新写的rgb2ycbcr函数以求对应到matlab的rgb2ycbcr函数
'''same as matlab rgb2ycbcr
only_y: only return Y channel
Input:
uint8, [0, 255]
float, [0, 1]
'''
in_img_type = img.dtype
img.astype(np.float32)
if in_img_type != np.uint8:
img *= 255.
# convert
if only_y:
rlt = np.dot(img, [65.481, 128.553, 24.966]) / 255.0 + 16.0
else:
rlt = np.matmul(img, [[65.481, -37.797, 112.0], [128.553, -74.203, -93.786],
[24.966, 112.0, -18.214]]) / 255.0 + [16, 128, 128]
if in_img_type == np.uint8:
rlt = rlt.round()
else:
rlt /= 255.
return rlt.astype(in_img_type)
is_train=False #预处理训练数据时时为True,预处理测试数据时改为False
scale=3; #插值规模
#定义一个保存图片的函数
def imsave(image, path):
return imageio.imwrite(path, image)
#将数据读入进来
if is_train:
dataset="Train"
filenames = os.listdir(dataset)
data_dir = os.path.join(os.getcwd(), dataset)
data = glob.glob(os.path.join(data_dir, "*.bmp"))
else:
dataset="Test"
data_dir = os.path.join(os.sep, (os.path.join(os.getcwd(), dataset)), "Set5")
data = glob.glob(os.path.join(data_dir, "*.bmp"))
if is_train:
image_size=33 #训练输入的图片的大小33*33
stride=14 #
label_size=21 #训练输入图片经过训练变成21*21大小
padding = abs(image_size - label_size) / 2 # 6像素点的边缘
sub_input_sequence = []
sub_label_sequence = []
for i in range(len(data)): #range产生随机数且优于range函数
image=imageio.imread(data[i])
image=rgb2ycbcr(image) #自己写rgb2ycbcr函数可取
#image=scipy.misc.imread(data[i], flatten=True, mode='YCbCr').astype(np.float) #提取Y通道
if len(image.shape) == 3:
h, w, _ = image.shape
h = h - np.mod(h, scale)
w = w - np.mod(w, scale)
label_ = image[0:h, 0:w, :] #长宽进行裁剪，第三维原样，但这里并没有变成33*33
else:
h, w = image.shape
h = h - np.mod(h, scale)
w = w - np.mod(w, scale)
label_ = image[0:h, 0:w]
image = image / 255.
label_ = label_ / 255.
#进行两次插值构造低分辨率图片
label_1=Image.fromarray(label_)
input_= label_1.resize(( w // scale,h // scale),Image.BICUBIC)
input_= input_.resize((w,h), Image.BICUBIC)
input_=np.float64(input_)
#保存四位小数
label_=np.around(label_, decimals=4)
input_=np.around(input_,decimals=4)
#下面这个插值函数与matlab中不一致，因此舍弃
#input_ = scipy.ndimage.interpolation.zoom(label_, (1./scale), prefilter=False)#进行双三次插值变为低分辨率图片
#input_ = scipy.ndimage.interpolation.zoom(input_, (scale/1.), prefilter=False)#再次进行双三次插值变为与高分辨率图片一样大小
if len(input_.shape) == 3: #取train图片的长宽，因为这里是在is_train=True下的if语句
h, w, _ = input_.shape
else:
h, w = input_.shape
for x in range(0, h-image_size+1, stride): #以stride为步长进行取子图片操作
for y in range(0, w-image_size+1, stride):
sub_input = input_[x:x+image_size, y:y+image_size] # [33 x 33]
sub_label = label_[x+int(padding):x+int(padding)+label_size, y+int(padding):y+int(padding)+label_size] # [21 x 21]
# Make channel value
sub_input = sub_input.reshape([image_size, image_size, 1])
sub_label = sub_label.reshape([label_size, label_size, 1])
sub_input_sequence.append(sub_input) #append为在列表末尾添加新的对象
sub_label_sequence.append(sub_label)
if not is_train:
image_size=33
stride=14
label_size=21
padding = abs(image_size - label_size) / 2 # 6
sub_input_sequence = []
sub_label_sequence = []
for i in range(len(data)): #range产生随机数且优于range函数
image=imageio.imread(data[i])
image=rgb2ycbcr(image) #自己写rgb2ycbcr函数可取
#image=scipy.misc.imread(data[i], flatten=True, mode='YCbCr').astype(np.float)#读取方法不可取
if len(image.shape) == 3:
h, w, _ = image.shape
h = h - np.mod(h, scale)
w = w - np.mod(w, scale)
label_ = image[0:h, 0:w, :] #长宽进行裁剪，第三维原样，但这里并没有变成33*33
else:
h, w = image.shape
h = h - np.mod(h, scale)
w = w - np.mod(w, scale)
label_ = image[0:h, 0:w]
image = image / 255.
label_ = label_ / 255. #此时label_为真图，后续进行psnr计算时与预测图片进行对比
#方法一最好，基本达到与matlab相同水平
label_1=Image.fromarray(label_)
input_= label_1.resize((w//scale,h//scale),Image.BICUBIC)
input_= input_.resize((w,h), Image.BICUBIC)
input_=np.float64(input_)
#方法二优于方法三
#input_ =smi.imresize(label_, (85,85),interp='bicubic')
#input_ =smi.imresize(input_, (255,255),interp='bicubic')/255.
#方法三不太行
#input_ = scipy.ndimage.interpolation.zoom(label_, (1./scale), prefilter=False)
#input_ = scipy.ndimage.interpolation.zoom(input_, (scale/1.), prefilter=False)
if len(input_.shape) == 3: #取输入图片的长宽
h, w, _ = input_.shape
else:
h, w = input_.shape
#保留四位小数
label_=np.around(label_, decimals=4)
input_=np.around(input_,decimals=4)
image_path = os.path.join(os.getcwd(),'sample')
image_path = os.path.join(image_path, "label_image.png")
imsave(label_, image_path) #保存真图
image_path = os.path.join(os.getcwd(),'sample')
image_path = os.path.join(image_path, "input_image.png")
imsave(input_, image_path) #保存输入图片
sub_input = input_.reshape([h, w, 1])
sub_label = label_.reshape([h, w, 1])
sub_input_sequence.append(sub_input) #append为在列表末尾添加新的对象
sub_label_sequence.append(sub_label)
arrdata = np.asarray(sub_input_sequence) # [?, 33, 33, 1]
arrlabel = np.asarray(sub_label_sequence) # [?, 21, 21, 1]
if is_train:
savepath = os.path.join(os.getcwd(), 'h5/train.h5') #os.getcwd()为获取当前工作目录
else:
savepath = os.path.join(os.getcwd(), 'h5/test.h5')
with h5py.File(savepath, 'w') as hf: #数据集的制作,图片大小不一样，不能转成h5，这里无效，可以在test时直接读取图片
hf.create_dataset('data', data=arrdata)
hf.create_dataset('label', data=arrlabel)