-
Notifications
You must be signed in to change notification settings - Fork 0
/
fast_msnn1.py
502 lines (421 loc) · 26.8 KB
/
fast_msnn1.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
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
#tensorflow 模块
import tensorflow as tf
#skimage模块下的io transform(图像的形变与缩放)模块
from skimage import io,transform
#glob 文件通配符模块
import glob
#os 处理文件和目录的模块
import os
#多维数据处理模块
import numpy as np
#
import time
#本地数据集地址
path1='F:/目标数据集/train/'
#path1="L:/增强眼底图像/train/"
#本地模型保存地址
#将所有的图片resize成100*100
w=128
h=128
c=3
def read_img(path):
#os.listdir(path) 返回path指定的文件夹包含的文件或文件夹的名字的列表
#os.path.isdir(path)判断path是否是目录
#b = [x+x for x in list1 if x+x<15 ] 列表生成式,循环list1,当if为真时,将x+x加入列表b
cate=[path+x for x in os.listdir(path) if os.path.isdir(path+x)]
imgs=[]
labels=[]
for idx,folder in enumerate(cate):
#glob.glob(s+'*.py') 从目录通配符搜索中生成文件列表
for im in glob.glob(folder+'/*.jpg'):
#输出读取的图片的名称
#print('reading the images:%s'%(im))
#io.imread(im)读取单张RGB图片 skimage.io.imread(fname,as_grey=True)读取单张灰度图片
#读取的图片
img=io.imread(im)
#skimage.transform.resize(image, output_shape)改变图片的尺寸
img=transform.resize(img,(w,h,c))
#将读取的图片数据加载到imgs[]列表中
imgs.append(img)
#将图片的label加载到labels[]中,与上方的imgs索引对应
labels.append(idx)
#将读取的图片和labels信息,转化为numpy结构的ndarr(N维数组对象(矩阵))数据信息
return np.asarray(imgs,np.float32),np.asarray(labels,np.int32)
#调用读取图片的函数,得到图片和labels的数据集
data1,label1=read_img(path1)
#打乱顺序
#读取data矩阵的第一维数(图片的个数)
num_example=data1.shape[0]
#产生一个num_example范围,步长为1的序列
arr=np.arange(num_example)
#调用函数,打乱顺序
np.random.shuffle(arr)
#按照打乱的顺序,重新排序
num_example=data1.shape[0]
#产生一个num_example范围,步长为1的序列
arr=np.arange(num_example)
#调用函数,打乱顺序
np.random.shuffle(arr)
#按照打乱的顺序,重新排序
data1=data1[arr]
label1=label1[arr]
x_train=data1
y_train=label1
#将所有数据分为训练集和验证集
#将所有数据分为训练集和验证集
# ratio=0.8
# s=np.int(num_example*ratio)
# x_train=data1[:s]
# print(x_train.shape)
# y_train=label1[:s]
# x_val=data1[s:]
# y_val=label1[s:]
#本地数据集地址
path2='F:/目标数据集/test/'
#path2="L:/增强眼底图像/test/"
# #调用读取图片的函数,得到图片和labels的数据集
data2,label2=read_img(path2)
x_val=data2
y_val=label2
#读取图片+数据预处理
#函数声明
#-----------------构建网络----------------------
#占位符,设置输入参数的大小和格式
x=tf.placeholder(tf.float32,shape=[None,w,h,c],name='x')
y_=tf.placeholder(tf.int32,shape=[None,],name='y_')
def weight_variable(shape):
initial = tf.truncated_normal(shape, stddev = 0.1)
return tf.Variable(initial)
# 设置阈值函数
def bias_variable(shape):
initial = tf.constant(0.1, shape=shape)
return tf.Variable(initial)
# 设置卷积层
def conv2d(x,W):
return tf.nn.conv2d(x,W,strides=[1,1,1,1],padding = "SAME")
# 设置池化层
def pool(x):
return tf.nn.max_pool(x,ksize=[1,2,2,1],strides = [1,2,2,1],padding = "SAME")
def SE_block(x,ratio):
shape = x.get_shape().as_list()
channel_out = shape[3]
out_shape=int(channel_out/ratio)
# print(shape)
with tf.variable_scope("squeeze_and_excitation"):
# 第一层,全局平均池化层
squeeze = tf.nn.avg_pool(x,[1,shape[1],shape[2],1],[1,shape[1],shape[2],1],padding = "SAME")
# 第二层,全连接层
w_excitation1 = weight_variable([1,1,channel_out,out_shape])
b_excitation1 = bias_variable([out_shape])
excitation1 = conv2d(squeeze,w_excitation1) + b_excitation1
excitation1_output = tf.nn.relu(excitation1)
# 第三层,全连接层
w_excitation2 = weight_variable([1, 1, out_shape, channel_out])
b_excitation2 = bias_variable([channel_out])
excitation2 = conv2d(excitation1_output, w_excitation2) + b_excitation2
excitation2_output = tf.nn.sigmoid(excitation2)
# 第四层,点乘
excitation_output = tf.reshape(excitation2_output,[-1,1,1,channel_out])
h_output = excitation_output * x
return h_output
def normalization(input,dim1):
inputt=input
a=tf.reduce_max(input,axis=1,keepdims=True)
b=tf.reduce_max(a,axis=2,keepdims=True)
max1=b
c=tf.reduce_min(input,axis=1,keepdims=True)
d=tf.reduce_min(c,axis=2,keepdims=True)
min1=d
# con=tf.constant(1e-10,shape=(batch_size,1,1,dim1))
con=1e-10
e=tf.divide(tf.subtract(input,min1),tf.add(tf.subtract(max1,min1),con)) #归一化
return e
# def thresholding(input):
# x=tf.zeros_like(input)
# y=tf.ones_like(input)
# out=tf.where(input>0.2*255,y,x)
# return out
def dilation_conv2d(input,kernel_size, num_o, dilation_factor,name): #空洞卷积
input_channel=input.shape[3].value
with tf.variable_scope(name):
conva1_weights=tf.get_variable('weight',[kernel_size,kernel_size,input_channel,num_o],initializer=tf.truncated_normal_initializer(stddev=0.1))
conva1_biases=tf.get_variable('bias',[num_o],initializer=tf.constant_initializer(0.0))
conva1=tf.nn.atrous_conv2d(input,conva1_weights,dilation_factor,padding='SAME')
output=tf.nn.relu(tf.nn.bias_add(conva1,conva1_biases))
return output
def _add(x_l, name):
return tf.add_n(x_l, name=name)
def ASPP(input, num_o, dilations):
o = []
for i, d in enumerate(dilations):
o.append(dilation_conv2d(input, 3, num_o, d, name='aspp/conv%d' % (i+1)))
return _add(o, name='aspp/add')
def conv2d1(input,kernel_size, num_o,name): #卷积
input_channel=input.shape[3].value
with tf.variable_scope(name):
conva1_weights=tf.get_variable('weight',[kernel_size,kernel_size,input_channel,num_o],initializer=tf.truncated_normal_initializer(stddev=0.1))
conva1_biases=tf.get_variable('bias',[num_o],initializer=tf.constant_initializer(0.0))
conva1=tf.nn.conv2d(input,conva1_weights,strides=[1,1,1,1],padding='SAME')
output=tf.nn.relu(tf.nn.bias_add(conva1,conva1_biases))
return output
def mlevel_block(input): #提取多级特征
o11=conv2d1(input,3,32,'11')
o12=conv2d1(o11,3,32,'12')
o13=conv2d1(o12,3,32,'13')
o21=conv2d1(input,5,32,'21')
o22=conv2d1(o21,5,32,'22')+o11
o23=conv2d1(o22,5,32,'23')+o12
o31=conv2d1(input,7,32,'31')
o32=conv2d1(o31,7,32,'32')+o21+o11
o33=conv2d1(o32,7,32,'33')+o22+o12
return o13+o23+o33
def inference(input_tensor, train, regularizer):
#-----------------------第一层----------------------------
branch1=input_tensor[:,:,:,0] #T1c序列
branch1=tf.expand_dims(branch1,axis=-1)
branch2=input_tensor[:,:,:,1] #T2序列
branch2=tf.expand_dims(branch2,axis=-1)
branch3=input_tensor[:,:,:,2] #T1序列
branch3=tf.expand_dims(branch3,axis=-1)
attention=tf.image.flip_left_right(branch2)
attention1=branch2-attention #获取差分特征图
attention1=normalization(attention1,1)
#attention1=thresholding(attention1)
##feature extraction stage 1
##branch1
with tf.variable_scope('layers-branch1-conv1'): #提取T1C特征
convb11_weights=tf.get_variable('weight',[5,5,1,32],initializer=tf.truncated_normal_initializer(stddev=0.1))
convb11_biases=tf.get_variable('bias',[32],initializer=tf.constant_initializer(0.0))
convb11=tf.nn.conv2d(branch1,convb11_weights,strides=[1,1,1,1],padding='SAME')
relub11=tf.nn.relu(tf.nn.bias_add(convb11,convb11_biases))
with tf.variable_scope('layers-branch1-conv2'):
convb12_weights=tf.get_variable('weight',[3,3,32,32],initializer=tf.truncated_normal_initializer(stddev=0.1))
convb12_biases=tf.get_variable('bias',[32],initializer=tf.constant_initializer(0.0))
convb12=tf.nn.conv2d(relub11,convb12_weights,strides=[1,1,1,1],padding='SAME')
relub12=tf.nn.relu(tf.nn.bias_add(convb12,convb12_biases))
# relub12=mlevel_block(branch1)
relub12=tf.multiply(relub12,attention1)+relub12
with tf.name_scope("branch1-pool1"):
#池化计算,调用tensorflow的max_pool函数,strides=[1,2,2,1],表示池化边界,2个对一个生成,padding="VALID"表示不操作。
pool11 = tf.nn.max_pool(relub12, ksize = [1,2,2,1],strides=[1,2,2,1],padding="VALID")
#branch2
with tf.variable_scope('layers-branch2-conv1'): #提取T2特征
convb21_weights=tf.get_variable('weight',[5,5,1,32],initializer=tf.truncated_normal_initializer(stddev=0.1))
convb21_biases=tf.get_variable('bias',[32],initializer=tf.constant_initializer(0.0))
convb21=tf.nn.conv2d(branch2,convb21_weights,strides=[1,1,1,1],padding='SAME')
relub21=tf.nn.relu(tf.nn.bias_add(convb21,convb21_biases))
with tf.variable_scope('layers-branch2-conv2'):
convb22_weights=tf.get_variable('weight',[3,3,32,32],initializer=tf.truncated_normal_initializer(stddev=0.1))
convb22_biases=tf.get_variable('bias',[32],initializer=tf.constant_initializer(0.0))
convb22=tf.nn.conv2d(relub21,convb22_weights,strides=[1,1,1,1],padding='SAME')
relub22=tf.nn.relu(tf.nn.bias_add(convb22,convb22_biases))
# relub22=mlevel_block(branch2)
relub22=tf.multiply(relub22,attention1)+relub22
with tf.name_scope("branch2-pool1"):
#池化计算,调用tensorflow的max_pool函数,strides=[1,2,2,1],表示池化边界,2个对一个生成,padding="VALID"表示不操作。
pool21 = tf.nn.max_pool(relub22, ksize = [1,2,2,1],strides=[1,2,2,1],padding="VALID")
#branch3
with tf.variable_scope('layers-branch3-conv1'): #提取T1特征
convb31_weights=tf.get_variable('weight',[5,5,1,32],initializer=tf.truncated_normal_initializer(stddev=0.1))
convb31_biases=tf.get_variable('bias',[32],initializer=tf.constant_initializer(0.0))
convb31=tf.nn.conv2d(branch3,convb31_weights,strides=[1,1,1,1],padding='SAME')
relub31=tf.nn.relu(tf.nn.bias_add(convb31,convb31_biases))
with tf.variable_scope('layers-branch3-conv2'):
convb32_weights=tf.get_variable('weight',[3,3,32,32],initializer=tf.truncated_normal_initializer(stddev=0.1))
convb32_biases=tf.get_variable('bias',[32],initializer=tf.constant_initializer(0.0))
convb32=tf.nn.conv2d(relub31,convb32_weights,strides=[1,1,1,1],padding='SAME')
relub32=tf.nn.relu(tf.nn.bias_add(convb32,convb32_biases))
# relub32=mlevel_block(branch3)
relub32=tf.multiply(relub32,attention1)+relub32
# relub32=relub32+relub22
with tf.name_scope("branch1-pool1"):
#池化计算,调用tensorflow的max_pool函数,strides=[1,2,2,1],表示池化边界,2个对一个生成,padding="VALID"表示不操作。
pool31 = tf.nn.max_pool(relub32, ksize = [1,2,2,1],strides=[1,2,2,1],padding="VALID")
##feature merge
feat=tf.concat([pool11,tf.multiply(pool21,2),pool31],axis=3)
with tf.name_scope("attention-attention2"):
#池化计算,调用tensorflow的max_pool函数,strides=[1,2,2,1],表示池化边界,2个对一个生成,padding="VALID"表示不操作。
attention2 = tf.nn.max_pool(attention1, ksize = [1,2,2,1],strides=[1,2,2,1],padding="VALID")
attention3 = tf.nn.max_pool(attention2, ksize = [1,2,2,1],strides=[1,2,2,1],padding="VALID")
feat=tf.multiply(feat,attention2)
feat1=SE_block(feat,ratio=4)
with tf.variable_scope('layers-skip1-conv1'): #提取T1特征
convs11_weights=tf.get_variable('weight',[1,1,32,96],initializer=tf.truncated_normal_initializer(stddev=0.1))
convs11_biases=tf.get_variable('bias',[96],initializer=tf.constant_initializer(0.0))
convs11=tf.nn.conv2d(pool21,convs11_weights,strides=[1,1,1,1],padding='SAME')
relus11=tf.nn.relu(tf.nn.bias_add(convs11,convs11_biases))
feat2=feat1+relus11
# feat2=ASPP(feat2,96,[1,2,4])
##feature extraction2
with tf.variable_scope('layer1-conv1'):
#初始化权重conv1_weights为可保存变量,大小为5x5,3个通道(RGB),数量为32个
conv1_weights = tf.get_variable("weight",[3,3,96,128],initializer=tf.truncated_normal_initializer(stddev=0.1))
#初始化偏置conv1_biases,数量为32个
conv1_biases = tf.get_variable("bias", [128], initializer=tf.constant_initializer(0.0))
#卷积计算,tf.nn.conv2d为tensorflow自带2维卷积函数,input_tensor为输入数据,conv1_weights为权重,strides=[1, 1, 1, 1]表示左右上下滑动步长为1,padding='SAME'表示输入和输出大小一样,即补0
conv1 = tf.nn.conv2d(feat2, conv1_weights, strides=[1, 1, 1, 1], padding='SAME')
#激励计算,调用tensorflow的relu函数
relu1 = tf.nn.relu(tf.nn.bias_add(conv1, conv1_biases))
# relu1=relu1+tf.multiply(relu1,attention2)
with tf.variable_scope('layers-skip2-conv1'): #提取T1特征
convs2_weights=tf.get_variable('weight',[1,1,32,128],initializer=tf.truncated_normal_initializer(stddev=0.1))
convs2_biases=tf.get_variable('bias',[128],initializer=tf.constant_initializer(0.0))
convs2=tf.nn.conv2d(pool21,convs2_weights,strides=[1,1,1,1],padding='SAME')
relus2=tf.nn.relu(tf.nn.bias_add(convs2,convs2_biases))
relu1=relu1+relus2
with tf.name_scope("layer1-pool1"):
#池化计算,调用tensorflow的max_pool函数,strides=[1,2,2,1],表示池化边界,2个对一个生成,padding="VALID"表示不操作。
pool1 = tf.nn.max_pool(relu1, ksize = [1,2,2,1],strides=[1,2,2,1],padding="VALID")
with tf.variable_scope('layer2-conv2'):
#初始化权重conv1_weights为可保存变量,大小为5x5,3个通道(RGB),数量为32个
conv2_weights = tf.get_variable("weight",[3,3,128,256],initializer=tf.truncated_normal_initializer(stddev=0.1))
#初始化偏置conv1_biases,数量为32个
conv2_biases = tf.get_variable("bias", [256], initializer=tf.constant_initializer(0.0))
#卷积计算,tf.nn.conv2d为tensorflow自带2维卷积函数,input_tensor为输入数据,conv1_weights为权重,strides=[1, 1, 1, 1]表示左右上下滑动步长为1,padding='SAME'表示输入和输出大小一样,即补0
conv2 = tf.nn.conv2d(pool1, conv2_weights, strides=[1, 1, 1, 1], padding='SAME')
#激励计算,调用tensorflow的relu函数
relu2 = tf.nn.relu(tf.nn.bias_add(conv2, conv2_biases))
with tf.variable_scope('layers-skip3-conv3'): #提取T1特征
convs3_weights=tf.get_variable('weight',[1,1,32,256],initializer=tf.truncated_normal_initializer(stddev=0.1))
convs3_biases=tf.get_variable('bias',[256],initializer=tf.constant_initializer(0.0))
convs3=tf.nn.conv2d(pool21,convs3_weights,strides=[1,1,1,1],padding='SAME')
relus3=tf.nn.relu(tf.nn.bias_add(convs3,convs3_biases))
pools3=tf.nn.max_pool(relus3, ksize = [1,2,2,1],strides=[1,2,2,1],padding="VALID")
relu2=relu2+pools3
with tf.name_scope("layer2-pool2"):
#池化计算,调用tensorflow的max_pool函数,strides=[1,2,2,1],表示池化边界,2个对一个生成,padding="VALID"表示不操作。
pool2 = tf.nn.max_pool(relu2, ksize = [1,2,2,1],strides=[1,2,2,1],padding="VALID")
with tf.variable_scope('layer3-conv3'):
#初始化权重conv1_weights为可保存变量,大小为5x5,3个通道(RGB),数量为32个
conv3_weights = tf.get_variable("weight",[3,3,256,512],initializer=tf.truncated_normal_initializer(stddev=0.1))
#初始化偏置conv1_biases,数量为32个
conv3_biases = tf.get_variable("bias", [512], initializer=tf.constant_initializer(0.0))
#卷积计算,tf.nn.conv2d为tensorflow自带2维卷积函数,input_tensor为输入数据,conv1_weights为权重,strides=[1, 1, 1, 1]表示左右上下滑动步长为1,padding='SAME'表示输入和输出大小一样,即补0
conv3 = tf.nn.conv2d(pool2, conv3_weights, strides=[1, 1, 1, 1], padding='SAME')
#激励计算,调用tensorflow的relu函数
relu3 = tf.nn.relu(tf.nn.bias_add(conv3, conv3_biases))
with tf.variable_scope('layers-skip4-conv4'): #提取T1特征
convs4_weights=tf.get_variable('weight',[1,1,32,512],initializer=tf.truncated_normal_initializer(stddev=0.1))
convs4_biases=tf.get_variable('bias',[512],initializer=tf.constant_initializer(0.0))
convs4=tf.nn.conv2d(pool21,convs4_weights,strides=[1,1,1,1],padding='SAME')
relus4=tf.nn.relu(tf.nn.bias_add(convs4,convs4_biases))
pools4=tf.nn.max_pool(relus4, ksize = [1,2,2,1],strides=[1,2,2,1],padding="VALID")
pools4=tf.nn.max_pool(pools4, ksize = [1,2,2,1],strides=[1,2,2,1],padding="VALID")
relu3=relu3+pools4
# with tf.name_scope("layer3-pool3"):
# #池化计算,调用tensorflow的max_pool函数,strides=[1,2,2,1],表示池化边界,2个对一个生成,padding="VALID"表示不操作。
# pool3 = tf.nn.max_pool(relu3, ksize = [1,2,2,1],strides=[1,2,2,1],padding="VALID")
# with tf.variable_scope('layer4-conv4'):
# #初始化权重conv1_weights为可保存变量,大小为5x5,3个通道(RGB),数量为32个
# conv4_weights = tf.get_variable("weight",[3,3,512,1024],initializer=tf.truncated_normal_initializer(stddev=0.1))
# #初始化偏置conv1_biases,数量为32个
# conv4_biases = tf.get_variable("bias", [1024], initializer=tf.constant_initializer(0.0))
# #卷积计算,tf.nn.conv2d为tensorflow自带2维卷积函数,input_tensor为输入数据,conv1_weights为权重,strides=[1, 1, 1, 1]表示左右上下滑动步长为1,padding='SAME'表示输入和输出大小一样,即补0
# conv4 = tf.nn.conv2d(pool3, conv4_weights, strides=[1, 1, 1, 1], padding='SAME')
# #激励计算,调用tensorflow的relu函数
# relu4 = tf.nn.relu(tf.nn.bias_add(conv4, conv4_biases))
with tf.name_scope("layer4-pool4"):
#池化计算,调用tensorflow的max_pool函数,strides=[1,2,2,1],表示池化边界,2个对一个生成,padding="VALID"表示不操作。
pool4 = tf.nn.max_pool(relu3, ksize = [1,2,2,1],strides=[1,2,2,1],padding="VALID")
pool4=tf.keras.layers.GlobalAveragePooling2D()(pool4)
# nodes = 4*4*512
# reshaped = tf.reshape(pool4,[-1,nodes])
# #-----------------------第五层----------------------------
# with tf.variable_scope('layer11-fc1'):
# #初始化全连接层的参数,隐含节点为1024个
# fc1_weights = tf.get_variable("weight", [nodes, 1024],
# initializer=tf.truncated_normal_initializer(stddev=0.1))
# if regularizer != None: tf.add_to_collection('losses', regularizer(fc1_weights))
# fc1_biases = tf.get_variable("bias", [1024], initializer=tf.constant_initializer(0.1))
# #使用relu函数作为激活函数
# fc1 = tf.nn.relu(tf.matmul(reshaped, fc1_weights) + fc1_biases)
# #采用dropout层,减少过拟合和欠拟合的程度,保存模型最好的预测效率
# if train: fc1 = tf.nn.dropout(fc1, 0.5)
# #-----------------------第六层----------------------------
# with tf.variable_scope('layer12-fc2'):
# #同上,不过参数的有变化,根据卷积计算和通道数量的变化,设置对应的参数
# fc2_weights = tf.get_variable("weight", [1024, 512],
# initializer=tf.truncated_normal_initializer(stddev=0.1))
# if regularizer != None: tf.add_to_collection('losses', regularizer(fc2_weights))
# fc2_biases = tf.get_variable("bias", [512], initializer=tf.constant_initializer(0.1))
# fc2 = tf.nn.relu(tf.matmul(fc1, fc2_weights) + fc2_biases)
# if train: fc2 = tf.nn.dropout(fc2, 0.5)
#-----------------------第七层----------------------------
with tf.variable_scope('layer13-fc3'):
#同上,不过参数的有变化,根据卷积计算和通道数量的变化,设置对应的参数
fc3_weights = tf.get_variable("weight", [512, 2],
initializer=tf.truncated_normal_initializer(stddev=0.1))
if regularizer != None: tf.add_to_collection('losses', regularizer(fc3_weights))
fc3_biases = tf.get_variable("bias", [2], initializer=tf.constant_initializer(0.1))
logit = tf.matmul(pool4, fc3_weights) + fc3_biases
return logit
#---------------------------网络结束---------------------------
#设置正则化参数为0.0001
regularizer = tf.contrib.layers.l2_regularizer(0.0001)
#将上述构建网络结构引入
logits = inference(x,False,regularizer)
#(小处理)将logits乘以1赋值给logits_eval,定义name,方便在后续调用模型时通过tensor名字调用输出tensor
b = tf.constant(value=1,dtype=tf.float32)
logits_eval = tf.multiply(logits,b,name='logits_eval')
#设置损失函数,作为模型训练优化的参考标准,loss越小,模型越优
loss=tf.nn.sparse_softmax_cross_entropy_with_logits(logits=logits, labels=y_)
loss1=tf.add_n(tf.get_collection('losses'))
loss2=loss+loss1
#设置整体学习率为α为0.001
# batch_size=16
# global_step=tf.Variable(0,trainable=False)
# decay_ste=label1.shape[0]/batch_size
# learning_rate_base=0.01
# decay_rate=0.99
# learning_rate=tf.train.exponential_decay(learning_rate_base,global_step,decay_ste,decay_rate)
# train_op=tf.train.AdamOptimizer(learning_rate).minimize(loss,global_step=global_step)
train_op=tf.train.AdamOptimizer(learning_rate=0.001).minimize(loss2)
#设置预测精度
correct_prediction = tf.equal(tf.cast(tf.argmax(logits,1),tf.int32), y_)
acc= tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
#定义一个函数,按批次取数据
def minibatches(inputs=None, targets=None, batch_size=None, shuffle=False):
assert len(inputs) == len(targets)
if shuffle:
indices = np.arange(len(inputs))
np.random.shuffle(indices)
for start_idx in range(0, len(inputs) - batch_size + 1, batch_size):
if shuffle:
excerpt = indices[start_idx:start_idx + batch_size]
else:
excerpt = slice(start_idx, start_idx + batch_size)
yield inputs[excerpt], targets[excerpt]
#训练和测试数据,可将n_epoch设置更大一些
#迭代次数
n_epoch=40
#每次迭代输入的图片数据
batch_size=16
saver=tf.train.Saver(max_to_keep=1)
sess=tf.Session()
#初始化全局参数
sess.run(tf.global_variables_initializer())
max_acc=0
#开始迭代训练,调用的都是前面设置好的函数或变量
for epoch in range(n_epoch):
start_time = time.time()
#training
train_loss, train_acc, n_batch = 0, 0, 0
i=0
for x_train_a, y_train_a in minibatches(x_train, y_train, batch_size, shuffle=True):
i=i+1
_,err,ac=sess.run([train_op,loss,acc], feed_dict={x: x_train_a, y_: y_train_a})
train_loss += err; train_acc += ac; n_batch += 1
print(" train loss: %f" % (np.sum(train_loss)/ n_batch))
print(" train acc: %f" % (np.sum(train_acc)/ n_batch))
#validation
val_loss, val_acc, n_batch = 0, 0, 0
for x_val_a, y_val_a in minibatches(x_val, y_val, batch_size, shuffle=False):
err, ac = sess.run([loss,acc], feed_dict={x: x_val_a, y_: y_val_a})
val_loss += err; val_acc += ac; n_batch += 1
print(" validation loss: %f" % (np.sum(val_loss)/ n_batch))
print(" validation acc: %f" % (np.sum(val_acc)/ n_batch))
if val_acc>=max_acc:
max_acc=val_acc;
saver.save(sess,'ckpt2/train.ckpt',global_step=epoch+1)
#保存模型及模型参数
sess.close()