-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathface_detect.py
206 lines (179 loc) · 7.68 KB
/
face_detect.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
#!D:/workplace/python
# -*- coding: utf-8 -*-
# @File : face_detect.py
# @Author: WangYe
# @Date : 2019/4/17
# @Software: PyCharm
import tensorflow as tf
import numpy as np
from .face_prepare import readData
from sklearn.model_selection import train_test_split
import time
max_steps = 3000
batch_size = 1
data = np.array(readData()[0])#(1028, 1750, 1000, 3)
label = [[10,10,5],[10,10,5],[10,10,5],[10,10,5],[10,10,5],[10,10,5],[10,10,5],[10,10,5],[10,10,5]]
# label = np.array(readData()[1])#(1028,250, 250, 3)
label = np.array(label)
# print(data[8].shape)
# print(data.shape)
#切分训练集和测试集
x_train,x_test,y_train,y_test = train_test_split(data,label,test_size=0.3,random_state=0)
print('x_train',len(x_train))
# print('x_test',x_test.shape())
# print('y_train',y_train.shape())
# print('y_test',y_test.shape())
def weight_variable(shape):
#这里是构建初始变量
initial = tf.truncated_normal(shape, mean=0,stddev=0.1)
#创建变量
return tf.Variable(initial)
def bias_variable(shape):
initial = tf.constant(0.1, shape=shape)
return tf.Variable(initial)
#常规卷积函数
def con_wy(input,shape_weight,shape_bias,strides):
weight = weight_variable(shape_weight)
kernel = tf.nn.conv2d(input, weight, strides, padding='SAME')
bias = tf.Variable(tf.constant(0.1, shape=[shape_bias]))
conv= tf.nn.relu(kernel + bias)
return conv
#空洞卷积函数
def atrous_conv_wy(input,shape_weight,shape_bias,rate):
weight = weight_variable(shape_weight)
kernel = tf.nn.atrous_conv2d(input, weight, rate=rate, padding='SAME') # 空洞卷积会把上面的卷积核编程 10*10的大小
bias = tf.Variable(tf.constant(0.1, shape=[shape_bias])) # 输出维度128
conv = tf.nn.relu(kernel + bias) # 输出为b*7*7*128
return conv
image_holder = tf.placeholder(tf.float32, [batch_size, 1750, 1000,1])
label_holder = tf.placeholder(tf.float32, [batch_size,None,None,None])
#输入为 1600*1000*3
weight1 = weight_variable(shape=[5,5, 1, 1])
kernel1 = tf.nn.conv2d(image_holder, weight1, [1,7, 4, 1], padding='SAME')
bias1 = tf.Variable(tf.constant(0.0, shape=[3]))
# conv1 = tf.nn.relu(tf.nn.bias_add(kernel1,bias1))
conv1 = tf.nn.relu(kernel1 + bias1) #250*250*3
pool1 = tf.nn.max_pool(conv1, ksize=[1, 3, 3, 1], strides=[1, 1, 1, 1], padding='SAME')
norm1 = tf.nn.lrn(pool1, 4, bias=1.0, alpha=0.001 / 9.0, beta=0.75) #400*250*64
# print(label_holder)
# def loss(logits, labels):
# labels1 = tf.cast(labels, tf.float64)
# logits1 = tf.cast(logits, tf.float64)
# logits2 = tf.nn.softmax(logits1)
# labels2 = tf.nn.softmax(labels1)
# print("predict::::::::::::", local3)
# print("label::::::::::::::", label_holder)
# cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits=logits2, labels=labels2,
# name='cross_entropy_per_example')
# cross_entropy_mean = tf.reduce_mean(cross_entropy, name='cross_entropy')
# tf.add_to_collection('losses', cross_entropy_mean)
# return tf.add_n(tf.get_collection('losses'), name='total_loss')
#两张图片的欧氏距离
# def image_dist():
# print("norm1",norm1)
norm0 = tf.reshape(norm1,[batch_size,-1])
dim = norm0.get_shape()[1].value
weight3 = weight_variable([dim, 250*250])
# dmi1 = np.ravel(dim)
bias3 = tf.Variable(tf.constant(0.1, shape=[250*250]))
# local3 = tf.matmul(norm0,weight3)
local3 = tf.matmul(norm0, weight3)+bias3
print("local3",local3)
print("label_holder",label_holder)
#
cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits=local3, labels=label_holder,
name='cross_entropy_per_example')
# print("local3",local3)#
# dis = tf.square(local3-label_holder)
# # print("dis",dis)#
# dis1 =tf.reduce_sum(dis)
# # print("dis1",dis1)
# loss = tf.sqrt(tf.sqrt(tf.reduce_sum(dis1)))
opt = tf.train.AdamOptimizer(1e-4).compute_gradients(cross_entropy)
#
# print("euclidean",loss)
# dis = tf.norm(local3,label_holder)
# print(dis)
# dis = tf.sqrt(tf.reduce_sum(tf.square(label_holder-local3), 2))
# print(dis)
# with tf.Session() as sess:
# # dis = tf.sqrt(tf.reduce_sum(tf.square(label_holder-local3), 2))
# image_batch, label_batch = sess.run([x_train, y_train])
# out1 = sess.run([local3-label_holder],feed_dict={image_holder: image_batch, label_holder: label_batch})
# out2 = sess.run(tf.reduce_sum(tf.square(out1), 2))
# out3 = sess.run(tf.sqrt(out2))
# # print("out1",out1)
# # print("out2",out2)
# print("out3",out3)
# loss = loss(local3,label_holder)
# cross_entropy = tf.reduce_mean(-tf.reduce_sum(label_holder * tf.log(local3), reduction_indices=[1])) # 损失函数，交叉熵
# print("okkkkkkkkkkkkkkkkkkkkkkkkkkkkk")
# train_op = tf.train.AdamOptimizer(1e-3).minimize(cross_entropy)
# top_k_op = tf.nn.in_top_k(norm1, tf.cast(label_holder, tf.int64), 1)
# def get_batch_data(x,y):
# images,label = x,y
# input_queue = tf.train.slice_input_producer([images, label], shuffle=False,num_epochs=2)
# image_batch, label_batch = tf.train.batch(input_queue, batch_size=1,
# num_threads=1, capacity=64,
# allow_smaller_final_batch=False)
# return image_batch,label_batch
#
# def next_batch(train_data, train_target, batch_size=4):
# #打乱数据集
# #index = [ i for i in range(0,len(train_target)) ]
# # np.random.shuffle(index)
# #建立batch_data与batch_target的空列表
# batch_data = []
# batch_target = []
# #向空列表加入训练集及标签
# # print(train_data.shape)
# for i in range(0,batch_size):
# # print(train_data[i].shape)
# batch_data.append(train_data[i])
# batch_target.append(train_target[i])
# return batch_data, batch_target #返回
sess = tf.InteractiveSession()
tf.global_variables_initializer().run()
tf.train.start_queue_runners()
for step in range(10000):
start_time = time.time()
# image_batch, label_batch = next_batch(data,label)
# print("image_batch",image_batch)
# print("label_batch",label_batch)
for k in range(4):
print('++++++++++++++')
# print(x_train.shape())
# print(y_train.shape())
opt,test_out = sess.run([loss,
opt],
feed_dict={image_holder: [data[k]],
label_holder: [label[k]]})
print(test_out,opt)
# _, loss_value = sess.run([train_op, cross_entropy], feed_dict={image_holder: image_batch, label_holder: label_batch})
# duration = time.time() - start_time
# if step % 10 == 0:
# examples_per_sec = batch_size / duration
# sec_per_batch = float(duration)
#
# format_str = ('step %d,loss=%.2f (%.1f examples/sec;%.3f sec/batch)')
# print(format_str % (step, loss_value, examples_per_sec, sec_per_batch))
#
# num_examples = 10000
# import math
#
# num_iter = int(math.ceil(num_examples / batch_size))
# true_count = 0
# total_sample_count = num_iter * batch_size
# step = 0
# # with tf.Session() as sess:
# while step < num_iter:
# image_batch, label_batch = sess.run([images_test, labels_test])
# predictions = sess.run([top_k_op], feed_dict={image_holder: image_batch,
# label_holder: label_batch})
# true_count += np.sum(predictions)
# step += 1
# if step % 10 == 0:
# print(true_count)
#
# precision = float(true_count) / total_sample_count
# print('precision @ 1 =%.3f' % precision)