-
Notifications
You must be signed in to change notification settings - Fork 16
/
product.py
291 lines (248 loc) · 12.8 KB
/
product.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
# -*- coding:utf-8 -*-
"""
@Brief
ACSCP model:model building, model training and testing
source:Crowd Counting via Adversarial Cross-Scale Consistency Pursuit
https://pan.baidu.com/s/1mjPpKqG
@Description
using trained ACSCP model to estimate crowd map, this faces with production environment
@Reference
@Author: Ling Bao
@Data: April 12, 2018
@Version: 0.1.0
"""
# 系统库
from math import ceil
import numpy as np
import time
from matplotlib import pyplot as plt
import cv2
# 项目库
from lib_ops.ops import *
# 机器学习库
import tensorflow as tf
slim = tf.contrib.slim
class ProductMap(object):
def __init__(self, use_dropout=True):
# 模型相关
self.sess = tf.Session()
self.checkpoints = "./gan_mp_bn_1_240/"
self.g_large_name = "g_large.ckpt"
# self.g_large_path = "./product_model/g_large_model_412/" # have dropout
self.g_large_path = "./product_model/g_large_model_414/" # no dropout
# 输入图像
self.x = tf.placeholder(tf.float32, [1, 720, 720, 3])
# 批量归一化——large生成器
self.g_L_bn_e1 = batch_norm(name='g_L_bn_e1')
self.g_L_bn_e2 = batch_norm(name='g_L_bn_e2')
self.g_L_bn_e3 = batch_norm(name='g_L_bn_e3')
self.g_L_bn_e4 = batch_norm(name='g_L_bn_e4')
self.g_L_bn_e5 = batch_norm(name='g_L_bn_e5')
self.g_L_bn_e6 = batch_norm(name='g_L_bn_e6')
self.g_L_bn_e7 = batch_norm(name='g_L_bn_e7')
self.g_L_bn_e8 = batch_norm(name='g_L_bn_e8')
self.g_L_bn_d1 = batch_norm(name='g_L_bn_d1')
self.g_L_bn_d2 = batch_norm(name='g_L_bn_d2')
self.g_L_bn_d3 = batch_norm(name='g_L_bn_d3')
self.g_L_bn_d4 = batch_norm(name='g_L_bn_d4')
self.g_L_bn_d5 = batch_norm(name='g_L_bn_d5')
self.g_L_bn_d6 = batch_norm(name='g_L_bn_d6')
self.g_L_bn_d7 = batch_norm(name='g_L_bn_d7')
# 构建模型
self.crowd_map = self.generator_large(self.x, use_dropout=use_dropout)
def generator_large(self, image, batch_size=1, use_dropout=True, reuse=False):
"""
Large生成器网络
:param image: 输入数据
:param batch_size 批量数,默认为1
:param use_dropout: 是否使用dropout
:param reuse:
:return: 生成图片
"""
with tf.variable_scope("generator_large"):
if reuse:
tf.get_variable_scope().reuse_variables()
else:
assert tf.get_variable_scope().reuse is False
# input image size is 240 x 240 x 3, input_c_dim = output_c_dim = 3
# (240 x 240 x input_c_dim) --> e1(120 x 120 x 64) --> e2(60 x 60 x 64) --> e3(30 x 30 x 64) -->
# e4(15 x 15 x 64) --> e5(8 x 8 x 64) --> e6(4 x 4 x 64) --> e7(2 x 2 x 64) --> e8(2 x 2 x 64) <--
# d1(2 x 2 x 64*2) <-- d2(4 x 4 x 64*2) <-- d3(8 x 8 x 64*2) <-- d4(15 x 15 x 64*2) <--
# d5(30 x 30 x 64*2) <-- d6(60 x 60 x 64*2) <-- d7(120 x 120 x 64*2) <-- (240 x 240 x output_c_dim)
# general method, input image size is w x h x c, limit to w, h more greater 120, c is equal to 3
# (w x h x c) --> e1(c[w/2] x c[h/2] x 64) --> e2(c[w/4] x c[h/4] x 64) --> e3(c[w/8] x c[h/8] x 64) -->
# e4(c[w/16] x c[h/16] x 64) --> e5(c[w/32] x c[h/32] x 64) --> e6(c[w/64] x c[h/64] x 64) -->
# e7(c[w/128] x c[h/128] x 64) --> e8(c[w/128] x c[h/128] x 64) <-- d1(c[w/128] x c[h/128] x 64*2) <--
# d2(c[w/64] x c[h/64] x 64*2) <-- d3(c[w/32] x c[h/32] x 64*2) <-- d4(c[w/16] x c[h/16] x 64*2) <--
# d5(c[w/8] x c[h/8] x 64*2) <-- d6(c[w/4] x c[h/4] x 64*2) <-- d7(c[w/2] x c[h/2] x 64*2) <-- (w x h x c)
w = int(np.array(self.x.shape[1]))
h = int(np.array(self.x.shape[2]))
if use_dropout:
e1 = self.g_L_bn_e2(conv2d(image, output_dim=64, k_h=6, k_w=6, d_h=2, d_w=2, name='g_L_e1_con'))
else:
e1 = self.g_L_bn_e1(conv2d(image, output_dim=64, k_h=6, k_w=6, d_h=2, d_w=2, name='g_L_e1_con'))
e2 = self.g_L_bn_e2(conv2d(lrelu(e1), output_dim=64, k_h=4, k_w=4, d_h=2, d_w=2, name='g_L_e2_con'))
e3 = self.g_L_bn_e3(conv2d(lrelu(e2), output_dim=64, k_h=4, k_w=4, d_h=2, d_w=2, name='g_L_e3_con'))
e4 = self.g_L_bn_e4(conv2d(lrelu(e3), output_dim=64, k_h=4, k_w=4, d_h=2, d_w=2, name='g_L_e4_con'))
e5 = self.g_L_bn_e5(conv2d(lrelu(e4), output_dim=64, k_h=4, k_w=4, d_h=2, d_w=2, name='g_L_e5_con'))
e6 = self.g_L_bn_e6(conv2d(lrelu(e5), output_dim=64, k_h=4, k_w=4, d_h=2, d_w=2, name='g_L_e6_con'))
e7 = self.g_L_bn_e7(conv2d(lrelu(e6), output_dim=64, k_h=4, k_w=4, d_h=2, d_w=2, name='g_L_e7_con'))
e8 = self.g_L_bn_e8(conv2d(lrelu(e7), output_dim=64, k_h=4, k_w=4, d_h=1, d_w=1, name='g_L_e8_con'))
d1, _, _ = deconv2d(lrelu(e8), [batch_size, int(ceil(w / 128.)), int(ceil(h / 128.)), 64], k_h=4, k_w=4,
d_h=1, d_w=1, name='g_L_d1', with_w=True)
if use_dropout:
d1 = tf.nn.dropout(self.g_L_bn_d1(d1), 0.5)
d1 = tf.concat([d1, e7], 3)
d2, _, _ = deconv2d(tf.nn.relu(d1), [batch_size, int(ceil(w / 64.)), int(ceil(h / 64.)), 64], k_h=4, k_w=4,
d_h=2, d_w=2, name='g_L_d2', with_w=True)
if use_dropout:
d2 = tf.nn.dropout(self.g_L_bn_d2(d2), 0.5)
d2 = tf.concat([d2, e6], 3)
d3, _, _ = deconv2d(tf.nn.relu(d2), [batch_size, int(ceil(w / 32.)), int(ceil(h / 32.)), 64], k_h=4, k_w=4,
d_h=2, d_w=2, name='g_L_d3', with_w=True)
if use_dropout:
d3 = tf.nn.dropout(self.g_L_bn_d3(d3), 0.5)
d3 = tf.concat([d3, e5], 3)
d4, _, _ = deconv2d(tf.nn.relu(d3), [batch_size, int(ceil(w / 16.)), int(ceil(h / 16.)), 64], k_h=4, k_w=4,
d_h=2, d_w=2, name='g_L_d4', with_w=True)
d4 = self.g_L_bn_d4(d4)
d4 = tf.concat([d4, e4], 3)
d5, _, _ = deconv2d(tf.nn.relu(d4), [batch_size, int(ceil(w / 8.)), int(ceil(h / 8.)), 64], k_h=4, k_w=4,
d_h=2, d_w=2, name='g_L_d5', with_w=True)
d5 = self.g_L_bn_d5(d5)
d5 = tf.concat([d5, e3], 3)
d6, _, _ = deconv2d(tf.nn.relu(d5), [batch_size, int(ceil(w / 4.)), int(ceil(h / 4.)), 64], k_h=4, k_w=4,
d_h=2, d_w=2, name='g_L_d6', with_w=True)
d6 = self.g_L_bn_d6(d6)
d6 = tf.concat([d6, e2], 3)
d7, _, _ = deconv2d(tf.nn.relu(d6), [batch_size, int(ceil(w / 2.)), int(ceil(h / 2.)), 64], k_h=4, k_w=4,
d_h=2, d_w=2, name='g_L_d7', with_w=True)
d7 = self.g_L_bn_d7(d7)
d7 = tf.concat([d7, e1], 3)
d8, _, _ = deconv2d(tf.nn.relu(d7), [batch_size, int(w), int(h), 3], k_h=6, k_w=6,
d_h=2, d_w=2, name='g_L_d8', with_w=True)
return tf.nn.relu(tf.nn.sigmoid(d8))
def generator_large_save(self):
"""
保存ACSCP模型中的generator_large模型
"""
with tf.Session() as sess:
# 载入ACSCP模型参数并对generator_large模型参数进行初始化
saver = tf.train.Saver()
ckpt = tf.train.get_checkpoint_state(self.checkpoints)
saver.restore(sess, ckpt.model_checkpoint_path)
# 保存vgg2模型
saver.save(sess, self.g_large_path + self.g_large_name)
# 关闭session
sess.close()
def generator_large_load(self):
"""
载入generator_large模型
"""
saver = tf.train.Saver()
ckpt = tf.train.get_checkpoint_state(self.g_large_path)
saver.restore(self.sess, ckpt.model_checkpoint_path)
def run(self, image):
"""
利用vgg2模型对images进行特征提取
:param image 待估计图像
:return: 图像特征
"""
start_time = time.time()
data = np.array([image]).astype(np.float32)
tmp_mp = self.sess.run(self.crowd_map, feed_dict={self.x: data})
run_time = time.time() - start_time
mp_crowd = np.mean(tmp_mp[0], axis=2)
return mp_crowd, run_time
class VGGFTest(tf.test.TestCase):
"""
对vgg_2模型进行单元测试
"""
def test_build(self):
batch_size = 1
height, width = 720, 720
with self.test_session():
inputs = tf.random_uniform((batch_size, height, width, 3))
g_large = ProductMap()
g_large.x = inputs
expected_names = [
'generator_large/g_L_bn_d2/moving_variance',
'generator_large/g_L_bn_d4/beta',
'generator_large/g_L_bn_d3/moving_variance',
'generator_large/g_L_bn_d7/gamma',
'generator_large/g_L_bn_e8/moving_variance',
'generator_large/g_L_bn_d5/gamma',
'generator_large/g_L_bn_e7/moving_mean',
'generator_large/g_L_bn_d3/gamma',
'generator_large/g_L_bn_e6/moving_variance',
'generator_large/g_L_bn_e6/beta',
'generator_large/g_L_bn_e4/beta',
'generator_large/g_L_bn_e5/moving_mean',
'generator_large/g_L_bn_d1/moving_variance',
'generator_large/g_L_bn_e7/gamma',
'generator_large/g_L_bn_d2/moving_mean',
'generator_large/g_L_bn_d4/moving_variance',
'generator_large/g_L_bn_d1/moving_mean',
'generator_large/g_L_bn_e5/gamma',
'generator_large/g_L_bn_e8/moving_mean',
'generator_large/g_L_bn_d5/moving_mean',
'generator_large/g_L_bn_e5/beta',
'generator_large/g_L_bn_e6/gamma',
'generator_large/g_L_bn_d1/beta',
'generator_large/g_L_bn_e4/moving_mean',
'generator_large/g_L_bn_e3/beta',
'generator_large/g_L_bn_e2/gamma',
'generator_large/g_L_bn_e8/gamma',
'generator_large/g_L_bn_d2/gamma',
'generator_large/g_L_bn_e4/gamma',
'generator_large/g_L_bn_d3/moving_mean',
'generator_large/g_L_bn_e6/moving_mean',
'generator_large/g_L_bn_e8/beta',
'generator_large/g_L_bn_d4/moving_mean',
'generator_large/g_L_bn_d4/gamma',
'generator_large/g_L_bn_d5/moving_variance',
'generator_large/g_L_bn_d7/beta',
'generator_large/g_L_bn_d6/moving_mean',
'generator_large/g_L_bn_d6/moving_variance',
'generator_large/g_L_bn_e4/moving_variance',
'generator_large/g_L_bn_e2/beta',
'generator_large/g_L_bn_d7/moving_variance',
'generator_large/g_L_bn_d5/beta',
'generator_large/g_L_bn_e7/beta',
'generator_large/g_L_bn_d3/beta',
'generator_large/g_L_bn_e3/moving_mean',
'generator_large/g_L_bn_e5/moving_variance',
'generator_large/g_L_bn_d2/beta',
'generator_large/g_L_bn_e7/moving_variance',
'generator_large/g_L_bn_e2/moving_mean',
'generator_large/g_L_bn_d1/gamma',
'generator_large/g_L_bn_e3/gamma',
'generator_large/g_L_bn_d7/moving_mean',
'generator_large/g_L_bn_d6/beta',
'generator_large/g_L_bn_e2/moving_variance',
'generator_large/g_L_bn_e3/moving_variance',
'generator_large/g_L_bn_d6/gamma']
model_variables = [v.op.name for v in slim.get_model_variables()]
self.assertSetEqual(set(model_variables), set(expected_names))
if __name__ == "__main__":
# ****************************************************模型测试***************************************************** #
# # TF模型结构单元测试
# tf.test.main()
# ****************************************************接口示例***************************************************** #
# 载入图像
img_path = "data/data_im/test_im/"
img_name = "IMG_2_A"
image = cv2.imread(img_path + img_name + ".jpg")
if image is None:
print("Please check image path!!")
exit(0)
# 人群密度估计
# product = ProductMap(True) # have dropout
product = ProductMap(False) # no dropout
# product.generator_large_save() # 仅用于提取generator_larger模型参数并保存
product.generator_large_load()
mp, time = product.run(image)
# 保存估计的人群密度图
mp_name = img_name
print("Time: %4.4f, Estimation numbers: %4d" % (time, round(sum(sum(mp)))))
plt.imsave(mp_name + ".png", mp, cmap=plt.get_cmap('jet'))