-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathrun.py
131 lines (126 loc) · 4.15 KB
/
run.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
import tensorflow as tf
import numpy as np
import os
import cv2
import model
import random
import matplotlib.pyplot as plt
import config as cfg
charset=cfg.charset
num_class=cfg.num_class
width=cfg.width
height=cfg.height
epoch=cfg.epoch
batch=cfg.batch
model_ckpt = cfg.model_ckpt
train_dir=cfg.train_dir
test_dir=cfg.test_dir
pretrained_path=cfg.pretrained_path
def read_data(dir_name):
train_dir =dir_name
img_names = os.listdir(train_dir)
data=[]
x=[]
y=[]
for img_name in img_names:
img=cv2.imread(train_dir+'/'+img_name,0)
img=cv2.resize(img,(width,height))
#plt.imshow(img,cmap='gray')
#plt.show()
img=img[:,:,np.newaxis]
label = img_name.split('.')[0].split('_')[1]#img_name:0_A.jpg
data.append([img,label])
#random.shuffle(data)
return data
def shuffle(data):
x=[]
y=[]
random.shuffle(data)
for i in range(len(data)):
x.append(data[i][0])
y.append(data[i][1])
length = len(y)
label = np.zeros([length, num_class])
for i in range(length):
index = charset.index(y[i])
label[i][index] = 1
return np.array(x),label
def read_test(dirs):
test_dir = dirs
img_names = os.listdir(test_dir)
x = []
y = []
for img_name in img_names:
img = cv2.imread(test_dir + '/' + img_name, 0)
img = cv2.resize(img, (width,height))
img = img[:, :, np.newaxis]
label = img_name.split('.')[0].split('_')[1]
x.append(img)
y.append(label)
length = len(y)
label = np.zeros([length, num_class])
for i in range(length):
index = charset.index(y[i])
label[i][index] = 1
#print(y[i],label[i])
return np.array(x),np.array(label)
def get_label(img,sess,cnn):#img shape:[img,img,img]
predict = sess.run(cnn.predict, feed_dict={cnn.x: img})
label=''
for i in range(len(img)):
temp=np.argwhere(predict[i]==1.0)
if(len(temp)==1):
index=int(np.argwhere(predict[i]==1.0))
label+=charset[index]
def label_img(testdir):#recognize images from an dir
img_names = os.listdir(testdir)
x = []
for img_name in img_names:
img = cv2.imread(testdir + '/' + img_name, 0)
img = cv2.resize(img, (width, height))
img = img[:, :, np.newaxis]
x.append(img)
x=np.array(x)
print('have load image',len(img_names))
sess = tf.InteractiveSession()
cnn = model.cnn_ocr()
saver = tf.train.Saver()
sess.run(tf.global_variables_initializer())
path =pretrained_path
saver.restore(sess, path)
print('model has been restored')
predict=sess.run(cnn.predict,feed_dict={cnn.x:x})
#decode label
for i in range(len(img_names)):
temp=np.argwhere(predict[i]==1.0)
if(len(temp)==1):
index=int(np.argwhere(predict[i]==1.0))
os.rename(testdir + '/' + img_names[i],
testdir + '/' + str(i) + '_' + charset[index] + '.jpg')
def train(Is_restore=False):
sess = tf.InteractiveSession()
cnn = model.cnn_ocr()
optimizer=tf.train.AdadeltaOptimizer(0.01)
train_step = optimizer.minimize(cnn.loss)
saver = tf.train.Saver()
sess.run(tf.global_variables_initializer())
if Is_restore:
saver.restore(sess,pretrained_path)
print('model has been restored')
data=read_data(train_dir)
print('have load image:',len(data))
x_test,y_test=read_test(test_dir)
print('begin training---------------------------------------------------------------')
for i in range(epoch):
x, y = shuffle(data)
x = x / 255
num_data = len(y)
for j in range(int(num_data/batch)):
_,trainloss,trainacc=sess.run([train_step,cnn.loss,cnn.acc],feed_dict={cnn.x:x[j*batch:(j+1)*batch],cnn.y:y[j*batch:(j+1)*batch]})
if j%50==0:
#trainacc=sess.run(cnn.acc,feed_dict={cnn.x:x[0:500],cnn.y:y[0:500]})
acc=sess.run(cnn.acc,feed_dict={cnn.x:x_test,cnn.y:y_test})
print('epoch:',i,' trainloss:',trainloss,' trainacc:',trainacc,' acc:',acc)
saver.save(sess,model_ckpt,global_step=i)
train(Is_restore=False)
#label_img('')