-
Notifications
You must be signed in to change notification settings - Fork 7
/
Copy pathcrack_number_captcha.py
178 lines (143 loc) · 7.49 KB
/
crack_number_captcha.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
# -*- coding: utf-8 -*-
"""
REF:http://blog.topspeedsnail.com/archives/10858#more-10858
"""
import random
import os
from captcha.image import ImageCaptcha # pip install captcha
from PIL import Image
import numpy as np
import tensorflow as tf
import tensorflow.contrib.slim as slim
NUMBERS = map(str, range(10))
CHAR_SET = NUMBERS
CHAR_SET_LEN = len(CHAR_SET)
CHAR_POS_MAP = {v: k for k, v in enumerate(CHAR_SET)}
POS_CHAR_MAP = {v: k for k, v in CHAR_POS_MAP.iteritems()}
CAPTCHA_LEN = 4
def random_captcha_text(char_set=NUMBERS, captcha_size=4):
return ''.join(random.choice(char_set) for _ in range(captcha_size))
def gen_captcha_text_and_image():
image = ImageCaptcha()
captcha_text = random_captcha_text()
captcha = image.generate(captcha_text)
# write the image to HD
# image.write(captcha_text, captcha_text + '.jpg')
captcha_image = Image.open(captcha)
captcha_image = np.array(captcha_image)
# captcha_grey_image = convert2gray(captcha_image)
# im = Image.fromarray(captcha_grey_image).convert('RGB')
# im.save('grey_' + captcha_text + '.jpeg')
return captcha_text, captcha_image
def convert2gray(img):
if len(img.shape) > 2:
# gray = np.mean(img, -1)
# 上面的转法较快，正规转法如下
r, g, b = img[:, :, 0], img[:, :, 1], img[:, :, 2]
gray = 0.2989 * r + 0.5870 * g + 0.1140 * b
return gray
else:
return img
def text2vec(text):
vec = np.zeros(CAPTCHA_LEN * CHAR_SET_LEN)
for i, c in enumerate(text):
pos = CHAR_POS_MAP[c] + i * CHAR_SET_LEN
vec[pos] = 1
return vec
def vec2text(vec):
nonzero_pos = vec.nonzero()[0]
chars = []
for i, p in enumerate(nonzero_pos):
chars.append(POS_CHAR_MAP[i])
return ''.join(chars)
IMAGE_HEIGHT = 60
IMAGE_WIDTH = 160
"""
cnn在图像大小是2的倍数时性能最高, 如果你用的图像大小不是2的倍数，可以在图像边缘补无用像素。
np.pad(image,((2,3),(2,2)), 'constant', constant_values=(255,)) # 在图像上补2行，下补3行，左补2行，右补2行
"""
def get_next_batch(batch_size=128):
batch_x = np.zeros([batch_size, IMAGE_HEIGHT * IMAGE_WIDTH])
batch_y = np.zeros([batch_size, CAPTCHA_LEN * CHAR_SET_LEN])
def loop_get_text_and_image():
while True:
text, image = gen_captcha_text_and_image()
if image.shape == (60, 160, 3):
return text, convert2gray(image)
else:
# print 'image shape is not expected:', image.shape
continue
for i in range(batch_size):
text, image = loop_get_text_and_image()
batch_x[i, :] = image.flatten() / 255
batch_y[i, :] = text2vec(text)
return batch_x, batch_y
X = tf.placeholder(tf.float32, [None, IMAGE_HEIGHT * IMAGE_WIDTH])
Y = tf.placeholder(tf.float32, [None, CAPTCHA_LEN * CHAR_SET_LEN])
keep_prob = tf.placeholder(tf.float32)
def build_crack_captcha_cnn():
x = tf.reshape(X, shape=[-1, IMAGE_HEIGHT, IMAGE_WIDTH, 1])
# 3 conv layer
conv_1 = slim.conv2d(x, 32, [3, 3], 1, padding='SAME', scope='conv1')
max_pool_1 = slim.max_pool2d(conv_1, [2, 2], [2, 2], padding='SAME')
conv_2 = slim.conv2d(max_pool_1, 64, [3, 3], padding='SAME', scope='conv2')
max_pool_2 = slim.max_pool2d(conv_2, [2, 2], [2, 2], padding='SAME')
conv_3 = slim.conv2d(max_pool_2, 64, [3, 3], padding='SAME', scope='conv3')
max_pool_3 = slim.max_pool2d(conv_3, [2, 2], [2, 2], padding='SAME')
# Fully connected layer
# with tf.variable_scope('layer4-fc1'):
# fc1_weights = tf.get_variable('weight', [8 * 20 * 64, 1024], initializer=tf.truncated_normal_initializer(stddev=0.1))
# # L1
# # tf.add_to_collection('losses', )
# fc1_biases = tf.get_variable('bias', [1024], initializer=tf.constant_initializer(0.1))
# dense = tf.reshape(max_pool_3, [-1, fc1_weights.get_shape().as_list()[0]])
# fc1 = tf.nn.relu(tf.matmul(dense, fc1_weights) + fc1_biases)
# fc1 = tf.nn.dropout(fc1, 0.5)
# with tf.variable_scope('layer5-fc2'):
# fc2_weights = tf.get_variable('weights', [1024, CAPTCHA_LEN * CHAR_SET_LEN])
# fc2_biases = tf.get_variable('bias', [CAPTCHA_LEN * CHAR_SET_LEN], initializer=tf.constant_initializer(0.1))
# logits = tf.matmul(fc1, fc2_weights) + fc2_biases
# return logits
flatten = slim.flatten(max_pool_3)
fc1 = slim.fully_connected(slim.dropout(flatten, keep_prob), 1024, activation_fn=tf.nn.tanh, scope='fc1')
logits = slim.fully_connected(fc1, CAPTCHA_LEN * CHAR_SET_LEN, activation_fn=None, scope='fc2')
return logits
MODEL_SAVED_DIR = 'crack_captcha.saved'
MONITOR_BOARD_DIR = 'monitor_board'
def train_crack_captcha_cnn():
output = build_crack_captcha_cnn()
loss = tf.reduce_mean(tf.nn.sparse_softmax_cross_entropy_with_logits(logits=output, labels=tf.argmax(Y, 1)))
global_step = tf.get_variable("step", [], dtype=tf.int64, initializer=tf.constant_initializer(0), trainable=False)
rate = tf.train.exponential_decay(1e-3, global_step, decay_steps=200, decay_rate=0.97, staircase=True)
optimizer = tf.train.AdamOptimizer(learning_rate=rate).minimize(loss, global_step=global_step)
predict = tf.reshape(output, [-1, CAPTCHA_LEN, CHAR_SET_LEN])
max_idx_p = tf.argmax(predict, 2)
max_idx_l = tf.argmax(tf.reshape(Y, [-1, CAPTCHA_LEN, CHAR_SET_LEN]), 2)
correct_pred = tf.equal(max_idx_p, max_idx_l)
accuracy = tf.reduce_mean(tf.cast(correct_pred, tf.float32))
saver = tf.train.Saver()
ckpt = tf.train.latest_checkpoint(MODEL_SAVED_DIR)
tf.summary.scalar('loss', loss)
tf.summary.scalar('accuracy', accuracy)
merged = tf.summary.merge_all()
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
board_writer = tf.summary.FileWriter(MONITOR_BOARD_DIR, sess.graph)
# TODO: board_writer.close()
if ckpt:
saver.restore(sess, ckpt)
print("restore from the checkpoint {0}".format(ckpt))
start_step = int(ckpt.split('-')[-1])
print start_step, 'from this.'
while True:
batch_x, batch_y = get_next_batch(256)
summary, _, loss_, step = sess.run([merged, optimizer, loss, global_step], feed_dict={X: batch_x, Y: batch_y, keep_prob: 0.75})
board_writer.add_summary(summary, step)
print('global_step:%s, loss:%s' % (step, loss_))
# 每100 step计算一次准确率
if step % 100 == 1:
batch_x_test, batch_y_test = get_next_batch(64)
acc = sess.run(accuracy, feed_dict={X: batch_x_test, Y: batch_y_test, keep_prob: 1.})
print('global_step:%s, acc:%s' % (step, acc))
saver.save(sess, os.path.join(MODEL_SAVED_DIR, 'model'), global_step=global_step)
train_crack_captcha_cnn()