本文转自WTF Daily Blog斗大的熊猫的##
TensorFlow练习20: 使用深度学习破解字符验证码
验证码是根据随机字符生成一幅图片,然后在图片中加入干扰象素,用户必须手动填入,防止有人利用机器人自动批量注册、灌水、发垃圾广告等等 。
验证码的作用是验证用户是真人还是机器人;设计理念是对人友好,对机器难。
上图是常见的字符验证码,还有一些验证码使用提问的方式。
我们先来看看破解验证码的几种方式:
人力打码(基本上,打码任务都是大型网站的验证码,用于自动化注册等等)
找到能过验证码的漏洞
最后一种是字符识别,这是本帖的关注点
我上网查了查,用Tesseract OCR、OpenCV等等其它方法都需把验证码分割为单个字符,然后识别单个字符。分割验证码可是人的强项,如果字符之间相互重叠,那机器就不容易分割了。
本帖实现的方法不需要分割验证码,而是把验证码做为一个整体进行识别。
相关论文:
Multi-digit Number Recognition from Street View Imagery using Deep CNN
CAPTCHA Recognition with Active Deep Learning
http://matthewearl.github.io/2016/05/06/cnn-anpr/
使用深度学习+训练数据+大量计算力,我们可以在几天内训练一个可以破解验证码的模型,当然前提是获得大量训练数据。
获得训练数据方法:
手动(累死人系列)
破解验证码生成机制,自动生成无限多的训练数据
打入敌人内部(卧底+不要脸+不要命+多大仇系列)
我自己做一个验证码生成器,然后训练CNN模型破解自己做的验证码生成器。
我觉的验证码机制可以废了,单纯的增加验证码难度只会让人更难识别,使用CNN+RNN,机器的识别准确率不比人差。Google已经意识到了这一点,他们现在使用机器学习技术检测异常流量。
验证码生成器
from captcha.image import ImageCaptcha # pip install captcha
import numpy as np
import matplotlib.pyplot as plt
from PIL import Image
import random
# 验证码中的字符, 就不用汉字了
number = ['0','1','2','3','4','5','6','7','8','9']
alphabet = ['a','b','c','d','e','f','g','h','i','j','k','l','m','n','o','p','q','r','s','t','u','v','w','x','y','z']
ALPHABET = ['A','B','C','D','E','F','G','H','I','J','K','L','M','N','O','P','Q','R','S','T','U','V','W','X','Y','Z']
# 验证码一般都无视大小写;验证码长度4个字符
def random_captcha_text(char_set=number+alphabet+ALPHABET, captcha_size=4):
captcha_text = []
for i in range(captcha_size):
c = random.choice(char_set)
captcha_text.append(c)
return captcha_text
# 生成字符对应的验证码
def gen_captcha_text_and_image():
image = ImageCaptcha()
captcha_text = random_captcha_text()
captcha_text = ''.join(captcha_text)
captcha = image.generate(captcha_text)
#image.write(captcha_text, captcha_text + '.jpg') # 写到文件
captcha_image = Image.open(captcha)
captcha_image = np.array(captcha_image)
return captcha_text, captcha_image
if __name__ == '__main__':
# 测试
text, image = gen_captcha_text_and_image()
f = plt.figure()
ax = f.add_subplot(111)
ax.text(0.1, 0.9,text, ha='center', va='center', transform=ax.transAxes)
plt.imshow(image)
plt.show()
左上角文本对应验证码图像
训练
from gen_captcha import gen_captcha_text_and_image
from gen_captcha import number
from gen_captcha import alphabet
from gen_captcha import ALPHABET
import numpy as np
import tensorflow as tf
text, image = gen_captcha_text_and_image()
print("验证码图像channel:", image.shape) # (60, 160, 3)
# 图像大小
IMAGE_HEIGHT = 60
IMAGE_WIDTH = 160
MAX_CAPTCHA = len(text)
print("验证码文本最长字符数", MAX_CAPTCHA) # 验证码最长4字符; 我全部固定为4,可以不固定. 如果验证码长度小于4,用'_'补齐
# 把彩色图像转为灰度图像(色彩对识别验证码没有什么用)
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
"""
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
"""
# 文本转向量
char_set = number + alphabet + ALPHABET + ['_'] # 如果验证码长度小于4, '_'用来补齐
CHAR_SET_LEN = len(char_set)
def text2vec(text):
text_len = len(text)
if text_len > MAX_CAPTCHA:
raise ValueError('验证码最长4个字符')
vector = np.zeros(MAX_CAPTCHA*CHAR_SET_LEN)
def char2pos(c):
if c =='_':
k = 62
return k
k = ord(c)-48
if k > 9:
k = ord(c) - 55
if k > 35:
k = ord(c) - 61
if k > 61:
raise ValueError('No Map')
return k
for i, c in enumerate(text):
idx = i * CHAR_SET_LEN + char2pos(c)
vector[idx] = 1
return vector
# 向量转回文本
def vec2text(vec):
char_pos = vec.nonzero()[0]
text=[]
for i, c in enumerate(char_pos):
char_at_pos = i #c/63
char_idx = c % CHAR_SET_LEN
if char_idx < 10:
char_code = char_idx + ord('0')
elif char_idx <36:
char_code = char_idx - 10 + ord('A')
elif char_idx < 62:
char_code = char_idx- 36 + ord('a')
elif char_idx == 62:
char_code = ord('_')
else:
raise ValueError('error')
text.append(chr(char_code))
return "".join(text)
"""
#向量(大小MAX_CAPTCHA*CHAR_SET_LEN)用0,1编码 每63个编码一个字符,这样顺利有,字符也有
vec = text2vec("F5Sd")
text = vec2text(vec)
print(text) # F5Sd
vec = text2vec("SFd5")
text = vec2text(vec)
print(text) # SFd5
"""
# 生成一个训练batch
def get_next_batch(batch_size=128):
batch_x = np.zeros([batch_size, IMAGE_HEIGHT*IMAGE_WIDTH])
batch_y = np.zeros([batch_size, MAX_CAPTCHA*CHAR_SET_LEN])
# 有时生成图像大小不是(60, 160, 3)
def wrap_gen_captcha_text_and_image():
while True:
text, image = gen_captcha_text_and_image()
if image.shape == (60, 160, 3):
return text, image
for i in range(batch_size):
text, image = wrap_gen_captcha_text_and_image()
image = convert2gray(image)
batch_x[i,:] = image.flatten() / 255 # (image.flatten()-128)/128 mean为0
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, MAX_CAPTCHA*CHAR_SET_LEN])
keep_prob = tf.placeholder(tf.float32) # dropout
# 定义CNN
def crack_captcha_cnn(w_alpha=0.01, b_alpha=0.1):
x = tf.reshape(X, shape=[-1, IMAGE_HEIGHT, IMAGE_WIDTH, 1])
#w_c1_alpha = np.sqrt(2.0/(IMAGE_HEIGHT*IMAGE_WIDTH)) #
#w_c2_alpha = np.sqrt(2.0/(3*3*32))
#w_c3_alpha = np.sqrt(2.0/(3*3*64))
#w_d1_alpha = np.sqrt(2.0/(8*32*64))
#out_alpha = np.sqrt(2.0/1024)
# 3 conv layer
w_c1 = tf.Variable(w_alpha*tf.random_normal([3, 3, 1, 32]))
b_c1 = tf.Variable(b_alpha*tf.random_normal([32]))
conv1 = tf.nn.relu(tf.nn.bias_add(tf.nn.conv2d(x, w_c1, strides=[1, 1, 1, 1], padding='SAME'), b_c1))
conv1 = tf.nn.max_pool(conv1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
conv1 = tf.nn.dropout(conv1, keep_prob)
w_c2 = tf.Variable(w_alpha*tf.random_normal([3, 3, 32, 64]))
b_c2 = tf.Variable(b_alpha*tf.random_normal([64]))
conv2 = tf.nn.relu(tf.nn.bias_add(tf.nn.conv2d(conv1, w_c2, strides=[1, 1, 1, 1], padding='SAME'), b_c2))
conv2 = tf.nn.max_pool(conv2, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
conv2 = tf.nn.dropout(conv2, keep_prob)
w_c3 = tf.Variable(w_alpha*tf.random_normal([3, 3, 64, 64]))
b_c3 = tf.Variable(b_alpha*tf.random_normal([64]))
conv3 = tf.nn.relu(tf.nn.bias_add(tf.nn.conv2d(conv2, w_c3, strides=[1, 1, 1, 1], padding='SAME'), b_c3))
conv3 = tf.nn.max_pool(conv3, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
conv3 = tf.nn.dropout(conv3, keep_prob)
# Fully connected layer
w_d = tf.Variable(w_alpha*tf.random_normal([8*20*64, 1024]))
b_d = tf.Variable(b_alpha*tf.random_normal([1024]))
dense = tf.reshape(conv3, [-1, w_d.get_shape().as_list()[0]])
dense = tf.nn.relu(tf.add(tf.matmul(dense, w_d), b_d))
dense = tf.nn.dropout(dense, keep_prob)
w_out = tf.Variable(w_alpha*tf.random_normal([1024, MAX_CAPTCHA*CHAR_SET_LEN]))
b_out = tf.Variable(b_alpha*tf.random_normal([MAX_CAPTCHA*CHAR_SET_LEN]))
out = tf.add(tf.matmul(dense, w_out), b_out)
#out = tf.nn.softmax(out)
return out
# 训练
def train_crack_captcha_cnn():
output = crack_captcha_cnn()
# loss
#loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(output, Y))
loss = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=output, labels=Y))
# 最后一层用来分类的softmax和sigmoid有什么不同?
# optimizer 为了加快训练 learning_rate应该开始大,然后慢慢衰
optimizer = tf.train.AdamOptimizer(learning_rate=0.001).minimize(loss)
predict = tf.reshape(output, [-1, MAX_CAPTCHA, CHAR_SET_LEN])
max_idx_p = tf.argmax(predict, 2)
max_idx_l = tf.argmax(tf.reshape(Y, [-1, MAX_CAPTCHA, 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()
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
step = 0
while True:
batch_x, batch_y = get_next_batch(64)
_, loss_ = sess.run([optimizer, loss], feed_dict={X: batch_x, Y: batch_y, keep_prob: 0.75})
print(step, loss_)
# 每100 step计算一次准确率
if step % 100 == 0:
batch_x_test, batch_y_test = get_next_batch(100)
acc = sess.run(accuracy, feed_dict={X: batch_x_test, Y: batch_y_test, keep_prob: 1.})
print(step, acc)
# 如果准确率大于50%,保存模型,完成训练
if acc > 0.5:
saver.save(sess, "crack_capcha.model", global_step=step)
break
step += 1
train_crack_captcha_cnn()
CNN需要大量的样本进行训练,由于时间和资源有限,测试时我只使用数字做为验证码字符集。如果使用数字+大小写字母CNN网络有462个输出,只使用数字CNN网络有410个输出。
TensorBoard是个好东西,既能用来调试也能帮助理解Graph。
训练完成时的准确率(超过50%我就不训练了):
使用训练的模型识别验证码:
def crack_captcha(captcha_image):
output = crack_captcha_cnn()
saver = tf.train.Saver()
with tf.Session() as sess:
saver.restore(sess, tf.train.latest_checkpoint('.'))
predict = tf.argmax(tf.reshape(output, [-1, MAX_CAPTCHA, CHAR_SET_LEN]), 2)
text_list = sess.run(predict, feed_dict={X: [captcha_image], keep_prob: 1})
text = text_list[0].tolist()
vector = np.zeros(MAX_CAPTCHA*CHAR_SET_LEN)
i = 0
for n in text:
vector[i*CHAR_SET_LEN + n] = 1
i += 1
return vec2text(vector)
text, image = gen_captcha_text_and_image()
image = convert2gray(image)
image = image.flatten() / 255
predict_text = crack_captcha(image)
print("正确: {} 预测: {}".format(text, predict_text))
loss和准确率曲线:
为了成为真正的码农,本熊猫要开始研习TensorFlow源代码了,应该能学到不少玩意。
如要转载,请保持本文完整,并注明作者@斗大的熊猫和本文原始地址: http://blog.topspeedsnail.com/archives/10858