上一篇文章tensorflow 实现端到端的OCR:二代身份证号识别实现了定长18位数字串的识别,并最终达到了98%的准确率。
但是实际应用场景中,常常需要面对无法确定字串长度的情况,这时候除了需要对识别字符模型参数进行训练外,还需要对字符划分模型进行训练,本文实现了上文提到的方法2,使用LSTM+CTC识别不定长的数字串。
环境依赖
环境依赖与上一篇基本一致
知识准备
-
LSTM(长短时记忆网络):是一种特殊结构的RNN,能够解决普通RNN不能解决的长期依赖问题。具体介绍可参看这篇译文[译] 理解 LSTM 网络
-
CTC :Connectionist Temporal Classifier 一般译为联结主义时间分类器 ,适合于输入特征和输出标签之间对齐关系不确定的时间序列问题,CTC可以自动端到端地同时优化模型参数和对齐切分的边界。
比如本文例子,32 x 256大小的图片,最大可切分256列,也就是输入特征最大256,而输出标签的长度最大设定是18,这种就可以用CTC模型进行优化。
关于CTC模型,笔者认为可以这样理解,假设32 x 256的图片,数字串标签是"123",把图片按列切分(CTC会优化切分模型),然后分出来的每块再去识别数字,找出这块是每个数字或者特殊字符的概率(无法识别的则标记为特殊字符"-"),这样就得到了基于输入特征序列(图片)的每一个相互独立建模单元个体(划分出来的块)(包括“-”节点在内)的类属概率分布。基于概率分布,算出标签序列是"123"的概率P(123),当然这里设定"123"的概率为所有子序列之和,这里子序列包括'-'和'1'、'2'、'3'的连续重复,如下图所示:
所有子序列概率和
本文采用TF框架的CTC封装实现,tf.nn.ctc_loss,我们最后的目标是最小化ctc_loss
官方定义如下:
ctc_loss(
labels,
inputs,
sequence_length,
preprocess_collapse_repeated=False,
ctc_merge_repeated=True,
time_major=True
)
inputs: 输入(训练)数据,是一个三维float型的数据结构[max_time_step , batch_size , num_classes],当修改time_major = False时,[batch_size,max_time_step,num_classes]。
总体的数据流:
image_batch
->[batch_size,max_time_step,num_features]->lstm
->[batch_size,max_time_step,cell.output_size]->reshape
->[batch_sizemax_time_step,num_hidden]->affine projection AW+b
->[batch_size*max_time_step,num_classes]->reshape
->[batch_size,max_time_step,num_classes]->transpose
->[max_time_step,batch_size,num_classes]
本文输入图片大小是(32,256),则num_features是32,max_time_step是256 代表最大划分序列,其中cell.output_size == num_hidden,num_hidden及num_classes的值见下文常量定义
labels:OCR识别结果的标签,是一个稀疏矩阵,下文训练数据生成部分会有相关解释
sequence_length: 一维数据,[max_time_step,…,max_time_step]长度为batch_size,值为max_time_step
因此我们需要做的就是将图片的标签label(需要OCR出的结果),图片数据,以及图片的长度转换为labels,inputs,和sequence_length
正文
定义一些常量
#定义一些常量
#图片大小,32 x 256
OUTPUT_SHAPE = (32,256)
#训练最大轮次
num_epochs = 10000
#LSTM
num_hidden = 64
num_layers = 1
obj = gen_id_card()
num_classes = obj.len + 1 + 1 # 10位数字 + blank + ctc blank
#初始化学习速率
INITIAL_LEARNING_RATE = 1e-3
DECAY_STEPS = 5000
REPORT_STEPS = 100
LEARNING_RATE_DECAY_FACTOR = 0.9 # The learning rate decay factor
MOMENTUM = 0.9
DIGITS='0123456789'
BATCHES = 10
BATCH_SIZE = 64
TRAIN_SIZE = BATCHES * BATCH_SIZE
训练数据集生成
训练数据集的生成基本与上文一致,唯一变化就是增加生成随机长度串的选项,对应方法如下:
def gen_text(self, is_ran=False):
text = ''
vecs = np.zeros((self.max_size * self.len))
//唯一变化,随机设定长度
if is_ran == True:
size = random.randint(1, self.max_size)
else:
size = self.max_size
for i in range(size):
c = random.choice(self.char_set)
vec = self.char2vec(c)
text = text + c
vecs[i*self.len:(i+1)*self.len] = np.copy(vec)
return text,vecs
# 生成一个训练batch
def get_next_batch(batch_size=128):
obj = gen_id_card()
#(batch_size,256,32)
inputs = np.zeros([batch_size, OUTPUT_SHAPE[1],OUTPUT_SHAPE[0]])
codes = []
for i in range(batch_size):
#生成不定长度的字串
image, text, vec = obj.gen_image(True)
#np.transpose 矩阵转置 (32*256,) => (32,256) => (256,32)
inputs[i,:] = np.transpose(image.reshape((OUTPUT_SHAPE[0],OUTPUT_SHAPE[1])))
#标签转成列表保存在codes
codes.append(list(text))
#比如batch_size=2,两条数据分别是"12"和"1",则targets [['1','2'],['1']]
targets = [np.asarray(i) for i in codes]
#targets转成稀疏矩阵
sparse_targets = sparse_tuple_from(targets)
#(batch_size,) sequence_length值都是256,最大划分列数
seq_len = np.ones(inputs.shape[0]) * OUTPUT_SHAPE[1]
return inputs, sparse_targets, seq_len
这里我们来了解一下什么是稀疏矩阵,下面是百度百科的定义
对于那些零元素数目远远多于非零元素数目,并且非零元素的分布没有规律的矩阵称为稀疏矩阵(sparse)
其实很容易理解这里为什么OCR识别训练的标签labels是一个稀疏矩阵,假设我们生成的batch_size 是64的样本,每个样本是长度为1~18的数字串,则生成一个(64,18)的矩阵,矩阵有数字的是非零元素,无数字的是零元素,且因为这个标签是不定长的,所以非零元素的分布没有规律,标签在存储数字串的同时还要存储位置信息。
我们来看一下tensorflow中,如何把targets转成一个稀疏矩阵的
#转化一个序列列表为稀疏矩阵
def sparse_tuple_from(sequences, dtype=np.int32):
"""
Create a sparse representention of x.
Args:
sequences: a list of lists of type dtype where each element is a sequence
Returns:
A tuple with (indices, values, shape)
"""
indices = []
values = []
for n, seq in enumerate(sequences):
indices.extend(zip([n] * len(seq), xrange(len(seq))))
values.extend(seq)
indices = np.asarray(indices, dtype=np.int64)
values = np.asarray(values, dtype=dtype)
shape = np.asarray([len(sequences), np.asarray(indices).max(0)[1] + 1], dtype=np.int64)
return indices, values, shape
indices:二维int64的矩阵,代表非0的坐标点
values:二维tensor,代表indice位置的数据值
dense_shape:一维,代表稀疏矩阵的大小
仍然拿刚才的两个串"12"和"1"做例子,转成的稀疏矩阵应该是
indecs = [[0,0],[0,1],[1,0]]
values = [1,2,1]
dense_shape = [2,2] (两个数字串,最大长度为2)
代表dense tensor:
[[1,2],[1,0]]
有了序列列表转稀疏矩阵的方法,反过来,当然也需要稀疏矩阵转序列列表的方法:
def decode_sparse_tensor(sparse_tensor):
decoded_indexes = list()
current_i = 0
current_seq = []
for offset, i_and_index in enumerate(sparse_tensor[0]):
i = i_and_index[0]
if i != current_i:
decoded_indexes.append(current_seq)
current_i = i
current_seq = list()
current_seq.append(offset)
decoded_indexes.append(current_seq)
result = []
for index in decoded_indexes:
result.append(decode_a_seq(index, sparse_tensor))
return result
def decode_a_seq(indexes, spars_tensor):
decoded = []
for m in indexes:
str = DIGITS[spars_tensor[1][m]]
decoded.append(str)
return decoded
构建网络,开始训练
数据准备工作完成,则开始构建LSTM+CTC的训练模型,其中TF实现LSTM的方法就不做过多解释,请读者自行百度。
def get_train_model():
inputs = tf.placeholder(tf.float32, [None, None, OUTPUT_SHAPE[0]])
#定义ctc_loss需要的稀疏矩阵
targets = tf.sparse_placeholder(tf.int32)
#1维向量 序列长度 [batch_size,]
seq_len = tf.placeholder(tf.int32, [None])
#定义LSTM网络
cell = tf.contrib.rnn.LSTMCell(num_hidden, state_is_tuple=True)
stack = tf.contrib.rnn.MultiRNNCell([cell] * num_layers, state_is_tuple=True)
outputs, _ = tf.nn.dynamic_rnn(cell, inputs, seq_len, dtype=tf.float32)
shape = tf.shape(inputs)
#[batch_size,256]
batch_s, max_timesteps = shape[0], shape[1]
#[batch_size*max_time_step,num_hidden]
outputs = tf.reshape(outputs, [-1, num_hidden])
W = tf.Variable(tf.truncated_normal([num_hidden,
num_classes],
stddev=0.1), name="W")
b = tf.Variable(tf.constant(0., shape=[num_classes]), name="b")
#[batch_size*max_timesteps,num_classes]
logits = tf.matmul(outputs, W) + b
#[batch_size,max_timesteps,num_classes]
logits = tf.reshape(logits, [batch_s, -1, num_classes])
#转置矩阵,第0和第1列互换位置=>[max_timesteps,batch_size,num_classes]
logits = tf.transpose(logits, (1, 0, 2))
return logits, inputs, targets, seq_len, W, b
训练模型
def train():
global_step = tf.Variable(0, trainable=False)
learning_rate = tf.train.exponential_decay(INITIAL_LEARNING_RATE,
global_step,
DECAY_STEPS,
LEARNING_RATE_DECAY_FACTOR,
staircase=True)
logits, inputs, targets, seq_len, W, b = get_train_model()
#tragets是一个稀疏矩阵
loss = tf.nn.ctc_loss(labels=targets,inputs=logits, sequence_length=seq_len)
cost = tf.reduce_mean(loss)
#optimizer = tf.train.MomentumOptimizer(learning_rate=learning_rate,momentum=MOMENTUM).minimize(cost, global_step=global_step)
optimizer = tf.train.AdamOptimizer(learning_rate=learning_rate).minimize(loss,global_step=global_step)
#前面说的划分块之后找每块的类属概率分布,ctc_beam_search_decoder方法,是每次找最大的K个概率分布
#还有一种贪心策略是只找概率最大那个,也就是K=1的情况ctc_ greedy_decoder
decoded, log_prob = tf.nn.ctc_beam_search_decoder(logits, seq_len, merge_repeated=False)
acc = tf.reduce_mean(tf.edit_distance(tf.cast(decoded[0], tf.int32), targets))
init = tf.global_variables_initializer()
def report_accuracy(decoded_list, test_targets):
original_list = decode_sparse_tensor(test_targets)
detected_list = decode_sparse_tensor(decoded_list)
true_numer = 0
if len(original_list) != len(detected_list):
print("len(original_list)", len(original_list), "len(detected_list)", len(detected_list),
" test and detect length desn't match")
return
print("T/F: original(length) <-------> detectcted(length)")
for idx, number in enumerate(original_list):
detect_number = detected_list[idx]
hit = (number == detect_number)
print(hit, number, "(", len(number), ") <-------> ", detect_number, "(", len(detect_number), ")")
if hit:
true_numer = true_numer + 1
print("Test Accuracy:", true_numer * 1.0 / len(original_list))
def do_report():
test_inputs,test_targets,test_seq_len = get_next_batch(BATCH_SIZE)
test_feed = {inputs: test_inputs,
targets: test_targets,
seq_len: test_seq_len}
dd, log_probs, accuracy = session.run([decoded[0], log_prob, acc], test_feed)
report_accuracy(dd, test_targets)
def do_batch():
train_inputs, train_targets, train_seq_len = get_next_batch(BATCH_SIZE)
feed = {inputs: train_inputs, targets: train_targets, seq_len: train_seq_len}
b_loss,b_targets, b_logits, b_seq_len,b_cost, steps, _ = session.run([loss, targets, logits, seq_len, cost, global_step, optimizer], feed)
print b_cost, steps
if steps > 0 and steps % REPORT_STEPS == 0:
do_report()
save_path = saver.save(session, "ocr.model", global_step=steps)
return b_cost, steps
with tf.Session() as session:
session.run(init)
saver = tf.train.Saver(tf.global_variables(), max_to_keep=100)
for curr_epoch in xrange(num_epochs):
print("Epoch.......", curr_epoch)
train_cost = train_ler = 0
for batch in xrange(BATCHES):
start = time.time()
c, steps = do_batch()
train_cost += c * BATCH_SIZE
seconds = time.time() - start
print("Step:", steps, ", batch seconds:", seconds)
train_cost /= TRAIN_SIZE
train_inputs, train_targets, train_seq_len = get_next_batch(BATCH_SIZE)
val_feed = {inputs: train_inputs,
targets: train_targets,
seq_len: train_seq_len}
val_cost, val_ler, lr, steps = session.run([cost, acc, learning_rate, global_step], feed_dict=val_feed)
log = "Epoch {}/{}, steps = {}, train_cost = {:.3f}, train_ler = {:.3f}, val_cost = {:.3f}, val_ler = {:.3f}, time = {:.3f}s, learning_rate = {}"
print(log.format(curr_epoch + 1, num_epochs, steps, train_cost, train_ler, val_cost, val_ler, time.time() - start, lr))
训练结果
训练到第80个epoch的时候,64个测试样本的准确率达到64%
训练第80次后测试结果
训练到第100个epoch的时候,64个测试样本的准确率达到100%了,后续基本上准确率都是100%了
测试准确率达到100%
后记
最后完整的代码托管在我的Github下
训练产生的图片数据属于比较理想,无噪音的环境下,所以才会在100个epoch的时候准确率就达到100%了,实际应用中图片可能会有些线段或者离散点的噪音,读者可以自行在生成训练集中增加一些噪音,测试模型训练效果
本文生成的串的所属类别仅是0~9的10个类别,如果后续加上26*2个大小写英文字母,或者加上3500+常用中文汉字去组成串,随着所属类别的不断扩大,模型还能不能很好的识别?模型收敛的速度如何?
在编写本文示例代码的过程中,较多的参考了别人的代码和模型,很多东西深层的原理基本没去了解。先做下记录,后续再研究那些数学模型和推导公式之类的。
参考链接
[译] 理解 LSTM 网络
tensorflow_lstm_ctc_ocr
tensorflow LSTM+CTC/warpCTC使用详解