詳解tensorflow訓(xùn)練自己的數(shù)據(jù)集實(shí)現(xiàn)CNN圖像分類(lèi)

更新時(shí)間：2018年02月07日 11:40:38 作者：康小武

本篇文章了tensorflow訓(xùn)練自己的數(shù)據(jù)集實(shí)現(xiàn)CNN圖像分類(lèi)，小編覺(jué)得挺不錯(cuò)的，現(xiàn)在分享給大家，也給大家做個(gè)參考。一起跟隨小編過(guò)來(lái)看看吧

利用卷積神經(jīng)網(wǎng)絡(luò)訓(xùn)練圖像數(shù)據(jù)分為以下幾個(gè)步驟

1.讀取圖片文件
2.產(chǎn)生用于訓(xùn)練的批次
3.定義訓(xùn)練的模型（包括初始化參數(shù)，卷積、池化層等參數(shù)、網(wǎng)絡(luò)）
4.訓(xùn)練

1 讀取圖片文件

def get_files(filename):
 class_train = []
 label_train = []
 for train_class in os.listdir(filename):
  for pic in os.listdir(filename+train_class):
   class_train.append(filename+train_class+'/'+pic)
   label_train.append(train_class)
 temp = np.array([class_train,label_train])
 temp = temp.transpose()
 #shuffle the samples
 np.random.shuffle(temp)
 #after transpose, images is in dimension 0 and label in dimension 1
 image_list = list(temp[:,0])
 label_list = list(temp[:,1])
 label_list = [int(i) for i in label_list]
 #print(label_list)
 return image_list,label_list

這里文件名作為標(biāo)簽，即類(lèi)別（其數(shù)據(jù)類(lèi)型要確定，后面要轉(zhuǎn)為tensor類(lèi)型數(shù)據(jù)）。

然后將image和label轉(zhuǎn)為list格式數(shù)據(jù)，因?yàn)楹筮呌玫降牡囊恍﹖ensorflow函數(shù)接收的是list格式數(shù)據(jù)。

2 產(chǎn)生用于訓(xùn)練的批次

def get_batches(image,label,resize_w,resize_h,batch_size,capacity):
 #convert the list of images and labels to tensor
 image = tf.cast(image,tf.string)
 label = tf.cast(label,tf.int64)
 queue = tf.train.slice_input_producer([image,label])
 label = queue[1]
 image_c = tf.read_file(queue[0])
 image = tf.image.decode_jpeg(image_c,channels = 3)
 #resize
 image = tf.image.resize_image_with_crop_or_pad(image,resize_w,resize_h)
 #(x - mean) / adjusted_stddev
 image = tf.image.per_image_standardization(image)
 
 image_batch,label_batch = tf.train.batch([image,label],
            batch_size = batch_size,
            num_threads = 64,
            capacity = capacity)
 images_batch = tf.cast(image_batch,tf.float32)
 labels_batch = tf.reshape(label_batch,[batch_size])
 return images_batch,labels_batch

首先使用tf.cast轉(zhuǎn)化為tensorflow數(shù)據(jù)格式，使用tf.train.slice_input_producer實(shí)現(xiàn)一個(gè)輸入的隊(duì)列。

label不需要處理，image存儲(chǔ)的是路徑，需要讀取為圖片，接下來(lái)的幾步就是讀取路徑轉(zhuǎn)為圖片，用于訓(xùn)練。

CNN對(duì)圖像大小是敏感的，第10行圖片resize處理為大小一致，12行將其標(biāo)準(zhǔn)化，即減去所有圖片的均值，方便訓(xùn)練。

接下來(lái)使用tf.train.batch函數(shù)產(chǎn)生訓(xùn)練的批次。

最后將產(chǎn)生的批次做數(shù)據(jù)類(lèi)型的轉(zhuǎn)換和shape的處理即可產(chǎn)生用于訓(xùn)練的批次。

3 定義訓(xùn)練的模型

（1）訓(xùn)練參數(shù)的定義及初始化

def init_weights(shape):
 return tf.Variable(tf.random_normal(shape,stddev = 0.01))
#init weights
weights = {
 "w1":init_weights([3,3,3,16]),
 "w2":init_weights([3,3,16,128]),
 "w3":init_weights([3,3,128,256]),
 "w4":init_weights([4096,4096]),
 "wo":init_weights([4096,2])
 }

#init biases
biases = {
 "b1":init_weights([16]),
 "b2":init_weights([128]),
 "b3":init_weights([256]),
 "b4":init_weights([4096]),
 "bo":init_weights([2])
 }

CNN的每層是y=wx+b的決策模型，卷積層產(chǎn)生特征向量，根據(jù)這些特征向量帶入x進(jìn)行計(jì)算，因此，需要定義卷積層的初始化參數(shù)，包括權(quán)重和偏置。其中第8行的參數(shù)形狀后邊再解釋。

（2）定義不同層的操作

 def conv2d(x,w,b):
 x = tf.nn.conv2d(x,w,strides = [1,1,1,1],padding = "SAME")
 x = tf.nn.bias_add(x,b)
 return tf.nn.relu(x)

def pooling(x):
 return tf.nn.max_pool(x,ksize = [1,2,2,1],strides = [1,2,2,1],padding = "SAME")

def norm(x,lsize = 4):
 return tf.nn.lrn(x,depth_radius = lsize,bias = 1,alpha = 0.001/9.0,beta = 0.75)

這里只定義了三種層，即卷積層、池化層和正則化層

（3）定義訓(xùn)練模型

def mmodel(images):
 l1 = conv2d(images,weights["w1"],biases["b1"])
 l2 = pooling(l1)
 l2 = norm(l2)
 l3 = conv2d(l2,weights["w2"],biases["b2"])
 l4 = pooling(l3)
 l4 = norm(l4)
 l5 = conv2d(l4,weights["w3"],biases["b3"])
 #same as the batch size
 l6 = pooling(l5)
 l6 = tf.reshape(l6,[-1,weights["w4"].get_shape().as_list()[0]])
 l7 = tf.nn.relu(tf.matmul(l6,weights["w4"])+biases["b4"])
 soft_max = tf.add(tf.matmul(l7,weights["wo"]),biases["bo"])
 return soft_max

模型比較簡(jiǎn)單，使用三層卷積，第11行使用全連接，需要對(duì)特征向量進(jìn)行reshape，其中l(wèi)6的形狀為[-1，w4的第1維的參數(shù)]，因此，將其按照“w4”reshape的時(shí)候，要使得-1位置的大小為batch_size，這樣，最終再乘以“wo”時(shí)，最終的輸出大小為[batch_size,class_num]

（4）定義評(píng)估量

 def loss(logits,label_batches):
  cross_entropy = tf.nn.sparse_softmax_cross_entropy_with_logits(logits=logits,labels=label_batches)
  cost = tf.reduce_mean(cross_entropy)
  return cost

　　首先定義損失函數(shù)，這是用于訓(xùn)練最小化損失的必需量
 def get_accuracy(logits,labels):
  acc = tf.nn.in_top_k(logits,labels,1)
  acc = tf.cast(acc,tf.float32)
  acc = tf.reduce_mean(acc)
  return acc

評(píng)價(jià)分類(lèi)準(zhǔn)確率的量，訓(xùn)練時(shí)，需要loss值減小，準(zhǔn)確率增加，這樣的訓(xùn)練才是收斂的。

（5）定義訓(xùn)練方式

 def training(loss,lr):
  train_op = tf.train.RMSPropOptimizer(lr,0.9).minimize(loss)
  return train_op

有很多種訓(xùn)練方式，可以自行去官網(wǎng)查看，但是不同的訓(xùn)練方式可能對(duì)應(yīng)前面的參數(shù)定義不一樣，需要另行處理，否則可能報(bào)錯(cuò)。

4 訓(xùn)練

def run_training():
 data_dir = 'C:/Users/wk/Desktop/bky/dataSet/'
 image,label = inputData.get_files(data_dir)
 image_batches,label_batches = inputData.get_batches(image,label,32,32,16,20)
 p = model.mmodel(image_batches)
 cost = model.loss(p,label_batches)
 train_op = model.training(cost,0.001)
 acc = model.get_accuracy(p,label_batches)
 
 sess = tf.Session()
 init = tf.global_variables_initializer()
 sess.run(init)
 
 coord = tf.train.Coordinator()
 threads = tf.train.start_queue_runners(sess = sess,coord = coord)
 
 try:
  for step in np.arange(1000):
   print(step)
   if coord.should_stop():
    break
   _,train_acc,train_loss = sess.run([train_op,acc,cost])
   print("loss:{} accuracy:{}".format(train_loss,train_acc))
 except tf.errors.OutOfRangeError:
  print("Done!!!")
 finally:
  coord.request_stop()
 coord.join(threads)
 sess.close()

神經(jīng)網(wǎng)絡(luò)訓(xùn)練的時(shí)候，我們需要將模型保存下來(lái)，方便后面繼續(xù)訓(xùn)練或者用訓(xùn)練好的模型進(jìn)行測(cè)試。因此，我們需要?jiǎng)?chuàng)建一個(gè)saver保存模型。

def run_training():
 data_dir = 'C:/Users/wk/Desktop/bky/dataSet/'
 log_dir = 'C:/Users/wk/Desktop/bky/log/'
 image,label = inputData.get_files(data_dir)
 image_batches,label_batches = inputData.get_batches(image,label,32,32,16,20)
 print(image_batches.shape)
 p = model.mmodel(image_batches,16)
 cost = model.loss(p,label_batches)
 train_op = model.training(cost,0.001)
 acc = model.get_accuracy(p,label_batches)
 
 sess = tf.Session()
 init = tf.global_variables_initializer()
 sess.run(init)
 saver = tf.train.Saver()
 coord = tf.train.Coordinator()
 threads = tf.train.start_queue_runners(sess = sess,coord = coord)
 
 try:
  for step in np.arange(1000):
   print(step)
   if coord.should_stop():
    break
   _,train_acc,train_loss = sess.run([train_op,acc,cost])
   print("loss:{} accuracy:{}".format(train_loss,train_acc))
   if step % 100 == 0:
    check = os.path.join(log_dir,"model.ckpt")
    saver.save(sess,check,global_step = step)
 except tf.errors.OutOfRangeError:
  print("Done!!!")
 finally:
  coord.request_stop()
 coord.join(threads)
 sess.close()

訓(xùn)練好的模型信息會(huì)記錄在checkpoint文件中，大致如下：

model_checkpoint_path: "C:/Users/wk/Desktop/bky/log/model.ckpt-100"
all_model_checkpoint_paths: "C:/Users/wk/Desktop/bky/log/model.ckpt-0"
all_model_checkpoint_paths: "C:/Users/wk/Desktop/bky/log/model.ckpt-100"

其余還會(huì)生成一些文件，分別記錄了模型參數(shù)等信息，后邊測(cè)試的時(shí)候程序會(huì)讀取checkpoint文件去加載這些真正的數(shù)據(jù)文件

構(gòu)建好神經(jīng)網(wǎng)絡(luò)進(jìn)行訓(xùn)練完成后，如果用之前的代碼直接進(jìn)行測(cè)試，會(huì)報(bào)shape不符合的錯(cuò)誤，大致是卷積層的輸入與圖像的shape不一致，這是因?yàn)樯掀拇a，將weights和biases定義在了模型的外面，調(diào)用模型的時(shí)候，出現(xiàn)valueError的錯(cuò)誤。

因此，我們需要將參數(shù)定義在模型里面，加載訓(xùn)練好的模型參數(shù)時(shí)，訓(xùn)練好的參數(shù)才能夠真正初始化模型。重寫(xiě)模型函數(shù)如下

def mmodel(images,batch_size):
 with tf.variable_scope('conv1') as scope:
  weights = tf.get_variable('weights', 
         shape = [3,3,3, 16],
         dtype = tf.float32, 
         initializer=tf.truncated_normal_initializer(stddev=0.1,dtype=tf.float32))
  biases = tf.get_variable('biases', 
         shape=[16],
         dtype=tf.float32,
         initializer=tf.constant_initializer(0.1))
  conv = tf.nn.conv2d(images, weights, strides=[1,1,1,1], padding='SAME')
  pre_activation = tf.nn.bias_add(conv, biases)
  conv1 = tf.nn.relu(pre_activation, name= scope.name)
 with tf.variable_scope('pooling1_lrn') as scope:
  pool1 = tf.nn.max_pool(conv1, ksize=[1,2,2,1],strides=[1,2,2,1],
        padding='SAME', name='pooling1')
  norm1 = tf.nn.lrn(pool1, depth_radius=4, bias=1.0, alpha=0.001/9.0,
       beta=0.75,name='norm1')
 with tf.variable_scope('conv2') as scope:
  weights = tf.get_variable('weights',
         shape=[3,3,16,128],
         dtype=tf.float32,
         initializer=tf.truncated_normal_initializer(stddev=0.1,dtype=tf.float32))
  biases = tf.get_variable('biases',
         shape=[128], 
         dtype=tf.float32,
         initializer=tf.constant_initializer(0.1))
  conv = tf.nn.conv2d(norm1, weights, strides=[1,1,1,1],padding='SAME')
  pre_activation = tf.nn.bias_add(conv, biases)
  conv2 = tf.nn.relu(pre_activation, name='conv2') 
 with tf.variable_scope('pooling2_lrn') as scope:
  norm2 = tf.nn.lrn(conv2, depth_radius=4, bias=1.0, alpha=0.001/9.0,
       beta=0.75,name='norm2')
  pool2 = tf.nn.max_pool(norm2, ksize=[1,2,2,1], strides=[1,1,1,1],
        padding='SAME',name='pooling2')
 with tf.variable_scope('local3') as scope:
  reshape = tf.reshape(pool2, shape=[batch_size, -1])
  dim = reshape.get_shape()[1].value
  weights = tf.get_variable('weights',
         shape=[dim,4096],
         dtype=tf.float32,
         initializer=tf.truncated_normal_initializer(stddev=0.005,dtype=tf.float32))
  biases = tf.get_variable('biases',
         shape=[4096],
         dtype=tf.float32, 
         initializer=tf.constant_initializer(0.1))
  local3 = tf.nn.relu(tf.matmul(reshape, weights) + biases, name=scope.name) 
 with tf.variable_scope('softmax_linear') as scope:
  weights = tf.get_variable('softmax_linear',
         shape=[4096, 2],
         dtype=tf.float32,
         initializer=tf.truncated_normal_initializer(stddev=0.005,dtype=tf.float32))
  biases = tf.get_variable('biases', 
         shape=[2],
         dtype=tf.float32, 
         initializer=tf.constant_initializer(0.1))
  softmax_linear = tf.add(tf.matmul(local3, weights), biases, name='softmax_linear')
 return softmax_linear

測(cè)試訓(xùn)練好的模型

首先獲取一張測(cè)試圖像

 def get_one_image(img_dir):
  image = Image.open(img_dir)
  plt.imshow(image)
  image = image.resize([32, 32])
  image_arr = np.array(image)
  return image_arr

加載模型，計(jì)算測(cè)試結(jié)果

def test(test_file):
 log_dir = 'C:/Users/wk/Desktop/bky/log/'
 image_arr = get_one_image(test_file)
 
 with tf.Graph().as_default():
  image = tf.cast(image_arr, tf.float32)
  image = tf.image.per_image_standardization(image)
  image = tf.reshape(image, [1,32, 32, 3])
  print(image.shape)
  p = model.mmodel(image,1)
  logits = tf.nn.softmax(p)
  x = tf.placeholder(tf.float32,shape = [32,32,3])
  saver = tf.train.Saver()
  with tf.Session() as sess:
   ckpt = tf.train.get_checkpoint_state(log_dir)
   if ckpt and ckpt.model_checkpoint_path:
    global_step = ckpt.model_checkpoint_path.split('/')[-1].split('-')[-1]
    saver.restore(sess, ckpt.model_checkpoint_path)
    print('Loading success)
   else:
    print('No checkpoint')
   prediction = sess.run(logits, feed_dict={x: image_arr})
   max_index = np.argmax(prediction)
   print(max_index)

前面主要是將測(cè)試圖片標(biāo)準(zhǔn)化為網(wǎng)絡(luò)的輸入圖像，15-19是加載模型文件，然后將圖像輸入到模型里即可

以上就是本文的全部?jī)?nèi)容，希望對(duì)大家的學(xué)習(xí)有所幫助，也希望大家多多支持腳本之家。

您可能感興趣的文章: