正文

73 - Image Segmentation using U-Net - Part1 -What is U-net-

卷积神经网络的架构：

png

常见的卷积神经网络：

png

U-Net：

png

VGG 示意图及其代码：

png

U-Net architecture is designed for Semantic segmentation（语义分割）

png

74 - Image Segmentation using U-Net - Part 2 -Defining U-Net in Python using Keras-

1
2
3
4
5
import tensorflow as tf

IMG_WIDTH = 128
IMG_HEIGHT = 128
IMG_CHANNELS = 3

建立模型

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
inputs = tf.keras.layers.Input((IMG_HEIGHT, IMG_WIDTH, IMG_CHANNELS))
s = tf.keras.layers.Lambda(lambda x: x / 255)(inputs)  # 将整数转换成浮点数

# Contraction path 收缩路径
c1 = tf.keras.layers.Conv2D(16, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(s)
c1 = tf.keras.layers.Dropout(0.1)(c1)
c1 = tf.keras.layers.Conv2D(16, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c1)
p1 = tf.keras.layers.MaxPooling2D((2, 2))(c1)

c2 = tf.keras.layers.Conv2D(32, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(p1)
c2 = tf.keras.layers.Dropout(0.1)(c2)
c2 = tf.keras.layers.Conv2D(32, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c2)
p2 = tf.keras.layers.MaxPooling2D((2, 2))(c2)
 
c3 = tf.keras.layers.Conv2D(64, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(p2)
c3 = tf.keras.layers.Dropout(0.2)(c3)
c3 = tf.keras.layers.Conv2D(64, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c3)
p3 = tf.keras.layers.MaxPooling2D((2, 2))(c3)
 
c4 = tf.keras.layers.Conv2D(128, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(p3)
c4 = tf.keras.layers.Dropout(0.2)(c4)
c4 = tf.keras.layers.Conv2D(128, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c4)
p4 = tf.keras.layers.MaxPooling2D(pool_size=(2, 2))(c4)
 
c5 = tf.keras.layers.Conv2D(256, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(p4)
c5 = tf.keras.layers.Dropout(0.3)(c5)
c5 = tf.keras.layers.Conv2D(256, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c5)

# Expansive path 延展路径
u6 = tf.keras.layers.Conv2DTranspose(128, (2, 2), strides=(2, 2), padding='same')(c5)
u6 = tf.keras.layers.concatenate([u6, c4])
c6 = tf.keras.layers.Conv2D(128, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(u6)
c6 = tf.keras.layers.Dropout(0.2)(c6)
c6 = tf.keras.layers.Conv2D(128, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c6)
 
u7 = tf.keras.layers.Conv2DTranspose(64, (2, 2), strides=(2, 2), padding='same')(c6)
u7 = tf.keras.layers.concatenate([u7, c3])
c7 = tf.keras.layers.Conv2D(64, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(u7)
c7 = tf.keras.layers.Dropout(0.2)(c7)
c7 = tf.keras.layers.Conv2D(64, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c7)
 
u8 = tf.keras.layers.Conv2DTranspose(32, (2, 2), strides=(2, 2), padding='same')(c7)
u8 = tf.keras.layers.concatenate([u8, c2])
c8 = tf.keras.layers.Conv2D(32, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(u8)
c8 = tf.keras.layers.Dropout(0.1)(c8)
c8 = tf.keras.layers.Conv2D(32, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c8)
 
u9 = tf.keras.layers.Conv2DTranspose(16, (2, 2), strides=(2, 2), padding='same')(c8)
u9 = tf.keras.layers.concatenate([u9, c1], axis=3)
c9 = tf.keras.layers.Conv2D(16, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(u9)
c9 = tf.keras.layers.Dropout(0.1)(c9)
c9 = tf.keras.layers.Conv2D(16, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c9)
 
outputs = tf.keras.layers.Conv2D(1, (1, 1), activation='sigmoid')(c9)
 
model = tf.keras.Model(inputs=[inputs], outputs=[outputs])
model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])
model.summary()
Model: "model"
__________________________________________________________________________________________________
 Layer (type)                   Output Shape         Param #     Connected to                     
==================================================================================================
 input_2 (InputLayer)           [(None, 128, 128, 3  0           []                               
                                )]                                                                
                                                                                                  
 lambda (Lambda)                (None, 128, 128, 3)  0           ['input_2[0][0]']                
                                                                                                  
 conv2d (Conv2D)                (None, 128, 128, 16  448         ['lambda[0][0]']                 
                                )                                                                 
                                                                                                  
 dropout (Dropout)              (None, 128, 128, 16  0           ['conv2d[0][0]']                 
                                )                                                                 
                                                                                                  
 conv2d_1 (Conv2D)              (None, 128, 128, 16  2320        ['dropout[0][0]']                
                                )                                                                 
                                                                                                  
 max_pooling2d (MaxPooling2D)   (None, 64, 64, 16)   0           ['conv2d_1[0][0]']               
                                                                                                  
 conv2d_2 (Conv2D)              (None, 64, 64, 32)   4640        ['max_pooling2d[0][0]']          
                                                                                                  
 dropout_1 (Dropout)            (None, 64, 64, 32)   0           ['conv2d_2[0][0]']               
                                                                                                  
 conv2d_3 (Conv2D)              (None, 64, 64, 32)   9248        ['dropout_1[0][0]']              
                                                                                                  
 max_pooling2d_1 (MaxPooling2D)  (None, 32, 32, 32)  0           ['conv2d_3[0][0]']               
                                                                                                  
 conv2d_4 (Conv2D)              (None, 32, 32, 64)   18496       ['max_pooling2d_1[0][0]']        
                                                                                                  
 dropout_2 (Dropout)            (None, 32, 32, 64)   0           ['conv2d_4[0][0]']               
                                                                                                  
 conv2d_5 (Conv2D)              (None, 32, 32, 64)   36928       ['dropout_2[0][0]']              
                                                                                                  
 max_pooling2d_2 (MaxPooling2D)  (None, 16, 16, 64)  0           ['conv2d_5[0][0]']               
                                                                                                  
 conv2d_6 (Conv2D)              (None, 16, 16, 128)  73856       ['max_pooling2d_2[0][0]']        
                                                                                                  
 dropout_3 (Dropout)            (None, 16, 16, 128)  0           ['conv2d_6[0][0]']               
                                                                                                  
 conv2d_7 (Conv2D)              (None, 16, 16, 128)  147584      ['dropout_3[0][0]']              
                                                                                                  
 max_pooling2d_3 (MaxPooling2D)  (None, 8, 8, 128)   0           ['conv2d_7[0][0]']               
                                                                                                  
 conv2d_8 (Conv2D)              (None, 8, 8, 256)    295168      ['max_pooling2d_3[0][0]']        
                                                                                                  
 dropout_4 (Dropout)            (None, 8, 8, 256)    0           ['conv2d_8[0][0]']               
                                                                                                  
 conv2d_9 (Conv2D)              (None, 8, 8, 256)    590080      ['dropout_4[0][0]']              
                                                                                                  
 conv2d_transpose (Conv2DTransp  (None, 16, 16, 128)  131200     ['conv2d_9[0][0]']               
 ose)                                                                                             
                                                                                                  
 concatenate (Concatenate)      (None, 16, 16, 256)  0           ['conv2d_transpose[0][0]',       
                                                                  'conv2d_7[0][0]']               
                                                                                                  
 conv2d_10 (Conv2D)             (None, 16, 16, 128)  295040      ['concatenate[0][0]']            
                                                                                                  
 dropout_5 (Dropout)            (None, 16, 16, 128)  0           ['conv2d_10[0][0]']              
                                                                                                  
 conv2d_11 (Conv2D)             (None, 16, 16, 128)  147584      ['dropout_5[0][0]']              
                                                                                                  
 conv2d_transpose_1 (Conv2DTran  (None, 32, 32, 64)  32832       ['conv2d_11[0][0]']              
 spose)                                                                                           
                                                                                                  
 concatenate_1 (Concatenate)    (None, 32, 32, 128)  0           ['conv2d_transpose_1[0][0]',     
                                                                  'conv2d_5[0][0]']               
                                                                                                  
 conv2d_12 (Conv2D)             (None, 32, 32, 64)   73792       ['concatenate_1[0][0]']          
                                                                                                  
 dropout_6 (Dropout)            (None, 32, 32, 64)   0           ['conv2d_12[0][0]']              
                                                                                                  
 conv2d_13 (Conv2D)             (None, 32, 32, 64)   36928       ['dropout_6[0][0]']              
                                                                                                  
 conv2d_transpose_2 (Conv2DTran  (None, 64, 64, 32)  8224        ['conv2d_13[0][0]']              
 spose)                                                                                           
                                                                                                  
 concatenate_2 (Concatenate)    (None, 64, 64, 64)   0           ['conv2d_transpose_2[0][0]',     
                                                                  'conv2d_3[0][0]']               
                                                                                                  
 conv2d_14 (Conv2D)             (None, 64, 64, 32)   18464       ['concatenate_2[0][0]']          
                                                                                                  
 dropout_7 (Dropout)            (None, 64, 64, 32)   0           ['conv2d_14[0][0]']              
                                                                                                  
 conv2d_15 (Conv2D)             (None, 64, 64, 32)   9248        ['dropout_7[0][0]']              
                                                                                                  
 conv2d_transpose_3 (Conv2DTran  (None, 128, 128, 16  2064       ['conv2d_15[0][0]']              
 spose)                         )                                                                 
                                                                                                  
 concatenate_3 (Concatenate)    (None, 128, 128, 32  0           ['conv2d_transpose_3[0][0]',     
                                )                                 'conv2d_1[0][0]']               
                                                                                                  
 conv2d_16 (Conv2D)             (None, 128, 128, 16  4624        ['concatenate_3[0][0]']          
                                )                                                                 
                                                                                                  
 dropout_8 (Dropout)            (None, 128, 128, 16  0           ['conv2d_16[0][0]']              
                                )                                                                 
                                                                                                  
 conv2d_17 (Conv2D)             (None, 128, 128, 16  2320        ['dropout_8[0][0]']              
                                )                                                                 
                                                                                                  
 conv2d_18 (Conv2D)             (None, 128, 128, 1)  17          ['conv2d_17[0][0]']              
                                                                                                  
==================================================================================================
Total params: 1,941,105
Trainable params: 1,941,105
Non-trainable params: 0
__________________________________________________________________________________________________

1
2
3
4
from tensorflow.keras import utils
from keras.utils.vis_utils import model_to_dot

utils.plot_model(model, 'model1.png',show_shapes=True,show_dtype=True,show_layer_names=True)

png

75 - Image Segmentation using U-Net - Part 3 -What are trainable parameters-

​ 介绍了神经网络中参数量是怎么算的。

76 - Image Segmentation using U-Net - Part 4 -Model fitting- checkpoints- and callbacks-

Modelcheckpoint

1
2
3
4
5
checkpointer = tf.keras.callbacks.ModelCheckpoint('model_for_nuclei.h5', verbose=1, save_best_only=True)
callbacks = [
    tf.keras.callbacks.EarlyStopping(patience=2, monitor='val_loss'),
    tf.keras.callbacks.TensorBoard(log_dir='logs')
]

77 - Image Segmentation using U-Net - Part 5 -Understanding the data-

获取数据集：2018 Data Science Bowl | Kaggle

78 - Image Segmentation using U-Net - Part 6 -Running the code and understanding results-

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
import tensorflow as tf
import os
import numpy as np

IMG_WIDTH = 128
IMG_HEIGHT = 128
IMG_CHANNELS = 3

TRAIN_PATH = 'stage1_train/'
TEST_PATH = 'stage1_test/'

train_ids = next(os.walk(TRAIN_PATH))[1]
test_ids = next(os.walk(TEST_PATH))[1]

X_train = np.zeros((len(train_ids), IMG_HEIGHT, IMG_WIDTH, IMG_CHANNELS), dtype=np.uint8)
Y_train = np.zeros((len(train_ids), IMG_HEIGHT, IMG_WIDTH, 1), dtype=np.bool_)  # y 是一个二值图像

Resizing training images and masks

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
import random
from tqdm import tqdm 
from skimage.io import imread, imshow
from skimage.transform import resize
import matplotlib.pyplot as plt

seed = 42
np.random.seed = seed

print('Resizing training images and masks')
for n, id_ in tqdm(enumerate(train_ids), total=len(train_ids)):   
    path = TRAIN_PATH + id_
    img = imread(path + '/images/' + id_ + '.png')[:,:,:IMG_CHANNELS]  # 读取图像
    img = resize(img, (IMG_HEIGHT, IMG_WIDTH), mode='constant', preserve_range=True)  # 将图像大小修改为 128x128
    X_train[n] = img  # Fill empty X_train with values from img
    mask = np.zeros((IMG_HEIGHT, IMG_WIDTH, 1), dtype=np.bool_)
    for mask_file in next(os.walk(path + '/masks/'))[2]:
        mask_ = imread(path + '/masks/' + mask_file)
        mask_ = np.expand_dims(resize(mask_, (IMG_HEIGHT, IMG_WIDTH), mode='constant',  
                                      preserve_range=True), axis=-1)
        mask = np.maximum(mask, mask_)  
            
    Y_train[n] = mask # Fill empty Y_train with values from img
Resizing training images and masks


100%|██████████| 670/670 [04:38<00:00,  2.41it/s]

Resizing test images

1
2
3
4
5
6
7
8
9
10
11
X_test = np.zeros((len(test_ids), IMG_HEIGHT, IMG_WIDTH, IMG_CHANNELS), dtype=np.uint8)
sizes_test = []
print('Resizing test images') 
for n, id_ in tqdm(enumerate(test_ids), total=len(test_ids)):
    path = TEST_PATH + id_
    img = imread(path + '/images/' + id_ + '.png')[:,:,:IMG_CHANNELS]
    sizes_test.append([img.shape[0], img.shape[1]])
    img = resize(img, (IMG_HEIGHT, IMG_WIDTH), mode='constant', preserve_range=True)
    X_test[n] = img

print('Done!')
Resizing test images


100%|██████████| 65/65 [00:01<00:00, 50.05it/s]

Done!

​
​

1
2
3
image_x = random.randint(0, len(train_ids))
imshow(X_train[image_x])
plt.show()

png

1
2
imshow(np.squeeze(Y_train[image_x]))
plt.show()

png

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
inputs = tf.keras.layers.Input((IMG_HEIGHT, IMG_WIDTH, IMG_CHANNELS))
s = tf.keras.layers.Lambda(lambda x: x / 255)(inputs)  # 将整数转换成浮点数

# Contraction path 收缩路径
c1 = tf.keras.layers.Conv2D(16, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(s)
c1 = tf.keras.layers.Dropout(0.1)(c1)
c1 = tf.keras.layers.Conv2D(16, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c1)
p1 = tf.keras.layers.MaxPooling2D((2, 2))(c1)

c2 = tf.keras.layers.Conv2D(32, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(p1)
c2 = tf.keras.layers.Dropout(0.1)(c2)
c2 = tf.keras.layers.Conv2D(32, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c2)
p2 = tf.keras.layers.MaxPooling2D((2, 2))(c2)
 
c3 = tf.keras.layers.Conv2D(64, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(p2)
c3 = tf.keras.layers.Dropout(0.2)(c3)
c3 = tf.keras.layers.Conv2D(64, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c3)
p3 = tf.keras.layers.MaxPooling2D((2, 2))(c3)
 
c4 = tf.keras.layers.Conv2D(128, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(p3)
c4 = tf.keras.layers.Dropout(0.2)(c4)
c4 = tf.keras.layers.Conv2D(128, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c4)
p4 = tf.keras.layers.MaxPooling2D(pool_size=(2, 2))(c4)
 
c5 = tf.keras.layers.Conv2D(256, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(p4)
c5 = tf.keras.layers.Dropout(0.3)(c5)
c5 = tf.keras.layers.Conv2D(256, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c5)

# Expansive path 延展路径
u6 = tf.keras.layers.Conv2DTranspose(128, (2, 2), strides=(2, 2), padding='same')(c5)
u6 = tf.keras.layers.concatenate([u6, c4])
c6 = tf.keras.layers.Conv2D(128, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(u6)
c6 = tf.keras.layers.Dropout(0.2)(c6)
c6 = tf.keras.layers.Conv2D(128, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c6)
 
u7 = tf.keras.layers.Conv2DTranspose(64, (2, 2), strides=(2, 2), padding='same')(c6)
u7 = tf.keras.layers.concatenate([u7, c3])
c7 = tf.keras.layers.Conv2D(64, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(u7)
c7 = tf.keras.layers.Dropout(0.2)(c7)
c7 = tf.keras.layers.Conv2D(64, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c7)
 
u8 = tf.keras.layers.Conv2DTranspose(32, (2, 2), strides=(2, 2), padding='same')(c7)
u8 = tf.keras.layers.concatenate([u8, c2])
c8 = tf.keras.layers.Conv2D(32, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(u8)
c8 = tf.keras.layers.Dropout(0.1)(c8)
c8 = tf.keras.layers.Conv2D(32, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c8)
 
u9 = tf.keras.layers.Conv2DTranspose(16, (2, 2), strides=(2, 2), padding='same')(c8)
u9 = tf.keras.layers.concatenate([u9, c1], axis=3)
c9 = tf.keras.layers.Conv2D(16, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(u9)
c9 = tf.keras.layers.Dropout(0.1)(c9)
c9 = tf.keras.layers.Conv2D(16, (3, 3), activation='relu', kernel_initializer='he_normal', padding='same')(c9)
 
outputs = tf.keras.layers.Conv2D(1, (1, 1), activation='sigmoid')(c9)
 
model = tf.keras.Model(inputs=[inputs], outputs=[outputs])
model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])

Modelcheckpoint

1
2
3
4
5
checkpointer = tf.keras.callbacks.ModelCheckpoint('model_for_nuclei.h5', verbose=1, save_best_only=True)
callbacks = [
    tf.keras.callbacks.EarlyStopping(patience=2, monitor='val_loss'),
    tf.keras.callbacks.TensorBoard(log_dir='logs')
]
1
results = model.fit(X_train, Y_train, validation_split=0.1, batch_size=16, epochs=25, callbacks=callbacks)
Epoch 1/25
38/38 [==============================] - 20s 280ms/step - loss: 0.5785 - accuracy: 0.7459 - val_loss: 0.3360 - val_accuracy: 0.7809
Epoch 2/25
38/38 [==============================] - 7s 173ms/step - loss: 0.2739 - accuracy: 0.8642 - val_loss: 0.2158 - val_accuracy: 0.9125
Epoch 3/25
38/38 [==============================] - 7s 174ms/step - loss: 0.1924 - accuracy: 0.9304 - val_loss: 0.1558 - val_accuracy: 0.9406
Epoch 4/25
38/38 [==============================] - 7s 173ms/step - loss: 0.1589 - accuracy: 0.9422 - val_loss: 0.1474 - val_accuracy: 0.9425
Epoch 5/25
38/38 [==============================] - 7s 175ms/step - loss: 0.1449 - accuracy: 0.9472 - val_loss: 0.1173 - val_accuracy: 0.9543
Epoch 6/25
38/38 [==============================] - 7s 176ms/step - loss: 0.1394 - accuracy: 0.9480 - val_loss: 0.1090 - val_accuracy: 0.9565
Epoch 7/25
38/38 [==============================] - 7s 176ms/step - loss: 0.1246 - accuracy: 0.9540 - val_loss: 0.1084 - val_accuracy: 0.9579
Epoch 8/25
38/38 [==============================] - 7s 177ms/step - loss: 0.1204 - accuracy: 0.9553 - val_loss: 0.1113 - val_accuracy: 0.9577
Epoch 9/25
38/38 [==============================] - 7s 177ms/step - loss: 0.1124 - accuracy: 0.9581 - val_loss: 0.0957 - val_accuracy: 0.9611
Epoch 10/25
38/38 [==============================] - 7s 177ms/step - loss: 0.1087 - accuracy: 0.9591 - val_loss: 0.0930 - val_accuracy: 0.9618
Epoch 11/25
38/38 [==============================] - 7s 179ms/step - loss: 0.1079 - accuracy: 0.9593 - val_loss: 0.0892 - val_accuracy: 0.9650
Epoch 12/25
38/38 [==============================] - 7s 179ms/step - loss: 0.1044 - accuracy: 0.9605 - val_loss: 0.0918 - val_accuracy: 0.9643
Epoch 13/25
38/38 [==============================] - 7s 180ms/step - loss: 0.1021 - accuracy: 0.9608 - val_loss: 0.0888 - val_accuracy: 0.9651
Epoch 14/25
38/38 [==============================] - 7s 180ms/step - loss: 0.0989 - accuracy: 0.9621 - val_loss: 0.0950 - val_accuracy: 0.9616
Epoch 15/25
38/38 [==============================] - 7s 180ms/step - loss: 0.1003 - accuracy: 0.9612 - val_loss: 0.0858 - val_accuracy: 0.9661
Epoch 16/25
38/38 [==============================] - 7s 180ms/step - loss: 0.1005 - accuracy: 0.9613 - val_loss: 0.0872 - val_accuracy: 0.9660
Epoch 17/25
38/38 [==============================] - 7s 181ms/step - loss: 0.0963 - accuracy: 0.9630 - val_loss: 0.0846 - val_accuracy: 0.9659
Epoch 18/25
38/38 [==============================] - 7s 184ms/step - loss: 0.0956 - accuracy: 0.9632 - val_loss: 0.0840 - val_accuracy: 0.9665
Epoch 19/25
38/38 [==============================] - 7s 183ms/step - loss: 0.0925 - accuracy: 0.9642 - val_loss: 0.0819 - val_accuracy: 0.9675
Epoch 20/25
38/38 [==============================] - 7s 183ms/step - loss: 0.0927 - accuracy: 0.9640 - val_loss: 0.0833 - val_accuracy: 0.9677
Epoch 21/25
38/38 [==============================] - 7s 183ms/step - loss: 0.0917 - accuracy: 0.9641 - val_loss: 0.0846 - val_accuracy: 0.9660

1
2
3
4
5
6
7
8
9
idx = random.randint(0, len(X_train))

preds_train = model.predict(X_train[:int(X_train.shape[0]*0.9)], verbose=1)
preds_val = model.predict(X_train[int(X_train.shape[0]*0.9):], verbose=1)
preds_test = model.predict(X_test, verbose=1)

preds_train_t = (preds_train > 0.5).astype(np.uint8)
preds_val_t = (preds_val > 0.5).astype(np.uint8)
preds_test_t = (preds_test > 0.5).astype(np.uint8)
19/19 [==============================] - 5s 161ms/step
3/3 [==============================] - 0s 91ms/step
3/3 [==============================] - 1s 417ms/step

Perform a sanity check on some random training samples

  • 对一些随机训练样本进行合理性检查
1
2
3
4
5
6
7
ix = random.randint(0, len(preds_train_t))
imshow(X_train[ix])
plt.show()
imshow(np.squeeze(Y_train[ix]))
plt.show()
imshow(np.squeeze(preds_train_t[ix]))
plt.show()

png

png

C:\Users\gzjzx\anaconda3\lib\site-packages\skimage\io\_plugins\matplotlib_plugin.py:150: UserWarning: Low image data range; displaying image with stretched contrast.
  lo, hi, cmap = _get_display_range(image)

png

Perform a sanity check on some random validation samples

  • 对一些随机的验证样本进行合理性检查
1
2
3
4
5
6
7
ix = random.randint(0, len(preds_val_t))
imshow(X_train[int(X_train.shape[0]*0.9):][ix])
plt.show()
imshow(np.squeeze(Y_train[int(Y_train.shape[0]*0.9):][ix]))
plt.show()
imshow(np.squeeze(preds_val_t[ix]))
plt.show()

png

png

png

可视化结果

1
!tensorboard --logdir=logs/ --host localhost --port 8088

png

Icon
Icon
Icon
Icon
上一篇：DIP-Python tutorials for image processing and machine learning(70-71)-Deep Learning
下一篇：DIP-Python tutorials for image processing and machine learning(69)-BOVW