我如何知道图像分类器中图像的最佳重塑尺寸？

Question

当我尝试创建只有两个数字（7 和 10）的手写数字图像数据集时，我尝试加载自定义图像（原始颜色：黑白，尺寸：251 x 54 请参见下面的示例）我在下面的 load_img 函数中遇到了这个错误：

from tensorflow.keras.preprocessing.image import load_img
from tensorflow.keras.preprocessing.image import img_to_array
from tensorflow.keras.models import load_model


# load and prepare the image
def load_image(filename):
    # load the image
    img = load_img(filename, color_mode="grayscale",interpolation='nearest')
    # convert to array
    img = img_to_array(img)
    # reshape into a single sample with 1 channel
    img = img.reshape(2, 200, 50, 1)
    # prepare pixel data
    img = img.astype('float32')
    img = img / 255.0
    return img




# load an image and predict the class
def run_example():
    # load the image
    img = load_image('C:/Users/ADEM/Desktop/msi_youssef/PFE/dataset/10/kz.png')
    # load model
    model = load_model('C:/Users/ADEM/Desktop/msi_youssef/PFE/other_shit/first_try.h5')
    # predict the class
    digit = model.predict_classes(img)
    print(digit[0])


# entry point, run the example
run_example()  

这是我得到的错误：

ValueError                                Traceback (most recent call last)
<ipython-input-2-5427252e970b> in <module>
     32 
     33 # entry point, run the example
---> 34 run_example()

<ipython-input-2-5427252e970b> in run_example()
     23 def run_example():
     24     # load the image
---> 25     img = load_image('C:/Users/ADEM/Desktop/msi_youssef/PFE/dataset/10/kz.png')
     26     # load model
     27     model = load_model('C:/Users/ADEM/Desktop/msi_youssef/PFE/other_shit/final_model.h5')

<ipython-input-2-5427252e970b> in load_image(filename)
     11     img = img_to_array(img)
     12     # reshape into a single sample with 1 channel
---> 13     img = img.reshape(2, 200, 50, 1)
     14     # prepare pixel data
     15     img = img.astype('float32')

ValueError: cannot reshape array of size 13554 into shape (2,200,50,1) 

请注意，在 final_model.h5 中，我将 img 的平均大小设置为 200，50
final_model.h5 的代码将在第一个答案中！

Answer 1

from keras.preprocessing.image import ImageDataGenerator
from keras.models import Sequential
from keras.layers import Conv2D, MaxPooling2D
from keras.layers import Activation, Dropout, Flatten, Dense
from keras import backend as K


# dimensions of our images.
img_width, img_height = 200, 55

train_data_dir = 'C:/Users/ADEM/Desktop/msi_youssef/PFE/test/numbers/data/train'
validation_data_dir = 'C:/Users/ADEM/Desktop/msi_youssef/PFE/test/numbers/data/valid'
nb_train_samples = 140
nb_validation_samples = 30
epochs = 10 # how much time you want to train your model on the data
batch_size = 16

if K.image_data_format() == 'channels_first':
    input_shape = (3, img_width, img_height)
else:
    input_shape = (img_width, img_height, 3)

model = Sequential()
model.add(Conv2D(32, (3, 3), input_shape=input_shape))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))

model.add(Conv2D(32, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))

model.add(Conv2D(64, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))

model.add(Flatten())
model.add(Dense(64))
model.add(Activation('relu'))
model.add(Dropout(0.5))
model.add(Dense(1))
model.add(Activation('sigmoid'))

model.compile(loss='binary_crossentropy',optimizer='rmsprop',metrics=['accuracy'])

# this is the augmentation configuration we will use for training
train_datagen = ImageDataGenerator(
    rescale=1. / 255,
    shear_range=0.1,
    zoom_range=0.05,
    horizontal_flip=False)

# this is the augmentation configuration we will use for testing:
# only rescaling
test_datagen = ImageDataGenerator(rescale=1. / 255)

train_generator = train_datagen.flow_from_directory(
    train_data_dir,
    target_size=(img_width, img_height),
    batch_size=batch_size,
    class_mode='binary')

validation_generator = test_datagen.flow_from_directory(
    validation_data_dir,
    target_size=(img_width, img_height),
    batch_size=batch_size,
    class_mode='binary')

model.fit_generator(
    train_generator,
    steps_per_epoch=nb_train_samples // batch_size,
    epochs=epochs,
    validation_data=validation_generator,
    validation_steps=nb_validation_samples // batch_size)

model.save('first_try.h5')  

Answer 2

2D 卷积层需要输入为 -

if using channels_last: (batch_size, imageside1, imageside2, channels)
if using channels_first: (batch_size, channels, imageside1, imageside2)

在你的情况下 batch_size = 不指定， imageside1 = 200, imageside1 = 50, channels = 1（灰度图）

所以修改您的load_image 函数并进行以下更改

# load and prepare the image
def load_image(filename):
    # load the image with target size
    img = load_img(filename, color_mode="grayscale",interpolation='nearest',target_size=(200,50))
    # convert to array
    img = img_to_array(img)
    # reshape into a single sample with 1 channel
    # img = img.reshape(2, 200, 50, 1) --> This is not required now and why batch size argument as 2?
    # prepare pixel data
    img = img.astype('float32')
    img = img / 255.0
    return img