1、ZCA白化
from keras.datasets import mnist
from keras.preprocessing.image import ImageDataGenerator
from matplotlib import pyplot as plt
from keras import backend
backend.set_image_data_format('channels_first')
# 导入数据
(X_train, y_train), (X_validation, y_validation) = mnist.load_data()
X_train = X_train.reshape(X_train.shape[0], 1, 28, 28).astype('float32')
X_validation = X_validation.reshape(X_validation.shape[0], 1, 28, 28).astype('float32')
# ZCA白化
imgGen = ImageDataGenerator(zca_whitening=True)
imgGen.fit(X_train)
# 调用 flow 函数来生成样本
# 或者flow_from_directory 函数来完成相应的样本生成
for X_batch, y_batch in imgGen.flow(X_train, y_train, batch_size=9):
for i in range(0, 9):
plt.subplot(331 + i)
plt.imshow(X_batch[i].reshape(28, 28), cmap=plt.get_cmap('gray'))
plt.show()
break
2、特征标准化
from keras.datasets import mnist
from keras.preprocessing.image import ImageDataGenerator
from matplotlib import pyplot as plt
from keras import backend
backend.set_image_data_format('channels_first')
# 导入数据
(X_train, y_train), (X_validation, y_validation) = mnist.load_data()
#X_train.shape (60000,28,28)
X_train = X_train.reshape(X_train.shape[0], 1, 28, 28).astype('float32')
X_validation = X_validation.reshape(X_validation.shape[0], 1, 28, 28).astype('float32')
# ImageDataGenerator()是keras.preprocessing.image模块中的图片生成器,
# 同时也可以在batch中对数据进行增强,扩充数据集大小,增强模型的泛化能力。
# 比如进行旋转,变形,归一化等等
# 图像特征化
# featurewise_center: Boolean. 对输入的图片每个通道减去每个通道对应均值。
# samplewise_center: Boolan. 每张图片减去样本均值, 使得每个样本均值为0。
imgGen = ImageDataGenerator(featurewise_center=True, featurewise_std_normalization=True)
imgGen.fit(X_train)
# 调用 flow 函数来生成样本
# 或者flow_from_directory 函数来完成相应的样本生成
for X_batch, y_batch in imgGen.flow(X_train, y_train, batch_size=9):
for i in range(0, 9):
plt.subplot(331 + i)
plt.imshow(X_batch[i].reshape(28, 28), cmap=plt.get_cmap('gray'))
plt.show()
break
3、随机旋转移动剪切和反转图像
from keras.datasets import mnist
from keras.preprocessing.image import ImageDataGenerator
from matplotlib import pyplot as plt
from keras import backend
backend.set_image_data_format('channels_first')
# 导入数据
(X_train, y_train), (X_validation, y_validation) = mnist.load_data()
X_train = X_train.reshape(X_train.shape[0], 1, 28, 28).astype('float32')
X_validation = X_validation.reshape(X_validation.shape[0], 1, 28, 28).astype('float32')
# 图像旋转
imgGen = ImageDataGenerator(rotation_range=90)
imgGen.fit(X_train)
for X_batch, y_batch in imgGen.flow(X_train, y_train, batch_size=9):
for i in range(0, 9):
plt.subplot(331 + i)
plt.imshow(X_batch[i].reshape(28, 28), cmap=plt.get_cmap('gray'))
plt.show()
break
# 图像移动
imgGen = ImageDataGenerator(width_shift_range=0.2, height_shift_range=0.2)
imgGen.fit(X_train)
# 调用 flow 函数来生成样本
# 或者flow_from_directory 函数来完成相应的样本生成
for X_batch, y_batch in imgGen.flow(X_train, y_train, batch_size=9):
for i in range(0, 9):
plt.subplot(331 + i)
plt.imshow(X_batch[i].reshape(28, 28), cmap=plt.get_cmap('gray'))
plt.show()
break
# 图像剪切
imgGen = ImageDataGenerator(shear_range=0.2)
imgGen.fit(X_train)
for X_batch, y_batch in imgGen.flow(X_train, y_train, batch_size=9):
for i in range(0, 9):
plt.subplot(331 + i)
plt.imshow(X_batch[i].reshape(28, 28), cmap=plt.get_cmap('gray'))
plt.show()
break
# 图像翻转
imgGen = ImageDataGenerator(horizontal_flip=True, vertical_flip=True)
imgGen.fit(X_train)
for X_batch, y_batch in imgGen.flow(X_train, y_train, batch_size=9):
for i in range(0, 9):
plt.subplot(331 + i)
plt.imshow(X_batch[i].reshape(28, 28), cmap=plt.get_cmap('gray'))
plt.show()
break