第五章 深度学习用于计算机视觉

""" 
2.1.猫狗大战 dogs-vs-cats (创建目录及分配图像) 
"""
import os, shutil

original_datasets_dir ='F:\datasets\dogs-vs-csts\dogs-vs-csts 数据集'       # 原始数据集解压目录的路径

base_dir = 'F:\datasets\dogs-vs-csts\猫狗大战'        # 保存较小数据集的目录 (创建目录)
os.mkdir(base_dir)

train_dir = os.path.join(base_dir, 'train')             # 划分后的训练集
os.mkdir(train_dir)
validation_dir = os.path.join(base_dir, 'validation')   # 划分后的验证集
os.mkdir(validation_dir)
test_dir = os.path.join(base_dir, 'test')               # 划分后的测试集
os.mkdir(test_dir)

train_cats_dir = os.path.join(train_dir, 'cats')         # 猫的训练图像目录
os.mkdir(train_cats_dir)
train_dogs_dir = os.path.join(train_dir, 'dogs')         # 狗的训练图像目录
os.mkdir(train_dogs_dir)

validation_cats_dir = os.path.join(validation_dir, 'cats')    # 猫的验证图像目录
os.mkdir(validation_cats_dir)
validation_dogs_dir = os.path.join(validation_dir, 'dogs')    # 狗的验证图像目录
os.mkdir(validation_dogs_dir)

test_cats_dir = os.path.join(test_dir, 'cats')          # 猫的测试图像目录
os.mkdir(test_cats_dir)
test_dogs_dir = os.path.join(test_dir, 'dogs')          # 狗的测试图像目录
os.mkdir(test_dogs_dir)

fnames = ['cat.{}.jpg'.format(i) for i in range(1000)]         # 猫的训练集：将前1000张猫的图像复制到train_cats_dir
for fname in fnames:
    src = os.path.join(original_datasets_dir, fname)
    dst = os.path.join(train_cats_dir, fname)
    shutil.copyfile(src, dst)
    
fnames = ['cat.{}.jpg'.format(i) for i in range(1000, 1500)]   # 毛的验证集：将接下来的500张猫的图像复制到validation_cats_dir
for fname in fnames:
    src = os.path.join(original_datasets_dir, fname)
    dst = os.path.join(validation_cats_dir, fname)
    shutil.copyfile(src, dst)
    
fnames = ['cat.{}.jpg'.format(i) for i in range(1500, 2000)]   # 猫的测试集：将接下来的500张猫的图像复制到test_cats_dir
for fname in fnames:
    src = os.path.join(original_datasets_dir, fname)
    dst = os.path.join(test_cats_dir, fname)
    shutil.copyfile(src, dst)

fnames = ['dog.{}.jpg'.format(i) for i in range(1000)]         # 狗的训练集：将前1000张狗的图像复制到train_dogs_dir
for fname in fnames:
    src = os.path.join(original_datasets_dir, fname)
    dst = os.path.join(train_dogs_dir, fname)
    shutil.copyfile(src, dst)

fnames = ['dog.{}.jpg'.format(i) for i in range(1000, 1500)]   # 狗的验证集:将接下来的500张狗的图像复制到validation_cdogs_dir
for fname in fnames:
    src = os.path.join(original_datasets_dir, fname)
    dst = os.path.join(validation_dogs_dir, fname)
    shutil.copyfile(src, dst)

fnames = ['dog.{}.jpg'.format(i) for i in range(1500, 2000)]   # 狗的测试集将前1000张狗的图像复制到test_dogs_dir
for fname in fnames:
    src = os.path.join(original_datasets_dir, fname)
    dst = os.path.join(test_dogs_dir, fname)
    shutil.copyfile(src, dst)

print('total training cat images:', len(os.listdir(train_cats_dir)))           # 猫的训练图像张数 1000
print('total training dog images:', len(os.listdir(train_dogs_dir)))           # 狗的训练图像张数 1000
print('total validation cat images:', len(os.listdir(validation_cats_dir)))    # 猫的验证图像张数 500
print('total validation dog images:', len(os.listdir(validation_dogs_dir)))    # 狗的验证图像张数 500
print('total test cat images:', len(os.listdir(test_cats_dir)))                # 狗的测试图像张数 500
print('total test dog images:', len(os.listdir(test_dogs_dir)))                # 狗的测试图像张数 500

""" 
2.2.猫狗大战 dogs-vs-cats 
"""

import os
base_dir = 'F:\datasets\dogs-vs-csts\猫狗大战'             # 保存较小数据集的目录

train_dir = os.path.join(base_dir, 'train')              # 划分后的训练集
validation_dir = os.path.join(base_dir, 'validation')    # 划分后的验证集
test_dir = os.path.join(base_dir, 'test')                # 划分后的测试集
train_cats_dir = os.path.join(train_dir, 'cats')         # 猫的训练集


# 1.将猫狗分类的小型卷积神经网络实例化
from keras import models
from keras import layers

model = models.Sequential()
model.add(layers.Conv2D(32, (3,3), activation='relu', input_shape=(150, 150, 3)))   # 3D
model.add(layers.MaxPooling2D((2,2)))
model.add(layers.Conv2D(64, (3,3), activation='relu'))
model.add(layers.MaxPooling2D((2,2)))
model.add(layers.Conv2D(128, (3,3), activation='relu'))
model.add(layers.MaxPooling2D((2