使用VGG16和MLP实现猫狗图像识别

数据集

数据集可以参考我之前那篇文章，取一部分数据每个300条即可：基于卷积神经网络（CNN）的猫狗图像分类系统实现-CSDN博客

1.目的

使用VGG16的结构提取图像特征，再根据特征建立MLP模型，实现猫狗图像识别。训练/测试数据：data  

1.对数据进行分离、计算测试数据准确率  

2.使用VGG16提取特征，再根据特征建立MLP模型，实现猫狗图像识别。  

3.从网站下载猫\狗图片，对其进行预测  

mlp模型一个隐藏层，10个神经元

2.模型训练

1.数据加载

#加载数据
from keras.preprocessing.image import load_img, img_to_array

img_path = '1.jpg'
img = load_img(img_path, target_size=(224, 224))
img = img_to_array(img)
type(img) #查看数据类型

运行结果：

2.对图像进行预处理

from keras.applications.vgg16 import VGG16
from keras.applications.vgg16 import preprocess_input
import numpy as np
model_vgg = VGG16(weights='imagenet', include_top=False)
x = np.expand_dims(img, axis=0)     #添加一个维度
x = preprocess_input(x)             #对图像进行预处理
print(x.shape)

运行结果

3.特征提取

#特征提取
features = model_vgg.predict(x)     #提取特征
print(features.shape)

运行结果

4.将特征展平

#将特征展平
features = features.reshape(1, 7*7*512) 
print(features.shape)

运行结果

3.可视化数据

#可视化数据
%matplotlib inline
import matplotlib.pyplot as plt
fig = plt.figure(figsize=(5, 5))
img = load_img(img_path, target_size=(224, 224))
plt.imshow(img)

运行结果

4.集中处理图像并生成寻来你数据集

#用vgg16结构加载图像并进行预处理
from keras.preprocessing.image import img_to_array,load_img
from keras.applications.vgg16 import VGG16
from keras.applications.vgg16 import preprocess_input
import numpy as np

model_vgg =VGG16(weights='imagenet',include_top=False)


#定义加载和预处理图像的方法
def modelProcess(img_path,model):
    img = load_img(img_path, target_size=(224, 224))
    img = img_to_array(img)
    x= np.expand_dims(img,axis=0)
    x= preprocess_input(x)
    x_vgg = model.predict(x)
    x_vgg =x_vgg.reshape(1,25088)
    return x_vgg


#列出训练数据集的文件名
import os
folder = "./data_vgg/cats2"
dirs = os.listdir(folder)
#为图像生成路径
img_path = []
for i in dirs:
    if os.path.splitext(i)[1] == ".jpg":
        img_path.append(i)
img_path = [folder+"//"+i for i in img_path]
#预处理多个图像
features1 =np.zeros([len(img_path),25088])
for i in range(len(img_path)):
    feature_i = modelProcess(img_path[i],model_vgg)
    print('preprocessed：',img_path[i])
    features1[i] = feature_i
    
folder ="./data_vgg/dogs2"
dirs = os.listdir(folder)
img_path = []
for i in dirs:
    if os.path.splitext(i)[1] == ".jpg":
        img_path.append(i)
img_path =[folder+"//"+i for i in img_path]
features2 =np.zeros([len(img_path),25088])
for i in range(len(img_path)):
    featute_i = modelProcess(img_path[i],model_vgg)
    print("preprocessed:",img_path[i])
    features2[i] = feature_i
#标记结果
print(features1.shape,features2.shape)
y1 = np.zeros(300)
y2= np.ones(300)
#生成训练数据
X = np.concatenate((features1,features2),axis=0)
y = np.concatenate((y1,y2),axis=0)
y = y.reshape(-1,1)
print(X.shape,y.shape)

运行结果

5.拆分训练和测试数据

#拆分训练和测试数据
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=50)
print(X_train.shape,X_test.shape,X.shape)

运行结果

6.建立MLP模型

#建立MLP模型
from keras.models import Sequential
from keras.layers import Dense
model = Sequential()
model.add(Dense(units=20,activation='relu',input_dim=25088))
model.add(Dense(units=1,activation='sigmoid'))
model.summary()

运行结果

7.配置并训练模型

#配置模型
model.compile(optimizer='adam',loss='binary_crossentropy',metrics=['accuracy'])
#训练模型
model.fit(X_train,y_train,epochs=20)

运行结果

8.导出模型

#导出模型
model.save('cat_vs_dog.keras')

3.模型评估

1.训练准确率

#评估模型
from sklearn.metrics import accuracy_score
y_train_predict = model.predict(X_train)
y_train_predict = np.argmax(y_train_predict,axis=1)
y_train_1 = np.argmax(y_train,axis=1)
accuracy_train = accuracy_score(y_train_1, y_train_predict)
print("训练集准确率：",accuracy_train)

运行结果

2.测试准确率

#测试准确率
y_test_predict = model.predict(X_test)
y_test_predict = np.argmax(y_test_predict,axis=1)
y_test_1 = np.argmax(y_test,axis=1)
accuracy_test = accuracy_score(y_test_1, y_test_predict)
print("测试集准确率：",accuracy_test)

运行结果

4.测试模型

1.测试单张图片

#测试单张图片
img_path = '1.jpg'
img = load_img(img_path, target_size=(224, 224))
img = img_to_array(img)
x= np.expand_dims(img,axis=0)
x= preprocess_input(x)
features = model_vgg.predict(x)
features = features.reshape(1, 7*7*512)
result = model.predict(features)
# result = np.argmax(result, axis=1)
result = (result > 0.5).astype(int)     #给一个阈值进行分类
print(result)

运行结果

2.多张进行图片分类

#进行图片分类
import matplotlib as mlp
font2 = {'family' : 'SimHei',
'weight' : 'normal',
'size'   : 20,
}
mlp.rcParams['font.family'] = 'SimHei'
mlp.rcParams['axes.unicode_minus'] = False
from matplotlib import pyplot as plt
from matplotlib.image import imread
from keras.preprocessing.image import load_img
from keras.preprocessing.image import img_to_array
from keras.models import load_model
import numpy as np
a = [i for i in range(1,10)]
fig = plt.figure(figsize=(10,10))
for i in a:
    img_name = str(i)+'.jpg'
    img_path = img_name
    img = load_img(img_path, target_size=(224, 224))
    img = img_to_array(img)
    x = np.expand_dims(img,axis=0)
    x = preprocess_input(x)
    x_vgg = model_vgg.predict(x)
    x_vgg = np.argmax(x_vgg,axis=0)
    x_vgg = x_vgg.reshape(1,7*7*512)
    result = model.predict(x_vgg)
    result = np.argmax(result,axis=1)
    print(result)
    img_ori = load_img(img_name, target_size=(250, 250))
    plt.subplot(3,3,i)
    plt.imshow(img_ori)
    plt.title('预测为：狗狗' if result[0] == 1 else '预测为：猫咪')
plt.show()

运行结果

5.总结

        使用VGG16和多层感知机（MLP）结合的方法来实现猫狗图像识别，是一种结合了迁移学习和传统机器学习分类器的有效策略。具体实现中，首先利用预训练的VGG16模型作为特征提取器，该模型已经在大型图像数据集（如ImageNet）上进行了训练，因此能够捕捉到图像中的高级特征。然后，从VGG16模型的某一层（通常是全连接层之前的全局平均池化层或卷积层的输出）提取特征，并将这些特征作为输入传递给一个较简单的MLP模型进行分类。