python+opencv人脸识别是否戴口罩

当今疫情情况下，检测是否戴口罩还是比较流行的，然后就搞啦一个检测人脸是否戴口罩的程序。程序调用笔记本摄像头，实时检测人脸是否佩戴口罩。

一、数据集

如果数据集大家有，就用自己的。没有的话，也可以找我要啊（3000+戴口罩/1000+不戴口罩）。在此就不分享出来了，我也是找别人要的。

二、开始上代码

1、导入keras库

import keras
keras.__version__

　　

 2、创建train、test、validation文件夹，并且三个文件夹里面都有have_mask和no_mask文件，运行之后将数据集中的照片放入对应的文件夹中。

import os, shutil
# The path to the directory where the original
# dataset was uncompressed
original_dataset_dir = 'F:\PyCharm\facemask\mask'

# The directory where we will
# store our smaller dataset
base_dir = 'F:\PyCharm\facemask\mask_and_nomask'
os.mkdir(base_dir)

# Directories for our training,
# validation and test splits
train_dir = os.path.join(base_dir, 'train')
os.mkdir(train_dir)
validation_dir = os.path.join(base_dir, 'validation')
os.mkdir(validation_dir)
test_dir = os.path.join(base_dir, 'test')
os.mkdir(test_dir)

# Directory with our training smile pictures
train_smile_dir = os.path.join(train_dir, 'have_mask')
os.mkdir(train_smile_dir)

# Directory with our training nosmile pictures
train_nosmile_dir = os.path.join(train_dir, 'no_mask')
os.mkdir(train_nosmile_dir)

# Directory with our validation smile pictures
validation_smile_dir = os.path.join(validation_dir, 'have_mask')
os.mkdir(validation_smile_dir)

# Directory with our validation nosmile pictures
validation_nosmile_dir = os.path.join(validation_dir, 'no_mask')
os.mkdir(validation_nosmile_dir)

# Directory with our validation smile pictures
test_smile_dir = os.path.join(test_dir, 'have_mask')
os.mkdir(test_smile_dir)

# Directory with our validation nosmile pictures
test_nosmile_dir = os.path.join(test_dir, 'no_mask')
os.mkdir(test_nosmile_dir)

 

3、记数，统计一下各个文件中照片的数量

print('total training smile images:', len(os.listdir(train_smile_dir)))
print('total training nosmile images:', len(os.listdir(train_nosmile_dir)))
print('total validation smile images:', len(os.listdir(validation_smile_dir)))
print('total validation nosmile images:', len(os.listdir(validation_nosmile_dir)))
print('total test smile images:', len(os.listdir(test_smile_dir)))
print('total test nosmile images:', len(os.listdir(test_nosmile_dir)))

 4、构建小型卷积网络

我们已经为MNIST构建了一个小型卷积网，所以您应该熟悉它们。我们将重用相同的通用结构:我们的卷积网将是一个交替的Conv2D(激活relu)和MaxPooling2D层的堆栈。然而，由于我们处理的是更大的图像和更复杂的问题，因此我们将使我们的网络相应地更大:它将有一个更多的Conv2D + MaxPooling2D阶段。这样既可以扩大网络的容量，又可以进一步缩小特征图的大小，这样当我们到达平坦层时，特征图就不会太大。在这里，由于我们从大小为150x150的输入开始(有点随意的选择)，我们在Flatten层之前得到大小为7x7的feature map。

注意：feature map的深度在网络中逐渐增加(从32到128)，而feature map的大小在减少(从148x148到7x7)。这是你会在几乎所有convnets中看到的模式。由于我们解决的是一个二元分类问题，我们用一个单一单元(一个大小为1的稠密层)和一个s型激活来结束网络。这个单元将对网络正在查看一个类或另一个类的概率进行编码。

from keras import layers
from keras import models

model = models.Sequential()
model.add(layers.Conv2D(32, (3, 3), activation='relu',
                        input_shape=(150, 150, 3)))
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, 2)))
model.add(layers.Conv2D(128, (3, 3), activation='relu'))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Flatten())
model.add(layers.Dense(512, activation='relu'))
model.add(layers.Dense(1, activation='sigmoid'))

5、输出模型各层的参数状况

model.summary()

 6、告知训练时用的优化器、损失函数和准确率评测标准

from keras import optimizers

model.compile(loss='binary_crossentropy',
              optimizer=optimizers.RMSprop(lr=1e-4),
              metrics=['acc'])

7、数据预处理

在将数据输入到我们的网络之前，应该将数据格式化为经过适当预处理的浮点张量。目前，我们的数据以JPEG文件的形式保存在硬盘上，因此将其导入网络的步骤大致如下:

• 读取图片文件
• 解码JPEG内容到RBG像素网格
• 把它们转换成浮点张量
• 将像素值(从0到255)缩放到[0,1]区间

from keras.preprocessing.image import ImageDataGenerator

# All images will be rescaled by 1./255
train_datagen = ImageDataGenerator(rescale=1./255)
test_datagen = ImageDataGenerator(rescale=1./255)

train_generator = train_datagen.flow_from_directory(
        # This is the target directory
        train_dir,
        # All images will be resized to 150x150
        target_size=(150, 150),
        batch_size=20,
        # Since we use binary_crossentropy loss, we need binary labels
        class_mode='binary')

validation_generator = test_datagen.flow_from_directory(
        validation_dir,
        target_size=(150, 150),
        batch_size=20,
        class_mode='binary')

让我们看看其中一个生成器的输出：它生成150×150 RGB图像的批次(Shape(20,150,150，3))和二进制标签(Shape(20，))。20是每批样品的数量(批次大小)。注意，生成器无限期地生成这些批：它只是无休止地循环目标文件夹中的图像。因此，我们需要在某个点中断迭代循环。

for data_batch, labels_batch in train_generator:
    print('data batch shape:', data_batch.shape)
    print('labels batch shape:', labels_batch.shape)
    break

 使用生成器使我们的模型适合于数据

history = model.fit_generator(
      train_generator,
      steps_per_epoch=100,
      epochs=50,
      validation_data=validation_generator,
      validation_steps=50)

 保存模型

model.save('F:\PyCharm\facemask\mask_and_nomask\test\mask_and_nomask.h5')

在训练和验证数据上绘制模型的损失和准确性

import matplotlib.pyplot as plt

acc = history.history['acc']
val_acc = history.history['val_acc']
loss = history.history['loss']
val_loss = history.history['val_loss']

epochs = range(len(acc))

plt.plot(epochs, acc, 'bo', label='Training acc')
plt.plot(epochs, val_acc, 'b', label='Validation acc')
plt.title('Training and validation accuracy')
plt.legend()

plt.figure()

plt.plot(epochs, loss, 'bo', label='Training loss')
plt.plot(epochs, val_loss, 'b', label='Validation loss')
plt.title('Training and validation loss')
plt.legend()
plt.show()

 8、数据准确性提高

datagen = ImageDataGenerator(
      rotation_range=40,
      width_shift_range=0.2,
      height_shift_range=0.2,
      shear_range=0.2,
      zoom_range=0.2,
      horizontal_flip=True,
      fill_mode='nearest')

查看提高后的图像

# This is module with image preprocessing utilities
from keras.preprocessing import image

fnames = [os.path.join(train_smile_dir, fname) for fname in os.listdir(train_smile_dir)]

# We pick one image to "augment"
img_path = fnames[3]

# Read the image and resize it
img = image.load_img(img_path, target_size=(150, 150))

# Convert it to a Numpy array with shape (150, 150, 3)
x = image.img_to_array(img)

# Reshape it to (1, 150, 150, 3)
x = x.reshape((1,) + x.shape)

# The .flow() command below generates batches of randomly transformed images.
# It will loop indefinitely, so we need to `break` the loop at some point!
i = 0
for batch in datagen.flow(x, batch_size=1):
    plt.figure(i)
    imgplot = plt.imshow(image.array_to_img(batch[0]))
    i += 1
    if i % 4 == 0:
        break

plt.show()

如果我们使用这种数据增加配置训练一个新的网络，我们的网络将永远不会看到两次相同的输入。然而，它看到的输入仍然是高度相关的，因为它们来自少量的原始图像——我们不能产生新的信息，我们只能混合现有的信息。因此，这可能还不足以完全消除过度拟合。

为了进一步对抗过拟合，我们还将在我们的模型中增加一个Dropout层，就在密集连接分类器之前：

model = models.Sequential()
model.add(layers.Conv2D(32, (3, 3), activation='relu',
                        input_shape=(150, 150, 3)))
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, 2)))
model.add(layers.Conv2D(128, (3, 3), activation='relu'))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Flatten())
model.add(layers.Dropout(0.5))
model.add(layers.Dense(512, activation='relu'))
model.add(layers.Dense(1, activation='sigmoid'))

model.compile(loss='binary_crossentropy',
              optimizer=optimizers.RMSprop(lr=1e-4),
              metrics=['acc'])

用数据增强和退出来训练我们的网络:

train_datagen = ImageDataGenerator(
    rescale=1./255,
    rotation_range=40,
    width_shift_range=0.2,
    height_shift_range=0.2,
    shear_range=0.2,
    zoom_range=0.2,
    horizontal_flip=True,)

# Note that the validation data should not be augmented!
test_datagen = ImageDataGenerator(rescale=1./255)

train_generator = train_datagen.flow_from_directory(
        # This is the target directory
        train_dir,
        # All images will be resized to 150x150
        target_size=(150, 150),
        batch_size=32,
        # Since we use binary_crossentropy loss, we need binary labels
        class_mode='binary')

validation_generator = test_datagen.flow_from_directory(
        validation_dir,
        target_size=(150, 150),
        batch_size=32,
        class_mode='binary')

history = model.fit_generator(
      train_generator,
      steps_per_epoch=100,
      epochs=150,
      validation_data=validation_generator,
      validation_steps=50)

这里程序会跑很久，我跑了几个小时，用GPU跑会快很多很多。

保存模型在convnet可视化部分使用：

model.save('F:\PyCharm\facemask\mask_and_nomask\test\mask_and_nomask.h5')

看一下结果：（挺好的）

acc = history.history['acc']
val_acc = history.history['val_acc']
loss = history.history['loss']
val_loss = history.history['val_loss']

epochs = range(len(acc))

plt.plot(epochs, acc, 'bo', label='Training acc')
plt.plot(epochs, val_acc, 'b', label='Validation acc')
plt.title('Training and validation accuracy')
plt.legend()

plt.figure()

plt.plot(epochs, loss, 'bo', label='Training loss')
plt.plot(epochs, val_loss, 'b', label='Validation loss')
plt.title('Training and validation loss')
plt.legend()
plt.show()

8、优化提高笑脸图像分类模型精度
构建卷积网络

from keras import layers
from keras import models
from keras import optimizers
model = models.Sequential()
#输入图片大小是150*150 3表示图片像素用(R,G,B)表示
model.add(layers.Conv2D(32, (3,3), activation='relu', input_shape=(150 , 150, 3)))
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((2,2)))
model.add(layers.Conv2D(128, (3,3), activation='relu'))
model.add(layers.MaxPooling2D((2,2)))
model.add(layers.Flatten())
model.add(layers.Dense(512, activation='relu'))
model.add(layers.Dense(1, activation='sigmoid'))
model.compile(loss='binary_crossentropy', optimizer=optimizers.RMSprop(lr=1e-4),
             metrics=['acc'])
model.summary()

 三、代码测试

import cv2
from keras.preprocessing import image
from keras.models import load_model
import numpy as np
import dlib
from PIL import Image
model = load_model('mask_and_nomask.h5')
detector = dlib.get_frontal_face_detector()
video=cv2.VideoCapture(0)
font = cv2.FONT_HERSHEY_SIMPLEX
def rec(img):
    gray=cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
    dets=detector(gray,1)
    if dets is not None:
        for face in dets:
            left=face.left()
            top=face.top()
            right=face.right()
            bottom=face.bottom()
            cv2.rectangle(img,(left,top),(right,bottom),(0,255,0),2)
            img1=cv2.resize(img[top:bottom,left:right],dsize=(150,150))
            img1=cv2.cvtColor(img1,cv2.COLOR_BGR2RGB)
            img1 = np.array(img1)/255.
            img_tensor = img1.reshape(-1,150,150,3)
            prediction =model.predict(img_tensor)    
            print(prediction)
            if prediction[0][0]>0.5:
                result='nomask'
            else:
                result='mask'
            cv2.putText(img, result, (left,top), font, 2, (0, 255, 0), 2, cv2.LINE_AA)
        cv2.imshow('mask detector', img)
while video.isOpened():
    res, img_rd = video.read()
    if not res:
        break
    rec(img_rd)
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break
video.release()
cv2.destroyAllWindows()

这里就不上运行结果了。

本文转载：https://blog.csdn.net/weixin_45137708/article/details/107142706