Table of content

1.Collect Data

Data Preprocessing
Visualizing Data
Build Model
Train model
Evaluating Trained Model
Save a trained model
Predict on custom data
Realtime detection

1.Collect data

In every machine learning, problem data is the main.
here, data is collected from Kaggle.
data is a subset of this dataset you can download it from here

2.Data Preprocessing

Before we process data, first structure our data in the right folder.
For this we have two option:

Use TensorFlow image_dataset_from_directory.
load CSV file.

Here we choose to load from a CSV file.
For that, we change the image name to withmask and withoutmask.

Withmask

# importing os module 
import os 

# Function to rename multiple files 
def main():
    for count, filename in enumerate(os.listdir("DATASET/")):
        dst ="withmask." + str(count) + ".jpeg"
        src ='DATASET/'+ filename 
        dst ='DATASET/'+ dst 
        # rename() function will
        # rename all the files 
        os.rename(src, dst) 

# Driver Code 
if __name__ == '__main__':
    # Calling main() function 
    main()

Withoutmask

# importing os module 
import os 

# Function to rename multiple files 
def main():
    for count, filename in enumerate(os.listdir("New folder/")):
        dst ="withoutmask." + str(count) + ".jpeg"
        src ='New folder/'+ filename 
        dst ='New folder/'+ dst 
        # rename() function will
        # rename all the files 
        os.rename(src, dst) 

# Driver Code 
if __name__ == '__main__':
    # Calling main() function 
    main()

Now move all images into one folder and create a pandas data frame

Generate DataFrame

import pandas as pd
filenames=os.listdir("FULL_DATA/")
categories=[]
for f_name in filenames:
    category=f_name.split('.')[0]
    if category=='withmask':
        categories.append('withmask')
    else:
        categories.append('withoutmask')
df=pd.DataFrame({
    'filename':filenames,
    'labels':categories
})

Save dataFrame into CSV file.

Now data is in the right structure we can load data

Read CSV file
Shuffle DataFrame with sample(frac=1)
Turn label into an Array of Boolean
Create a validation set with train_test_split
Turning images into Tensor

# Define image size
IMG_SIZE = 224

# Function
def process_image(image_path, image_size=IMG_SIZE):
  """
  Takes an image file path and turns the image into a Tensor.
  """
  # Read in an image file
  image = tf.io.read_file(image_path)
  # Turn the jpg image into numerical Tensor with 3 colour channel(RGB)
  image = tf.image.decode_jpeg(image,channels=3)
  # Convert the color channel values to (0-1) values
  image = tf.image.convert_image_dtype(image,tf.float32)
  # Resize the image to (224,224)
  image = tf.image.resize(image, size=[image_size,image_size])

  return image

Turning data into Batches

# Create a function to return a tuple (image, label)
def get_image_lable(image_path,label):
  """
  Takes an image file path name and the label,
  processes the image and return a tuple (image, label).
  """
  image = process_image(image_path)

  return image, label

# Define the batch size
BATCH_SIZE = 32

# Function to convert data into batches
def create_data_batches(X,y=None, batch_size=BATCH_SIZE,valid_data=False):
  """
  Creates batches of data of image (X) and label (y) pairs.
  Shuffle the data if it's training data but doesn't shuffle if it's validation data.
  """
  # If data is valid dataset (NO SHUFFLE)
  if valid_data:
    print("Creating valid data batches.........")
    data = tf.data.Dataset.from_tensor_slices((tf.constant(X),
                                               tf.constant(y)))
    data_batch = data.map(get_image_lable).batch(batch_size)
    return data_batch

  else:
    print("Creating train data batches.........")
    # Turn filepaths and labels into Tensors
    data = tf.data.Dataset.from_tensor_slices((tf.constant(X),
                                               tf.constant(y)))
    # Shuffling pathname and labels before mapping image processor fun
    data = data.shuffle(buffer_size=len(X))
    data_batch = data.map(get_image_lable).batch(batch_size)

    return data_batch

3.Visulizing Data

import matplotlib.pyplot as plt
# Create fun for viewing in a data batch
def show_images(images, labels):
  """
  Displays a plot of 25 images and their labels from a data batch.
  """
  plt.figure(figsize=(20, 20))
  for i in range(25):
    # Subplot
    ax = plt.subplot(5,5,i+1)
    plt.imshow(images[i])
    plt.title(unique_category[labels[i].argmax()])
    plt.axis("Off")

Call this fun

For Train data

train_images, train_labels = next(train_data.as_numpy_iterator())
show_images(train_images,train_labels)

For valid Data

val_images, val_labels = next(val_data.as_numpy_iterator())
show_images(val_images, val_labels)

4.Building a model

Here we can use the TensorFlow hub for pre-trained models.
For this task, we use MobileNet V2 which is a small model.

Set input_shape = [none, 224,224,3]
Set output_shape = 2
Use Sequential model from tf.keras

# Create a fun to build a keras model
def create_model(input_shape=INPUT_SHAPE,output_shape=OUTPUT_SHAPE, model_url=MODEL_URL):
  print("Building model with:", model_url)

  # Setup the model
  model = tf.keras.Sequential([
                               hub.KerasLayer(model_url),
                               tf.keras.layers.Dense(units=output_shape, 
                                                     activation="softmax")
  ])

  # Compile the model
  model.compile(
      loss = tf.keras.losses.BinaryCrossentropy(),
      optimizer = tf.keras.optimizers.Adam(),
      metrics = ["accuracy"]
  )

  # Build the model
  model.build(input_shape)

  return model

5. Train a model

Train a model on train_data and valid_data for 25 EPOCHS

Also, add an Early stopping callback

model = create_model()
model.summary()

With this model, val_loss is 0.0096 and Accuracy is almost 99.99 %

6.Evaluating prediction

Using model.predict() on val_data model return NumPy array of shape (_ , 2)

7.Saving and reloading a trained model

Save a trained model using save_model from keras.

Loading a model is a bit different from regular load_model

model = load_model(
    'model/model.h5', custom_objects={"KerasLayer": hub.KerasLayer})

here we have to provide custom_objects={“KerasLayer”: hub.KerasLayer} in load_model function alongside model_path.

8.Predict on custom data

Before predicting the new data make sure it is in the right shape as well as the right size.

def test_data(path):
  demo = imread(path)
  demo = tf.image.convert_image_dtype(demo,tf.float32)
  demo = tf.image.resize(demo,size=[224,224])
  demo = np.expand_dims(demo,axis=0)

  pred = model.predict(demo)
  result = unique_category[np.argmax(pred)]

  return result