Understanding data

  • Chest X-ray images (anterior-posterior) taken of pediatric patients, aged one to five years old

  • from Guangzhou Women and Children’s Medical Center, Guangzhou.

  • physicians labelled images with pneumonia, part of normal diagnosis

Prepare data

from tensorflow.keras.preprocessing import image_dataset_from_directory

train_batches = image_dataset_from_directory(train_path)

val_batches = image_dataset_from_directory(test_path)

#Found 5216 files belonging to 2 classes.
#Found 624 files belonging to 2 classes.

  • image_dataset_from_directory function reads data in format:

  • main_directory/

    • class_a/
      • a_image_1.jpg
      • a_image_2.jpg
    • class_b/
      • b_image_1.jpg
      • b_image_2.jpg

Model construction

Rescaling Layer

  • standarising layer : CNN converges faster on [0 to 1] data than on [0 to 255]

  • tf.keras.layers.Rescaling(1./255) added into model to standarise pixel values

model = Sequential()
model.add(Rescaling(1./255))
model.add(Conv2D(32,(3,3),strides=(1, 1),activation='relu',padding='same'))
model.add(MaxPool2D(pool_size=(2,2)))
model.add(Conv2D(64,(3,3),strides=(1, 1) ,padding='same',activation='relu'))
model.add(MaxPool2D(pool_size=(2,2)))
model.add(Conv2D(128,(3,3),strides=(1, 1),padding='same', activation='relu'))
model.add(MaxPool2D(pool_size=(2,2)))
model.add(Conv2D(256,(3,3),strides=(1, 1),padding='same', activation='relu'))
model.add(MaxPool2D(pool_size=(2,2)))
model.add(Flatten())

model.add(Dense(128, activation='relu'))
model.add(Dense(64, activation='relu'))
model.add(Dense(2))

model.compile(
  optimizer='adam',
  loss=tf.losses.SparseCategoricalCrossentropy(from_logits=True),
  metrics=['accuracy'])

earlystopping = callbacks.EarlyStopping(monitor ="val_loss", 
                                        mode ="min", patience = 3, 
                                        restore_best_weights = True)

history = model.fit(
  train_batches,
  validation_data=val_batches,
  epochs=50,
  callbacks=[earlystopping]
)

Model Evaluation

alt

alt