How to feed Nifti Images in 3d CNN for classification?

Question

I have 142 Nifti CT images of the brain, I converted them from Dicom. Every NIfti file has the dimension of 512×512×40. My plan is to work with 3d Conv Neural Network for multi-class classification. How should I feed Nifti images in a 3d CNN?

Answer 1

If you wish to use TensorFlow, you might consider the folowing steps:

• Load your dataset using tf.data

train_loader = tf.data.Dataset.from_tensor_slices((x_train, y_train))
validation_loader = tf.data.Dataset.from_tensor_slices((x_val, y_val))

• Apply your preprocessing steps

   train_dataset = (
   train_loader.shuffle(len(x_train))
   .map(train_preprocessing)
   .batch(1)
   .prefetch(2))
  
  validation_dataset = (
       validation_loader.shuffle(len(x_val))
       .map(validation_preprocessing)
       .batch(1)
       .prefetch(2)
   )
  

Build your 3D CNN model:

def 3D_model(width= 512, height= 512, depth=40):
 
    inputs = keras.Input((width, height, depth, 1))

    x = layers.Conv3D(filters=84, kernel_size=3, activation="relu")(inputs)
    x = layers.MaxPool3D(pool_size=2,padding="same")(x)
    x = layers.BatchNormalization()(x)

    x = layers.Conv3D(filters=64, kernel_size=3, activation="relu")(x)
    x = layers.MaxPool3D(pool_size=2,padding="same")(x)
    x = layers.BatchNormalization()(x)
    outputs = layers.Dense(units=n_classes, activation="softmax")(x)
    model = keras.Model(inputs, outputs)
    return model
model = get_model(width=512, height=512, depth=40)

• Train your model:

3D_model.compile(..)
3D_model.fit(
    train_dataset,
    validation_data=validation_dataset,
    epochs=epochs,
    shuffle=True)

You can also refer to this example