10. Data Preprocessing

Lambda Layers

In Keras, lambda layers can be used to create arbitrary functions that operate on each image as it passes through the layer.

(Using Keras v1.2.1 from the starter kit? See archived documentation over lambda layers here. Instead of keras.layers.Lambda as in Keras v2, you'll use keras.layers.core.Lambda.)

In this project, a lambda layer is a convenient way to parallelize image normalization. The lambda layer will also ensure that the model will normalize input images when making predictions in drive.py.

That lambda layer could take each pixel in an image and run it through the formulas:

pixel_normalized = pixel / 255

pixel_mean_centered = pixel_normalized - 0.5

A lambda layer will look something like:

Lambda(lambda x: (x / 255.0) - 0.5)

Below is some example code for how a lambda layer can be used.

from keras.models import Sequential, Model
from keras.layers import Lambda

# set up lambda layer
model = Sequential()
model.add(Lambda(lambda x: (x / 255.0) - 0.5, input_shape=(160,320,3)))
...