You’ve already seen how to extract color and shape features from an image. In the examples you’ve gone through, it was up to you to decide what features and filters were the most useful for grouping pixel data into similar clusters or classes. This is similar to how Convolutional Neural Networks or CNN’s learned to recognize patterns in images. For example, say we’re creating a CNN for image classification, this CNN should take in an image as input, and output a distribution of class scores from which we can get the predicted class for that image. This CNN is made up of a series of layers that learn to extract relevant features out of any image. The backbone of a CNN is the convolutional layer. A convolutional layer applies a series of different image filters, also known as convolutional kernels, to an input image. The resulting filtered images have different appearances. The filters may have extracted features like the edges of objects in that image or the colors that distinguish the different classes of images. As the CNN trains, it updates the weights that define the image filters in this convolutional layer using backpropagation. The end result is a classifier with convolutional layers that have learned to filter images to extract distinguishing features. In the next section, we’ll learn how convolutional neural networks learn to extract features from images, go over the types of layers that make up a complete convolutional neural network, and learn some techniques for visualizing the inner workings of a CNN.