Hands-On Automated Machine Learning

The convolution layer

This is the layer where the convolution operation takes place between the input image and the filters that we just discussed. This step is carried out to reduce the overall size of the image, so that it is easier to process the image in the next layers. Let's understand the function of this layer with a simple question, how do we identify a dog or a cat? It comes naturally to us within a second when we see them. We don't analyze all their features to determine whether a dog is a dog or a cat is a cat.

We recognize the important features, such as their eyes, ear, or tail, and then identify the creature. This is what is also done in a convolution layer. In this layer, the important features are identified and the rest ...