Introduction

This chapter covers CNNs. CNNs use convolutional layers that are well-suited to extracting features from images. They use learning filters that correlate with the task at hand. Simply put, they are very good at finding patterns in images.

In the previous chapter, you explored regularization and hyperparameter tuning. You used L1 and L2 regularization and added dropout to a classification model to prevent overfitting on the connect-4 dataset.

You will now be shifting gears quite a bit as you dive into deep learning with CNNs. In this chapter, you will learn the fundamentals of how CNNs process image data and how to apply those concepts to your own image classification problem. This is truly where TensorFlow shines.