In this chapter, we learned about computer vision and its association with deep learning. We explored a specific type of deep learning algorithm, CNNs, that is widely used in computer vision. We studied an open source deep learning framework called MXNet. After a detailed discussion of the MNIST dataset, we built models using various network architectures and successfully classified the handwritten digits in the MNIST dataset. At the end of the chapter, we delved into the concept of transfer learning and explored its association with computer vision. The last project we built in this chapter classified images using an Inception-BatchNorm pretrained model. 

In the next chapter, we will explore an unsupervised learning algorithm called the autoencoder neural network. I am really excited to implement a project to capture credit card fraud using autoencoders. Are you game? Let's go!