In this chapter, we'll discuss the classical convolutional neural networks (CNN) that greatly influence computer vision. We will present two advanced architectures: the residual neural network, which solves the problem of training deep neural networks, and the inception network, or GoogLeNet, which dramatically improves computation efficiency through the use of one-by-one convolution.

Next up, we'll gain insights into transfer learning and explore several ways to use it to train neural networks efficiently. Finally, we'll use transfer learning techniques and the VGG-16 architecture to build an animal recognizer Java application and run it through a graphical user interface with several examples.

The following topics will be covered in this chapter:

• Working with classical networks
• Using residual networks for image recognition
• The power of 1 x 1 convolutions and the inception network
• Applying transfer learning
• Building an animal image classification application