The application of a deep neural network is not only restricted to finding an object in an image (which we learned about in the previous chapters) – it can also be used to segment images into spatial regions, thereby producing artificial images and transferring style from one image to another.
In this chapter, we will use TensorFlow Colab to perform all these tasks. Semantic segmentation predicts whether each pixel of an image belongs to a certain class. It is a useful technique for image overlaying. You will learn about TensorFlow DeepLab so that you can perform semantic segmentation on images. Deep Convolutional Generative Adversarial Networks (DCGANs) are powerful tools that are used to produce artificial images such as human faces and handwritten digits. They can also be used for image inpainting. We will also discuss...
