Summary

In this chapter, we have learned about generating images from the contours of an image using the Pix2Pix GAN. Further, we learned about the various loss functions in CycleGAN to convert images of one class to another. Next, we learned about how StyleGAN can be used to generate realistic faces and also copy the style from one image to another, depending on how the generator is trained. Finally, we learned about using the pre-trained SRGAN model to generate high-resolution images. All of these techniques lay a strong foundation as we advance to learn about more modern ways of transferring image attributes in Chapters 16 and 17.

In the next chapter, we will switch gears and learn about combining computer vision techniques with other prominent techniques in reinforcement learning.