Mastering Computer Vision with TensorFlow 2.x

Summary

In this chapter, we learned how to use TensorFlow 2.0 and Google Colab to train a neural network to perform a lot of complex image manipulation tasks such as semantic segmentation, image inpainting, generating an artificial image, and neural style transfer. We learned about the functionality of generative and discriminator networks, as well as how neural networks can be trained simultaneously in a balanced fashion to create fake output images. We also learned how to use Atrous convolution, spatial pooling, and encoder-decoder networks to develop semantic segmentation. Finally, we used Google Colab to train a neural network to perform neural style transfer.

In the next chapter, we will use a neural network for activity recognition. ...