TensorFlow Tutorial - DeepDream Algorithm

 

Notes by Magnus Erik Hvass Pedersen: https://github.com/Hvass-Labs/TensorFlow-Tutorials/blob/master/14_DeepDr...

We have seen in the previous tutorials how to use the gradient of a neural network to generate images. Tutorials #11 and #12 showed how to generate adversarial noise using the gradient. Tutorial #13 showed how to use the gradient for generating images that caused specific features inside the neural network to respond.

This tutorial uses a similar idea as the previous tutorials. But we will now use the gradient of the neural network to amplify patterns in the input image. This is commonly called the DeepDream algorithm, but there are actually many different variations of the technique.

This builds on the previous tutorials. You should be familiar with neural networks in general (e.g. Tutorial #01 and #02).

Resource Type: