So in my journey of machine learning i recently stumbled upon a bunch of really cool looking images generated by something called a "deep dream".
Deep Dream is a computer vision program created by google's engineer "Alexander Mordvintsev" that uses Convolutional neural networks to find and enhance image features to make them look creepy and phenomenal.
It basically makes image have "more" of a certain feature.
With the research I've done and I'm no machine learning face with tongue, What i was able to understand is that this program focuses on finding certain features in an image and "multiply" it ...i guess.. rather than focusing on the networks "Cost Function" like most neural networks do.
More resources can be found here:
deepdreaming development website