A simple walk-through keras implementation to learn style transfer using CNN/VGG19.
It illustrates 5 examples and for each there is separate IPython Notebook:
As a part of this tutorial we will learn about
-
single style transfer
-
multiple style transfer
-
masked style transfer
-
color preservation
For understanding the basics of neural style transfer, go through this Neural-Style-Transfer.pdf
Kindly look for the outputs of our little style transfer experiment here: pictorial-summary-of-all-5-examples
References:
A Neural Algorithm of Artistic Style
Improving the Neural Algorithm of Artistic Style
Show, Divide and Neural: Weighted Style Transfer
Preserving Color in Neural Artistic Style Transfer
Very Deep Convolutional Networks for Large-Scale Image Recognition by Simonyan and Zisserman (2015)
Incorporating Long-Range Consistency in CNN-based Texture Generation