Portraitizer
Creates portraits in a hand-drawn style using DeepLab, Neural Style Transfer, and my own original art
Set Up
In order for the model to run smoothly, a GPU is recommended.
Download the VGG-19 model weights (see the "VGG-VD models from the Very Deep Convolutional Networks for Large-Scale Visual Recognition project" section).
After downloading, copy the weights file imagenet-vgg-verydeep-19.mat
to the project directory.
Usage
-
Create directories image_input, image_output, image2convert,and styles in project directory
-
Place any image you want to the styles folder to be used to transfer onto the content image
-
Outputs can then be found in the image_output folder
Instructions on how to use the two Jupyter Notebooks can be found within the notebooks themselves
Results
Here is one example:
More results can be found here
Also contained in this repository:
Credits
DeepLab documentation can be found here
Neural style transfer uses cysmith's implemenation which can be found here