Skip to content

wangmoyu/Neural-Style-Demonstration

Repository files navigation

Neural-Style-Demonstration

Introduction

This is my post-graduate project in University College Cork.

In this project, I fellow an article by Gatys to implement the neural style which is a method to achieve artist painting style transfer and also tried a domain transfer method to represent style.

This project is based on Tensorflow with Python under Mac OS. If you are using Window Tensorflow to run this code directly, there could be a transcoding problem.

All the involved techniques have been introduced in my project report.

Requirement

This implement is based on a Convolution Neural Network.
A VGG-19 network should be include in the file.

Related Work

If you are interested in neural style, here are some useful articles that could be helpful.
Neural Style Transfer: A Review
Neural-Style-Transfer Papers and related work

Results

Running python style_transfer.py --help
to view all the arguments.


Running python style_transfer.py --content content image path --styles style images path --output output path
to quickly start a style transfer process.

A simple result:



Releases

No releases published

Packages

No packages published

Languages