Arbitrary Style Transfer with Deep Feature Reshuffle

The major contributors of this repository include Shuyang Gu, Congliang Chen, Jing Liao, Lu Yuan at Microsoft Research.

Introduction

Deep Feature Reshuffle is a technique to using reshuffling deep features of style image for arbitrary style transfer. It connects both global and local style constrain respectively used by most parametric and non-parametric neural style transfer methods.

Disclaimer

This is an official C++ combined with CUDA implementation of "Arbitrary Style Transfer with Deep Feature Reshuffle". It is worth noticing that:

• Our codes are based on Caffe.
• Our codes only have been tested on Windows 10 and Windows Server 2012 R2 with CUDA 8 or 7.5.
• Our codes only have been tested on several Nvidia GPU: Titan X, Titan Z, K40, GTX770.

License

© Microsoft, 2018. Licensed under a MIT license.

Citation

If you find Deep Feature Reshuffle helpful for your research, please consider citing:

@inproceedings{gu2018arbitrary, 
title={Arbitrary Style Transfer with Deep Feature Reshuffle}, 
author={Gu, Shuyang and Chen, Congliang and Liao, Jing and Yuan, Lu}, 
booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition}, 
pages={8222--8231}, 
year={2018} 
} 

Getting Started

Prerequisite

• Windows 7/8/10
• CUDA 8 or 7.5
• Visual Studio 2013

Build

• Build Caffe at first. Just download and follow the tutorial here.
• Put style_feature_reshuflle under windows/
• Edit style_feature_reshuffle.vcxproj under style_feature_reshuffle to make the CUDA version in it match yours .
• Open solution Caffe and add style_feature_reshuffle project.
• Build project style_feature_reshuffle.

Running code

-style_feature_reshuffle content_image_name style_image_name output_image_name gpu_id