Home

Awesome

<table width="100%" border="0" cellspacing="15" cellpadding="0"> <tbody> <tr> <td> <b>Multi-style Generative Network for Real-time Transfer</b> [<a href="https://arxiv.org/pdf/1703.06953.pdf">arXiv</a>] [<a href="http://computervisionrutgers.github.io/MSG-Net/">project</a>] <br> <a href="http://hangzh.com/">Hang Zhang</a>, <a href="http://eceweb1.rutgers.edu/vision/dana.html">Kristin Dana</a> <pre> @article{zhang2017multistyle, title={Multi-style Generative Network for Real-time Transfer}, author={Zhang, Hang and Dana, Kristin}, journal={arXiv preprint arXiv:1703.06953}, year={2017} } </pre> </td> <td width="440"><a><img src ="https://raw.githubusercontent.com/zhanghang1989/MSG-Net/master/images/figure1.jpg" width="420px" border="1"></a></td> </tr> </tbody> </table>

Installation

We also provide PyTorch implementation and MXNet implementation. Please install Torch7 with cuda and cudnn support. The code has been tested on Ubuntu 16.04 with Titan X Pascal and Maxwell.

luarocks install https://raw.githubusercontent.com/zhanghang1989/MSG-Net/master/texture-scm-1.rockspec

Test and Demo

  1. Clone the repo and download pre-trained models
    git clone git@github.com:zhanghang1989/MSG-Net.git
    cd MSG-Net/experiments
    bash models/download_models.sh 
    
  2. Web Camera Demo
    qlua webcam.lua
    
  3. Test on Image
    th test.lua -input_image images/content/venice-boat.jpg -image_size 1024
    eog stylized
    

<img src ="images/stylized/1.jpg" width="260px" /> <img src ="images/stylized/2.jpg" width="260px" /> <img src ="images/stylized/3.jpg" width="260px" /> <img src ="images/stylized/4.jpg" width="260px" /> <img src ="images/stylized/5.jpg" width="260px" /> <img src ="images/stylized/6.jpg" width="260px" /> <img src ="images/stylized/7.jpg" width="260px" /> <img src ="images/stylized/8.jpg" width="260px" /> <img src ="images/stylized/9.jpg" width="260px" />

[More Example Results]

Train Your Own Model

Please follow this tutorial to train a new model.

Release Timeline

Acknowledgement

The code benefits from outstanding prior work and their implementations including: