awong1900 / Nonlinear_Face_3DMM

Source code for "Nonlinear 3D Face Morphable Model"

Nonlinear 3D Face Morphable Model

[Project page] [CVPR'18 Paper] [CVPR'19 Paper]

Library requirements

• Tensorflow

Data

Download following pre-processed training data (10GB) and unzip into ./data/300W_LP/

Filelist Images Textures Masks

Download following 3DMM definition and unzip into current folder (./) 3DMM_definition.zip

Compile the rendering layer - CUDA code

$ # Compile
$ cd TF_newop/
$ ./compile_op_v2_sz224.sh
$ # Run an example
$ python rendering_example.py

Currently the code is working but not optimal (i.e see line 139 of TF_newop/cuda_op_kernel_v2_sz224.cu.cc) also the image size is hard-coded. Any contribution is welcome!

Run the code

Pretraining

Finetunning

Citation

If you find this work useful, please cite our papers with the following bibtex:

@inproceedings{ tran2019towards, 
  author = { Luan Tran and Feng Liu and Xiaoming Liu },
  title = { Towards High-fidelity Nonlinear 3D Face Morphable Model },
  booktitle = { In Proceeding of IEEE Computer Vision and Pattern Recognition },
  address = { Long Beach, CA },
  month = { June },
  year = { 2019 },
}

@article{ tran2018on, 
  author = { Luan Tran and Xiaoming Liu },
  title = { On Learning 3D Face Morphable Model from In-the-wild Images },
  journal = { IEEE Transactions on Pattern Analysis and Machine Intelligence },
  month = { July },
  year = { 2019 },
}

@inproceedings{ tran2018nonlinear, 
  author = { Luan Tran and Xiaoming Liu },
  title = { Nonlinear 3D Face Morphable Model },
  booktitle = { IEEE Computer Vision and Pattern Recognition (CVPR) },
  address = { Salt Lake City, UT },
  month = { June },
  year = { 2018 },
}

Contacts

If you have any questions, feel free to drop an email to [e-mail].