Facial Expression Retargeting from Human to Avatar Made Easy
Dependency
OpenMesh 6.3
Keras 2.2.0
Tensorflow 1.14.0
Numpy 1.16.4
Scipy 1.2.0
pyigl
openmesh (python)
Pre-trained Model
Due to large file limitation on github, please download the pre-trained models from GDrive: https://drive.google.com/file/d/12UQFhxCibE0DWai6sVerzzQAQGwTDpbJ/view?usp=sharing
https://drive.google.com/file/d/1vuqaK54O3JkmrKwMCOqvSNP5axbzy0Ik/view?usp=sharing
After extracting the zip file, please copy the "model" folder to each branch, i.e., Mery_human_transfer, Chubby_human_transfer, Conan_human_transfer.
Testing
(1) Compute deformation representation feature (RIMD feature) -> ./RIMD_Reconstruct
(2) Concate DR feature into testing data -> See example in ./Mery_human_transfer/data/people/test_data.npy
(3) Run testing phase of each model -> For example: "cd Mery_human_transfer; python main.py -l -t -m triplemoji;"
Help
1). The libigl library has been replaced with a new version of python binding (beta version). However the newer version does not support well for this code. Please refer to the older version in "./libigl" and compile the pyigl.so library by: "cd libigl; mkdir build; cmake ..; make python -j4;".
2). To compile and run the RIMD_Reconstruct project, please first check the profile settings in "./RIMD_Reconstruct/RIMD.pro". Specifically, please replace the include path of OpenMesh library to your local one.
3). In each avatar filefolder, please re-compile the get_mesh_py module; For example, "cd ./Mery_human_transfer/src/; mkdir build; cd build; cmake ..; make -j4;".
Update 3). Please pre-compile the get_mesh_py and get_dr_py module at the root directory (cd get_mesh_py; mkdir build; cd build; cmake ..; make -j4;) and then copy the compiled libraries to each workspace folder (e.g. Mery_human_transfer).
Acknowledgement
The repository is partially built on get_dr_py and DR-Learning-for-3D-Face.