The CE-CLM algorithm is now a core part of OpenFace 2.0. Please visit OpenFace Github. CE-CLM will only be maintained as part of OpenFace 2.0 and is not maintained here.
The code for CE-CLM published in CVPR 2017. If you use this software for academic purposes please cite the following:
Zadeh, A., Lim, Y. C., Baltrušaitis, T., & Morency, L. P. (2017). Convolutional experts constrained local model for 3d facial landmark detection. In Proceedings of the IEEE International Conference on Computer Vision Workshops (Vol. 7).
Binaries for Windows are available here:
- 32-bit: (https://github.com/A2Zadeh/CE-CLM/releases/download/v0.2.3/OpenFace_CE-CLM_win32.zip)
- 64-bit: (https://github.com/A2Zadeh/CE-CLM/releases/download/v0.2.3/OpenFace_CE-CLM_win_x64.zip)
For the binaries to work you need to have Visual Studio 2015 installed or need to install the 64-bit Visual C++ redistributable package.
Explanation of how to use the command line binaries can be found here.
Note: As the CE-CLM model files are large (total ~1GB), we do not include the files in this release. When the programs are first run, they will offer to download the model files from (here)[http://sorena.multicomp.cs.cmu.edu/cen_patches/]. If you wish to manually download the model files, save the .dat files to model/patch_experts.
To generate the patches, follow the instructions in patch_generation/.
First ensure that you have generated the patches using the scripts in patch_generation/. The CEN training script is in cen_training/.
This distribution has been tested with Windows 10, using Visual Studio 2015. To compile, open "OpenFace.sln" using Visual Studio 2015 and compile the code. The project was built and tested on Visual Studio 2015, in Windows 10.