liuguoyou / fadenn

Source code of Robust Facial Alignment with Internal Denoising Auto-Encoder

FADeNN

This is the Repository of our paper of Robust Facial Alignment with Internal Denoising Auto-Encoder.

Requirements :

1. MTCNN requirements : https://github.com/DuinoDu/mtcnn
2. Tensorflow GPU : https://www.tensorflow.org/install/install_linux
3. Pytorch GPU : https://pytorch.org/
4. Other packages, can be installed by cloning my environment file on src : env.yml

This Repository hold the Facial Localiser from this repo : https://github.com/deckyal/RT

Example of the facial alignment

usage :

python DenoiseLocalise.py

Change : config.py :

1. img_name = 'indoor_120.png' #Your image name
2. trySynthetic = True/False #Whether to provide the demo on the synthetic or real data(300VW)
3. image_directory = 'ex_images' #the directory of the image
4. cl_type = 0#0,1,2 the type of denoising.

Citation :

D. Aspandi, O. Martinez, F. Sukno, X. Binefa, Robust facial align-ment with internal denoising auto-encoder, in: 2019 16th Confer-ence on Computer and Robot Vision (CRV), 2019, pp. 143–150.doi:10.1109/CRV.2019.00027.