This project is a C++ reimplementation of face alignment in 3000fps in the CVPR 2014 paper: Face Alignment at 3000 FPS via Regressing Local Binary Features. .
I modify my code to support openMP. You can use it in GCC(Linux) or in VS (Windows).
If you use it in Linux, you should comment or uncomment FIND_PACKAGE( OpenMP REQUIRED)
in CmakeLists.txt.
If you use it in Windows, you can directly use it.
I add a VS project.
Download datasets and get Path_Images.txt as jwyang/face-alignment.
To compiler the program: go to folder build
and
cmake .
make
To train a new model: set global parameters, model path, train database name in LBF.cpp
. Use "LBF.out TrainModel"
.
To test a model on dataset: set model path, test dataset name in LBF.cpp
. Use "LBF.out TestModel"
.
###Model I have trained a model on AFW, HELEN,LFPW dataset. You can download it from here or google drive.
How to get the bounding box of image ?
I use the face detector in OpenCV to get the bounding box.You can use any detector to get the bounding box but you must provide a bounding box of similar measure with the training data.
How about the liblinear?
I add the liblinear source code as the project code. So you can directly compiler this project and don't need to consider to compiler this library.
If you have any question, you can create an issue
on GitHub.
Or you can email yulequan@zju.edu.cn
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。