This is an implementation of 3000fps.

Before you download, copy, or view sources file of this project, you must obeyed the following regulations,

1) Using for personal applications or study are freedom.
2) Using for bussiness must ask us for agreements.

This project built on ubuntu 14.04, depending on opencv-2.4.x. If your environment are linux and you opencv was installded under /usr/local. You can use make in the root directory to build this project. There will be one excutable file in directory bin.

bin/align  show the effection, usage: ./align [image list] [out dir]

bin/align_model.dat the align model 
bin/detect_model.dat the detect model 

The algorithm in this project to detect face is 3000fps in [1].

If you like this project and want to help us, you may joint or denote us.

• Copyright: Hangzhou Qiantu Technology Copyright
• email: [email protected]
• website: http:https://www.qiantuai.com
• qq群： 535810799

If you want more help, you may send email to us.

• [1] Face Alignment at 3000 FPS via Regressing Local Binary Features, Shaoqing Ren, Xudong Cao, Yichen Wei, Jian Sun
• [2] Supervised Descent Method and its Applications to Face Alignment, Xuehan Xiong, Fernando De la Torre
• [3] Face Alignment by Explicit Shape Regression. Xudong Cao, Yichen Wei, Fang Wen, Jian Sun