ncjie/sdm

Supervised Descent Method Apply to Face Alignment, and Head Pose Estimation with Linear Regression. It is cross-platfrom, easily compile in windows, ubuntu, even in Android & iOS.

SDM and Head Pose Estimation

example

git clone https://github.com/RoboPai/sdm.git

cd sdm/example/
./Roboman-SDM-TEST.exe

You will look like as follows:

How to compile on Windows？

Make sure you have installed the cmake-gui & Qt5.3.2 (mingw will be needed).

mkdir build
cd build
cmake-gui
mingw32-make

How to compile on Ubuntu？

Maybe you should modify the CMakeLists.txt about OpenCV libs

mkdir build
cd build
cmake ..
make
cp Roboman-SDM-TEST ../example/
cd ../example/
./Roboman-SDM-TEST

How to compile on iOS？

1.Copy "haar_roboman_ff_alt2.xml" & "roboman-landmark-model.bin" to your iOS Project.

2.Add the OpenCV2.framework to you iOS Project, make sure you can easily read camera and show frame with OpenCV.

3.Copy the include folder under src folder to you iOS Project source code, just copy, not anything else.

4.And then "#include "include/ldmarkmodel.h"" in your .mm file, make sure not in .h file, use it like test_model.cpp, Compile & enjoy it.

5.Make sure iOS project Build Settings option "Enable Bitcode" is "NO" (above XCode7).

How to compile on Android？

According to Android-cmake-Tool.