Description
This work is used the model of mtcnn which trained by pytorch and processed by C++ to realize a face detect.
Prerequisites
You need to know loading a torchScript model in C++ ,you can learn from the website of pytorch. You also need to know how to use opencv.
Dependencies
C++
opencv
libtorch
Ubuntu/windows
Cuda 9/10 it is depend on the version of libtorch
Ubuntu
If you want to test on Ubuntu. Just Install all the dependencies and use make to build the project.
windows
If you want to test on windows. You need to install cmake and vs2015/2017 .
create a directory named build .
The command of cmake is : cmake -DCMAKE_PREFIX_PATH=D:/opencv/build/x64/vc14/lib;D:/libtorch -DCMAKE_BUILD_TYPE=Release -G "Visual Studio 15 Win64 " ..