marvis / ysqfastfd

A fast binary library for face detection and face landmark detection in images. No float point operations, especially suit for low cost ARM CPUs, The highest accuracy on FDDB among non deep learning methods

YSQ face detect algorithm demo on ARM (Linux) platform.

Performance on ARM platform

Test on Allwinner R40 (ARM32) and Rockchip RK3399 (ARM64)

Allwinner R40 A7 platform

Rockchip RK3399 A72 platform

Platform preparation :

• ARM32 platform hardware board
• Ubuntu 16.04 with GTK3 library
• USB camera (480P/720P), YUYV

Build

Check the gtk+ devel library is installed

sudo apt install build-essential vim-gtk libgtk-3-0 libgtk-3-dev libegl1-mesa-dev

Build the runtime shared libraries

cd YSQfastfd
make -j4

Copy libraries

sudo mkdir /usr/local/ysqfd
sudo cp 3rdlibs/libysqfd32.so /usr/local/ysqfd/.
sudo cp 3rdlibs/libysqfd64.so /usr/local/ysqfd/.
sudo cp middle/libaaid.so /usr/local/ysqfd/.
export LD_LIBRARY_PATH=/usr/local/ysqfd

run the demo

fddemo/ysqfddemo

NOTE : Type key 'q' to quit the display window.