wincle / NPD

C++ implementation of Fast-and-Accurate-Unconstrained-Face-Detector

NPD

The C++ implementation of A Fast and Accurate Unconstrained Face Detector.

The result is trained by 200k pos data and the template is 24*24, stages number is 620, model size is 540kb.

minFaceSize	speed(ms)	cores
80*80	30	1
24x24	500	1
24*24	60	16

the detection result is test on FDDB data set (average 400*400)

NOTICE

The "1226model" is dump from matlab code which is from References, this model has 1226 stages.

You must change the code in detection/LearnGAB.cpp:58-64. Because the difference between matlab and OpenCV. You should also change the coefficient in detection/LearnGAB.cpp:262-265 to fit the model.

How to use

• you should mkdir data first

In data folder, you should creat two file named FaceDB.txt and NonFaceDB.txt.

FaceDB.txt
../data/face/00001.jpg x1 y1 x2 y2
../data/face/00002.jpg x1 y1 x2 y2
....
....

NonfaceDB.txt
../data/bg/000001.jpg
../data/bg/000002.jpg
../data/bg/000003.jpg
....
....

hd.txt(Optional)
../data/hd/000001.jpg
../data/hd/000002.jpg
../data/hd/000003.jpg
...

the hd image is hard negative for init training , the size of it should to be the same with your model template(24 for me).

The config is in common.cpp

License

BSD 3-Clause

Python Implemention

https://github.com/wincle/NPD_python

(It's much slow than C++ implementisn. PR is welcomed)

References

http://www.cbsr.ia.ac.cn/users/scliao/projects/npdface/index.html