Zhedong Zheng
Zhedong Zheng
@Excaliburyz You can use "net = vl_simplenn_move(net,'gpu')" Yes.
You can get face data from http://lrs.icg.tugraz.at/research/aflw/ following the instruction on website. In this database, it provide bounding boxes for each face in the sqlite, which you can easily find...
My email is [email protected]
@luoyetx I forget whether only professor can apply this dataset. You may refer to http://lrs.icg.tugraz.at/research/aflw/
@luanps It seems ok. You may change 12net structure code in "f12net.m" which I didn't update to the latest. There is a error in it. You may change filter size...
@ikvision yes. Sometimes there are no faces in one certain pic at certain scale. So you can not get a face detection bounding box according to the threshold. It's necessary...
@ikvision thank you! You're right. I don't change some filter's weight from gpu array to single. I think Matlab will do it automatically although it gives warning.
@Anubhav7 Do you download 'test' branch rather than 'master' branch?(https://github.com/layumi/2015_Face_Detection/tree/test) The pic folder is called "test". There are some pictures in it already.
@issac8huxley No preprocessing is needed. AFLW data is enough strong to converge and easy to train 2 class classification. You may check your data whether there are wrong labeled picture...
@ronneycjm First, you may check http://www.vlfeat.org/matconvnet/install/#compiling to compile the library on Win. There are some information. (Perhaps you should install visual studio and mex.) Or run this code by cpu...