Question: making a new dataset

[yuasaonrails]

Thank you for your great work. This tool is really amazing.
I want to tune the dataset to get better result, but I am struggling now.

My questions are creating a new dataset and usage of it: how to make it and how to use it.
I am guessing that I need to make a xml file, and I looked up the original dataset. But I couldn't find it.

Can you help me?

Steps I took:

1. I have tried to make a new dataset using animeface-ruby/face_collector.rb.
2. The script made several images.
3. I have deleted images in which face wasn't shown.

[nagadomi]

I have not published the training code of this project. several people requested me to publish training dataset, so I described how to make it.
If you want a better model, you need to develop it yourself.
dilib: http://blog.dlib.net/2014/02/dlib-186-released-make-your-own-object.html
opencv: http://docs.opencv.org/2.4.13.2/doc/user_guide/ug_traincascade.html
or some deep learning projects support training face detector.